Proceedings of the Sixth

International Exergy, Energy and

Environment Symposium (IEEES-6)

Edited by
A. MİDİLLİ and H. KÜÇÜK

1-4 July 2013

ISBN: 978-605-85878-0-9

Recep Tayyip Erdoğan University

Rize, Turkey
Contents
APPLICATION OF ENERGY AND EXERGY ANALYSIS TO INCRESE EFFICIENCY OF A 1
HOT WATER GAS FIRED BOILER

ENERGY AND EXERGY ANALYSIS OF A DIRECT STEAM GENERATION SOLAR 12

THERMAL-POWER PLANT IN CYPRUS

ENERGY AND EXERGY ANALYSES OF A NEW INTEGRATED SOLAR ENERGY SYSTEM 23

FOR HYDROGEN PRODUCTION

EXERGY ANALYSIS OF A HYBRID SOLAR-WIND-BIOMASS SYSTEM WITH THERMAL 38

AND ELECTRICAL ENERGY STORAGE FOR A COMMUNITY

ENERGY AND EXERGY ANALYSES OF AN INTEGRATED SOLAR BASED HYDROGEN 48

PRODUCTION AND LIQUEFACTION SYSTEM

ENERGY AND EXERGY ANALYSES OF A ZERO EMISSION POWER PLANT FOR CO- 58
PRODUCTION OF ELECTRICITY AND METHANOL

INVESTIGATING THE EXERGETIC AND ENVIRONMENTAL EFFECTS OF SUBCOOLING 70

AND SUPERHEATING PROCESSES ON THE PERFORMANCE OF DIRECT EXPANSION
SYSTEMS

INVESTIGATION OF ENERGY EFFICIENCY BY MAKING EXERGY ANALYSIS IN THE 81

CEMENT FACTORY

INVESTIGATION OF EXERGETIC AND ENVIRONMENTAL PERFORMANCE OF HFCs 91

WITH 20 YEAR AND 100 YEAR GWP

A COMPACT DESIGN OF WATER HEATING-HUMIDIFICATION PROCESSES FOR 101

SOLAR HDD SYSTEMS

PERFORMANCE EVALUATION OF THE BIO-CHAR HEAVY METAL REMOVAL 107

PRODUCED FROM TOMATO FACTORY WASTE

EFFICIENT ANIONIC DYES ADSORPTION ON ACTIVATED ALGERIAN CLAYS 114

EXERGOECONOMIC ANALYSIS OF A CASCADE ACTIVE MAGNETIC REGENERATIVE 121

REFRIGERATION SYSTEM

THERMODYNAMIC AND COST ANALYSES OF A RESIDENTIAL HYBRID PV-FUEL 130

CELL-BATTERY SYSTEM FOR A CANADIAN HOUSE

RETROFITTING OF A HEAT EXCHANGER NETWORK BY USING EXERGY ANALYSIS 141

COMPARISON OF THE ENERGY PERFORMANCE AND STATIC LOADS ON A BUILDING 154

INTEGRATED WIND, SOLAR AND RAIN WATER HARVESTE

INVESTIGATING UTILITY OF LATENT HEAT STORAGE SYSTEMS FOR GREEN 163

BUILDING APPLICATIONS

ENVIRONMENTAL EFFECT EVALUATION OF TOPOGRAPHY AND NATURAL GAS 173

USAGE ON AIR QUALITY: A CASE STUDY OF K.MARAŞ

I
THERMODYNAMIC ANALYSIS OF GEOTHERMALLY DRIVEN HIGH-TEMPERATURE 185
STEAM ELECTROLYSIS SYSTEM FOR HYDROGEN PRODUCTION

ENERGY AND EXERGY ANALYSES OF A COMBINED MULTIGENERATION SYSTEM 196

PERFORMANCE ANALYSIS OF A SOLAR POND FOR DIFFERENT DIMENSIONS 208

EFFICIENCY ASSESSMENT OF CRUDE OIL DISTILLATION SYSTEMS 217

THERMODYNAMIC ANALYSIS OF A SOLAR DRIVEN TRI-GENERATION SYSTEM FOR 225

BUILDING APPLICATIONS

THERMODYNAMIC ANALYSIS OF AN AUTOMOTIVE AIR CONDITIONING SYSTEM 237

USING R1234YF and R134a

PARAMETRIC ANALYSIS OF PUMPED STORAGE HYDROPOWER-COUPLED 251

WIND TURBINE PLANTS

DESIGN OF A MICRO HYBRID TRIGENERATION SYSTEM FOR MEETING ENERGY 263

DEMANDS OF AN ENERGY EFFICIENT HOUSE USING RENEWABLE ENERGY SOURCES

ASPHALTENE AS LIGHT HARVESTING MATERIAL IN DYE-SENSITIZED SOLAR CELL 275

DETERMINATION OF A GEOTHERMAL ENERGY FIELD WITH AUDIO- 283

MAGNETOTELLURIC (AMT) DATA AT THE SOUTH OF MANISA

ENERGY PRODUCTION FROM MUNICIPAL SOLID WASTE USING PLASMA 291

GASIFICATION

THE IMPLEMENTATION AND ANALYSIS OF A THERMOELECTRIC GENERATOR FOR 299

GEOTHERMAL AND WASTE HEAT AREAS

ELECTROSYNTHESIS, CHARACTERIZATION AND ELECTROCATALYTIC PROPERTIES 305

OF Pt-Fe NANOPARTICLES TOWARDS THE OXIDATION OF METHANOL AND
FORMIC ACID

A NEW CARBON NANOTUBE-SUPPORTED PT-RU ANODIC CATALYST BY REVERSE 312

MICROEMULSION FOR DIRECT METHANOL ELECTRO-OXIDATION

ELECTROOXIDATION OF C1-C4 ON THE PT-SN ALLOY NANOPARTICLES MODIFIED 318

CARBON-CERAMIC ELECTRODE

PREPARATION AND CHARACTERIZATION OF NANOSIZED NiCo/Al2O3 CATALYSTS 325

PREPARED BY POLYOL METHOD FOR PARTIAL OXIDATION OF METHANE

CAPACITIVE PROPERTIES OF POLYPYRROLE/CARBON NANOTUBE COMPOSITES 337

ENERGY SAVING WITH DOUBLE-SKIN GLAZED FACADES IN MULTI-STOREY 349

BUILDINGS

WIND ENERGY RESORCE ASSESSMENT OF ERGAN MOUNTAIN SKI CENTER - 357

ERZINCAN
ASSESSMENT OF DESALINATION TECHNOLOGIES INTEGRATED WITH RENEWABLE 369
ENERGY SOURCES IN TURKEY

II
TORREFACTION OF AGRICULTURE AND FORESTRY BIOMASS USING TGA-FTIR-MS 386

A SOLAR ENERGY CALCULATION STUDY FOR THE BUILDINGDS IN BAYBURT IN 395

ORDER TO GET OPTIMUM BENEFIT FROM THE SUN DIRECTLY

PREDICTION OF DAILY AVERAGE GLOBAL SOLAR RADIATION AND PARABOLIC 406

MONTHLY IRRADIATION MODEL PARAMETERS FOR TURKEY

COMPARISON OF ALTERNATING-CURRENT LOSSES IN TWO-LAYER 413

SUPERCONDUCTING CABLES CONSTRUCTED BY SHELL-TYPE AND SOLID-CORE
CYLINDRICAL WIRES

MULTI MODAL STRUCTURE FOR THE MANAGEMENT OF ENERGIES IN A 420

RESIDENTIAL HOME

QUANTUM DOTS CONJUGATED E. COLI LIVING CELLS AS FLUORESCENT 429

REPORTERS TO DETECT CYTOTOXICITY OF CHEMICALS

THE EVALUATION OF TURKISH ELECTRICITY DEMAND PROJECTIONS 433

EX EXPERIMENTAL STUDY OF THIN LAYER DRYING BEHAVIOR OF A FISH ( 442

MYKISS)

THE USE OF TRANSPARENT INSULATION MATERIALS AS ONE OF THE MEASURES 450

OF IMPROVING THE ENERGY OF THE STRUCTURES

EVALUATION OF BIOHYDROGEN PRODUCTION POTENTIAL FROM SEWAGE 459

SLUDGE

DIRECT PRODUCTION OF HYDROGEN AND SYNTHESIS GAS FROM SOUR NATURAL 469
GAS: NOVEL TRENDS

A DYNAMIC SIMULATION STUDY OF A SMALL SCALE HYDROGEN PRODUCTION 474

SYSTEM FOR A HIGH TEMPERATURE PROTON EXCHANGE FUEL CELL

IONIC CONDUCTIVITY INVESTIGATION OF Ce1-xMxO2-x/2 (M=Sm, Gd, Nd, La) 485

BASED ELECTROLYTES FOR IT-SOFCs

ENHANCED BIOHYDROGEN AND 5-ALA PRODUCTIONS WITH THE OPTIMIZATION 495

OF CULTURE CONDITIONS

OPTIMIZATION OF THE OPERATION CONDITIONS IN A DIRECT BOROHYDRIDE 500

FUEL CELL WITH CARBON SUPPORTED Au ANODE

DESİGN AND ASSESSMENT OF A NET ZERO ENERGY HOUSE 509

NUMERICAL INVESTIGATION OF THE HEATING DISTRIBUTION EFFECT ON THE 526

BOILING FLOW IN THE BUBBLE PUMPS

OPTIMIZED ANALYSIS OF COLD ENERGY UTILIZATION FOR COLD STORAGE 534

PROJECT OF XINGTAN LNG SATELLITE STATION
A COMPUTER BASED ANALYSIS OF SINGLE-STAGE VAPOR ABSORPTION 543
REFRIGERATION SYSTEM

PERFORMANCE EVALUATION OF A FLOWING-ELECTROLYTE DIRECT METHANOL 554

FUEL CELL THROUGH NUMERICAL MODELING

III
INVESTIGATION OF HUMIDITY EFFECTS ON THE THERMAL COMFORT AND HEAT 560
BALANCE OF THE BODY

INFLUENCES OF FERROMAGNETIC DEFLECTORS BETWEEN LAYERS OF 573

SUPERCONDUCTING POWER TRANSMISSION CABLES ON TRANSPORT CURRENT
LOSSES

THERMAL COMFORT ANALYSIS OF NOVEL LOW EXERGY RADIANT HEATING 579

COOLING SYSTEM AND ENERGY POTANTIAL COMPARING TO CONVENTIONAL
SYSTEMS

EXPERIMENTAL INSULATION PERFORMANCE EVALUATION OF AEROGEL FOR 591

HOUSEHOLD REFRIGERATOR

ECOLOGICAL PERFORMANCE ANALYSIS OF IRREVERSIBLE BRAYTON CYCLE 602

SHORT AND LONG-TERM SOLAR RADIATION ESTIMATION METHOD 611

ENERGY AND EXERGY ANALYSES OF COPPER-CHLORINE (Cu-Cl) BASED 620

INTEGRATED SYSTEMS FOR HYDROGEN PRODUCTION

EXERGETIC PERFORMANCE OF A LOW BYPASS TURBOFAN ENGINE AT TAKEOFF 629

CONDITION

DEVELOPING SOME NEW EXERGETIC INDICES FOR ASSESSING THE PERFORMANCE 640
OF BUILDINGS

EXERGETIC SUSTAINABILITY ANALYSIS OF A SINGLE LAYER SOLAR DRYING 648

PROCESS

EXERGY ANALYSIS OF WIND TURBINE SYSTEM: A CASE STUDY OF BERGAMA, IZMIR 655

ENERGY AND EXERGY ANALYSIS OF A PERLITE EXPANSION FURNACE 665

MEASURING ENERGY EFFICIENCY OF A RESIDENTIAL AREA USING EXERGY 672

ANALYSIS: A COMPARISON OF MAVIŞEHIR I AND MAVIŞEHIR II, IZMIR

PERFORMANCE ASSESSMENT OF A SMALL SOLAR POND STRATIFIED WITH 684

MAGNESIUM CHLORIDE WATER

PERFORMANCE ASSESSMENT OF A RECUPERATIVE HELIUM GAS TURBINE SYSTEM 692

COMPARISON OF DIFFERENT TURBULENT MODELS IN TURBULENT FORCED 705

CONVECTIVE FLOW AND HEAT TRANSFER INSIDE RECTANGULAR CROSS-
SECTIONED DUCT HEATING AT THE BOTTOM WALL

PERFROMANCE EVALUATION OF ECO-FRIENDLY HUDMIDIFYING MATERIAL USING 714

HWANGTO
THE DEVELOPMENT OF WALL SYSTEMS TO IMPROVE THE ENVIRONMENTAL 720
PERFORMANCE OF TRADITIONAL BUILDINGS

THE USE OF MAGNETIC PROPULSION SYSTEM IN PLACE OF INDUSTRIAL 727

CONVEYORS

IV
FLOW AND HEAT TRANSFER CHARACTERISTICS OF AN EMPTY REFRIGERATED 736
CONTAINER

CERTIFICATION AND MAINTENANCE OF INDUSTRIAL INSTALLATIONS IN THE 748

COMPANY TREFILEST ACCORDING TO STANDARD ISO 9001 /2000

PLASMA-FUEL SYSTEMS UTILIZATION FOR ECOLOGICAL AND ENERGY EFFICIENCY 763

OF THERMAL POWER PLANTS

PLASMA TECHNOLOGIES OF SOLID AND GASEOUS FUELS PROCESSING 776

EXERGY ANALYSIS OF AN ETHYLENE GYLCOL RECOVERY SYTEM USING BATCH 789

DISTILLATION

INVESTIGATION OF EXERGY RATIOS OF A SOLAR POND AT VARIOUS REFERENCE 796

TEMPERATURES

ENERGY AND EXERGY ANALYSIS OF AN R134A AUTOMOTIVE HEAT PUMP SYSTEM 808
FOR VARIOUS HEAT SOURCES IN COMPARISON WITH BASELINE HEATING SYSTEM

EXERGY AND ENERGY EVALUATION OF DISTRICT HEATING SYSTEM 820

EXERGETIC PERFORMANCE ASSESSMENT OF A BINARY GEOTHERMAL POWER 827

PLANT

EXERGETIC EVALUATION OF A HIGH-PRESSURE HYDROGEN PRODUCTION SYSTEM 838

POLYPYRROLE/CARBON SUPPORTED Platınum CATALYSTS FOR PEM FUEL CELLS 846

COMPARISON of NATURAL GAS FIRED and INDUCTION HEATING FURNACES 852

THE USE OF HAZELNUT OIL ETHYL ESTER AS A FUEL IN PRE-CHAMBER DIESEL 859
ENGINE

DEVELOPMENT OF A REDUCED MECHANISM FOR N-HAPTANE FUEL IN HCCI 869

ENGINES

THREE DIMENSIONAL NUMERICAL MODELLING OF HYDROGEN COMBUSTION IN A 877

SPHERICAL MODEL COMBUSTOR

PRE-COMBUSTION CO2 CAPTURE BY PD BASED COMPOSITE MEMBRANE 885

DEPOSITED ON POROUS NICKEL SUPPORT
894
MATHEMATICAL MODEL OF PETROLEUM DYNAMICS IN CLOSED CONDUIT

A STATISTICAL ANALYSIS OF LEAN MISFIRES IN A GASOLINE ENGINE AND THE 906

EFFECT OF HYDROGEN ADDITION

A NEW APPROACH FOR TURBOMACHINERY MODELLING BY USING ANFIS 914

STRUCTURE

OVERVIEW OF IONIC LIQUIDS USED AS WORKING FLUIDS IN ABSORPTION CYCLES 926

FULL SCALE EXPERIMENTAL STUDIES OF A PASSIVE COOLING ROOF IN HOT ARID 937
AREAS

V
EXERGETIC ANALYSIS OF A VERTICAL GROUND-SOURCE HEAT PUMP SYSTEM 945
WITH WALL HEATING/COOLING

THE RELATION BETWEEN THE EXERGETIC PERFORMANCE OF HEAT PUMP 954

SYSTEM AND EXERGETIC PERFORMANCE OF ITS COMPONENTS

MODEL BASED PERFORMANCE ANALYSIS OF A CONCENTRATING PARABOLIC 967

TROUGH COLLECTOR ARRAY

SELECTION OF KAPLAN AND FRANCIS TURBINE IN HYDRAULIC POWER PLANTS: A 980

CASE STUDY

COMPARING CORN DRYING IN FLUIDIZED BED DRYER AND CONVECTIVE TRAY 989
DRYER

MATHEMATICAL MODELING OF A SMALL SCALE COMPRESSED AIR ENERGY 999

STORAGE SYSTEM

INVESTIGATION OF THE USE OF SOLAR THERMAL BUFFER ZONE IN BUILDINGS 1009

PERFORMANCE EVALUATION FOR DIFFERENT CONFIGURATED HRSGs 1018

THERMODYNAMIC PERFORMANCE EVALUATION OF A GEOTHERMAL DRYING 1024

SYSTEM

CHEMICAL THERMODYNAMICS OF HYDROCARBON COMPOSITIONS IN NATURAL 1035

GAS FIELD, NORTHEAST OF THAILAND

PERFORMANCE ESTIMATION OF GAS TURBINE SYSTEM VIA DEGREE-DAY METHOD 1045

COMBINED EFFECT OF VISCOUS DISSIPATION ON THE COUPLING OF CONDUCTION 1051

AND MIXED CONVECTION ALONG A VERTICAL SLENDER HOLLOW CYLINDER

THREE DIMENSIONAL NUMERICAL ANALYSIS OF THERMAL OUTPUT OF A STEEL 1063

PANEL RADIATOR

NUMERICAL SIMULATION OF COMPRESSION STROKE USING SMOOTHED PARTICLE 1073

HYDRODYNAMICS

HYDRODYNAMIC ASPECTS OF OSCILLATING FLOW THROUGH POROUS MEDIA 1084

CONSISTING OF STEEL SPHERES

AN EXPERIMENTAL STUDY ON ENHANCEMENT OF HEAT TRANSFER IN A SOLAR 1092

AIR HEATER COLLECTOR BY USING POROUS MEDIUM

NOVEL FIN DESIGN FOR THE PANEL TYPE RADIATORS USING CFD 1100
EFFECT OF RADIATION-CONDUCTION INTERACTION ON MIXED CONVECTION 1112
FROM A SLENDER CYLINDER

CONCEPTUAL DESIGN OFSMALL WIND TURBINES WITH MORPHING BLADE 1119

EVALUATION OF THERMAL CHARACTERISTICS OF A BOREHOLE THERMAL ENERGY 1128

STORAGE SYSTEM

VI
INVESTIGATION OF DRYING CHARACTERISTICS OF BLUEBERRY USING A NEW 1141
SOLAR AIR COLLECTOR DESIGN

INVESTIGATION OF TURBIDITY EFFECT ON TRANSMISSION OF HEAT IN A SOLAR 1148

POND

ENHANCEMENT OF GAS TURBINE PERFORMANCE USING COOLED INLET AIR AND 1158
SIMULATION WITH EES SOFTWARE
1165
A STUDY ON SINGLE LAYER DRYING CURVE EQUATIONS

VII
 APPLICATION OF ENERGY AND EXERGY ANALYSIS TO INCRESE EFFICIENCY OF A HOT
WATER GAS FIRED BOILER

Milena Todorović, Dragoljub Živković, Marko Mančić, Gradimir Ilić

Faculty of Mechanical Engineering, University of Nis in Serbia
Street: Aleksandra Medvedeva 14, 18000 Nis, Serbia
milenatod1@yahoo.com

ABSTRACT
In engineering practice exergy can be used for technical and economic optimization of energy conversion
processes. The problem of increasing energy consumption suggests that heating plants, i.e. hot water boilers
as energy suppliers for household heating should be subjected to exergy and energy analysis. Heating
plants are typically designed to meet energy demands, without the distinguished difference between quality
and quantity of the produced heat. In this paper the energy and exergy analysis of a gas fired hot water
boiler is conducted. Energy analysis gives only quantity results, while exergy analysis provides insight of the
actually available useful energy with respect to the system environment. In this paper, a hot water boiler was
decomposed into control volumes with respect to its functional components. Energy and exergy of the
created physical model of the hot water boiler is performed and destruction of exergy and energy loss in
each of the components are calculated. It is found that the greatest exergy destruction is in flame pipe, while
the greatest energy loss is a consequence of hot flue gases flow to the environment through the chimney.
The paper describes the current state of energy and exergy efficiency of hot water boiler. Additionally, the
largest energy and exergy losses are identified. The obtained results are analyzed and used to investigate
possibilities for improvement of availability and reliability of the boiler. A comparison between actual and
proposed more reliable solution is made.

Key words: hot water boiler, exergy, energy, availability, reliability

INTRODUCTION
Improving energy efficiency as well as energy saving represent one of the major problems of the modern
developed countries worldwide. Hot water boilers, common in district heating systems, should be subjected
to reliability, safety and efficiency research (Hasanuzzaman et al., 2011). The efficiency of hot water boiler
has a large impact on thermal performance in district heating systems. In order to improve efficiency, heat
transfer from flue gases to the water is increased to reduce energy losses of hot water boilers. In order to
optimize the processes in hot water boilers, it is necessary to pinpoint areas with the greatest losses (Saidur
et al., 2010). Since hot water boilers operate in conditions of high temperature and pressure, in addition to
their efficiency, it is necessary to consider safety and reliability issues as well, in order to avoid the major
disruption and downtime in operation (Katavić et al., 2007).
The first law of thermodynamics is mainly used for the analysis of energy utilization, but with the inability to
analyze qualitative aspects of used or consumed energy. In these cases the exergy analysis becomes
relevant. Exergy represents the consequence of the second law of thermodynamics. The concept of exergy
defines the maximum available work of a body or system, with clearly defined control volumes. Exergy can
also be seen as measure of deviation between the current state of a system and its equilibrium with
environment. Exergy analysis appears in modern thermodynamics as a very important tool, especially for
technical thermodynamics (Bejan et al., 1996). The complete analysis of thermodynamic characteristics of
the process usually requires using both energy and exergy analysis, since they provide different information
about the characteristics of the process. Such comprehensive analysis represents the suitable approach of
the performance of the system and finding ways to improve the process (Kaushik et al., 2011). A substantial
number of papers can be found in the literature dealing with energy and exergy analysis of power plants, i.e.
steam boilers using solid fuel. The paper that presents the comparative analysis of steam boilers in Turkey
(Hassan et al., 2009) uses this comparison of energy and exergy analysis. The performance of the power
plant is determined by modelling certain parts of the plant and detailed analysis of energy and exergy losses
considering the functionality of the power plant in Jordan (Aljundi et al., 2009). Detailed exergy analysis of
steam blocks were studied, decomposing the plant into several areas, considered independently (Datta et
al., 2007). The analysis of these steam boilers also included measurement of boiler losses, as well as the
turbine plant (Naterer et al., 2010).
Exergy and energy analysis is used in the cycles that use natural gas. Impact on the exergy efficiency has
been observed in relation to the input parameters of pressure and temperature of the gas turbine in the
combined cycle natural gas fired power plants (Reddy et al., 2007). The simplified method of exergy analysis

1
based on the second law of thermodynamics is presented by (Can et al., 2010) and has been used to
analyse the useful output of Rankin cycle, with respect to the parameters of exergy analysis of gas turbine.
As previously stated, the comprehensive analysis of energetic balance, as well as detailed exergy analysis
could provide better insight into the process, as well as identify new solutions to improve the process.
Understanding energy and exergy efficiency is crucial for designing, analyzing, optimizing and improving the
energy systems through variety of measures and strategies for sustainable development (Kanoglu et al.,
2007). Exergy analysis is intended to determine the maximum work ability of the system as well as to identify
the places where losses of “useful energy” occur, i.e. exergy. Identifying the locations where the exergy
losses occur, where destruction of exergy takes place, cause of its destruction, the real scale of exergy
losses represent the directions of future potential improvements of the system and its components (Kanoglu
et al., 2005). According to the previously discussed examples from the literature it is clear that exergy
analysis is essential for energy planning, resource optimization and the global reduction of harmful emission
of greenhouse gases.
In this paper energy and exergy balance of gas fired hot water boiler is presented. A model of an actual
operating boiler is created. The boiler is subjected to failure issues due to thermo-mechanical strain and
stress of the piping on the back of the flue gas diverter chamber. Decomposition of the boiler is done by
determining control volumes corresponding to the functional units of the boiler, thus creating a physical
model of the boiler. The analysis presents energy and exergy efficiency of each part of the hot water boiler,
as well as the destruction of exergy and energy losses in these components. In this way, the paper describes
the properties of energy efficiency of hot water boiler component wise, and identifies the areas with the major
energy and exergy losses and proposes. The obtained results are used to address the reliability and
availability issues of this type of boilers, caused by severe thermal stress in the piping of the flue gas
distribution chamber. A modified physical model of the boiler which should insure improvement in terms of
boiler reliability and availability is analyzed using the same methodology, and results are compared to the
actual state model of the boiler. Performance of the boiler without the cooling pipes is analyzed, and
efficiencies are compared to the original model performance.

THE ENERGY AND EXERGY ANALYSIS - METHODOLOGY

The term exergy was first mentioned at the conference in Lindau in 1953 by Z. Rant. Name of exergy
originates from the prefix “ex” before the root “ergy”, and the meaning of from and work and characterize the
work that can be obtained from the system (Baehr, 1965). This indicates that exergy represents the
maximum amount of work that can be extracted from the system in the process of establishing equilibrium
with its environment.
Opportunity to acquire work from the process whenever there is pressure, temperature, concentration, speed
and elevation gradient between the system and the surrounding area. As the system changes the state in the
direction of the environment, availability of work decreases, ceasing to exist when equilibrium is achieved.
The referent state is defined as the state of mechanical, thermal and chemical equilibrium between the
system and its surroundings. In addition, the relative speed and elevation of the system in relation to the
environment are equal to zero. Under these conditions there is no possibility of spontaneous changes within
the system or the environment, or there is no possibility to have any interaction between them. The second
type of equilibrium state between the system and the environment can also be identified. This is the limited
form of equilibrium, where only the conditions of mechanical and thermal equilibrium are satisfied and it is
called limited referent state.
Energy analysis is one of the most important aspects of engineering analysis. Energy can be transformed
from one form to another and can be transferred from one system to another, but the total amount of energy
remains constant. The general balance of the system can be written as:
INLET + CREATION – OUTLET – CONSUMPTION = ACCUMULATION
Inlet and outlet are related to the quantities that going in and out of the system boundaries. General
equations of equilibrium can be written for the mass and the energy balance:

(1)

(2)

The total exergy of the system can be represented as the total amount of physical exergy , kinetic exergy
, potential exergy and chemical exergy (Bejan,1998):

(3)

2
Analogue to this, the total exergy of the system can be expressed in specific value of exergy :

(4)

Physical exergy of the closed system is given:

(5)

or:

(6)

Kinetic and potential exergy are given:

(7)

(8)

Typically, kinetic and potential exergy are ignored in exergy analysis.

The chemical exergy of the system represents the measure of the distance of the observed chemical
composition from the composition of the system environment. Standard chemical exergy is based on the
standard value of environmental temperature and pressure , for example and
. The referent composition of ambient air is given in Table 1.

Table 1. The molar composition of the atmospheric air (Erlach et al., 2001).
Component Nitrogen N2 Oxygen O2 Carbon dioxide CO2 Water H2O (g)
The molar share % 77.48 20.59 0.03 1.9

In order to determine the chemical exergy of the fuel it is necessary to know all the mutual relations of
chemical elements in the fuel. However, in thermodynamic practice there is no need to have high accuracy,
hence semi-empirical formulas are commonly used to determine the approximate fuel exergy. Rant proposed
to determine chemical exergy of fuel through relations . The Rant method of determining chemical
exergy of liquid and gaseous fuels is used in this paper (Rant, 1961).
Exergy of gaseous hydrocarbons can be expressed by Rant method and by using the approximate equations
for gas hydrocarbons (Szargut, 1964):

(9)

Where:
– ratio of number of atoms of hydrogen and carbon in the molecule,

– reciprocal value of the number of carbon atoms in the molecule of fuel.

In the process of combustion of fuel, necessary quantity of oxygen for combustion is provided by bringing the
right amount of air. The air, with respect to pressure and temperature, possesses certain exergy, or its
exergy is assumed zero if its state is in equilibrium with environment. Exergy of air can be calculated from
the general expression of the exergy (Baehr, 1961):

(10)

In hot water boiler thermal plants the following processes can be distinguished: fuel combustion and the heat
exchange with the working fluid, i.e. water. In combustion processes chemical energy is converted into
thermal energy of combustion products. Combustion process is, as well as many other processes,
irreversible in nature. On one hand side, this is caused by irreversible mixing of reaction components and on
the other by a low medium temperature of combustion. Irreversibility leads to significant portions of input fuel
exergy to the combustion process to degenerate into useless energy – anenergy. The energy in the
combustion process, just as in any other process, remains constant while the exergy decreases transforming

3
into anenergy from which the work cannot be achieved. Calculating exergy of combustion products is quite
complex and can be also calculated according to the general form of exergy Eq.(10).
In the process of combustion exergy losses are significant. It is necessary to introduce an indicator that will
directly or indirectly represent their value. If the values of fuel exergy and exergy of combustion
products are known, the exergy efficiency of the combustion process can be determined as the ratio of
exergy of combustion products and fuel exergy:

(11)

The exergy efficiency is dimensionless quantity that indicates which part of the fuel exergy appears to
be useful after the combustion process, i.e. the quantity defines the exergy losses in the process.
Besides combustion, the most common process is the process of heat transfer from hot gases to the water in
boiler. Components of hot water boiler, where the process of heat transfer take place, can be considered as
cross heat exchangers. The process in heat exchanger is complex because it simultaneously runs two
different processes, heat transfer and fluid flow. Since ideal working bodies (fluids) cannot be related to
natural (real) process, there are real gases and liquids instead, during process of fluid flow exergy losses will
occur inevitable as consequence of friction in the fluid and between the fluid and the surface of heat
exchangers (Malić et al., 1970). Another source of exergy losses in the heat exchanger occurs due to finite
temperature differences during heat transfer. For calculating energy and exergy losses, standard equations
for exergy Eq.(10) and general equations for mass and energy balance Eq.(1) and Eq.(2) are used in the
paper. As for exergy efficiency in heat exchangers, so-called conventional exergy efficiency is used.
Conventional exergy efficiency represents the simplest form of exergy efficiency, and its formulation is based
on the exergy balance of incoming and out coming flows. It represents the ratio of total output and input of
exergy flow:

(12)

where reflects the irreversibility of the process, and other forms of exergy in the control volume are
included with incoming and out coming exergy.
Conventional exergy efficiency can be written in the form of:

(13)

GENERAL DESCTRIPTION OF THE HOT WATER PLANT

Thermal plant “Technical faculties“ in Nis, with total capacity of 25,7 MW, uses natural gas fired hot water
boiler type “TE110V”, manufactured by “Minel – kotlogradnja“. Below is the Table 2 with basic specifications
of hot water boiler.

Table 2. Technical data of the boiler

Technical specification of the boiler „Minel – kotlogradnja”
Boiler type “TE 110V“
Capacity 8700kW
Efficiency 0.91
Furnace resistance 10.7 mbar
Mass 41000 kg
Boiler length 8060 mm
The temperature level of water 130/70 C
Maximum allowed overpressure 12 bar
Operating pressure 6 bar
Total surface for heat transfer 351 m2
Water content in boiler 27,8 m3

The boiler is shaped as a cylindrical tube, closed with chambers on both sides and thermally insulated all
over the volume. The boiler has three pressurized gas channels. The flame pipe (first pass), located in the
pressurized water body, is heated by thermal energy generated by combustion of natural gas. After
combustion, exhaust gasses pass through a diverter chamber coated by water piping, go into the second

4
pass gas piping (II pass gas pipes), placed above the flame tube. On the front side of the boiler is a frontal
diverting chamber that redirects the exhaust gasses and leads them into the third pass gas piping, located
on the sides of the boiler (III pass gas pipes). After leaving III pass gas pipes, exhaust gasses splash the
water-cooled front of the diverting chamber, passing through the deflection chamber and then leaving the
chamber through the chimney. During operations hot water boilers are completely filled with water normally.
As for the water flow, the boiler is being filled with water on the bottom side of the boiler, below the flame
tube. After filling the boiler, the portion of the water goes into the water-cooled front of reversing chamber,
while the remaining volume of water is heated by radiation and convection passing through flue pipes (flame
tube, gas pipes II and III pass). This water flow at the exit of the boiler is mixed with the water from the water-
cooled front of diverting chamber and is then pumped to the consumers.

Fig.1. Schematic overview of the hot water boiler “Minel - kotlogradnja“

Fig.2. Current state of the hot water boiler “Minel - kotlogradnja“ – front view

Fig. 3. Schematic representation of water-cooled front hot gas reversing chamber (left), the current state of water-cooled pipes (right)

5
Gas and water flow of the boiler can be distinguished, as shown below. The simplified view of the gas flow of
the boiler is on the Fig. 4.

Fig. 4. Gas flow of the hot water boiler

Since the boiler volume is full of water, water flow of the boiler can not be precisely distinguished, because
there are no clearly defined boundaries of the boiler’s component. For further analysis the only logical
arrangement of the components of the water flow is adopted, which follows the arrangement of the
components of the gas flow where the water flows. So the adopted arrangement of components of water flow
is presented on Fig. 5.

Fig. 5. Water flow of the hot water boiler

For nominal operation mode and for each of the components of hot water boiler, shown in Fig. 4. and Fig. 5,
as control volume, quantity of the irreversibility of the process (exergy loss) and exergy efficiency (Table 3)
can be defined. Exergy efficiency of the total cycle can be defined in different ways. Table 3 represents
irreversibility of the process of heat transfer from the combustion products to the hot water, and the
destruction of exergy during the combustion of natural gas and internal friction of the fluid flow (water and
gas), as well as exergy loss with flue gas leaving the boiler (Kanoglu et al., 2007).

Table 3. Exergy analysis of the components of the hot water boiler

Component Exergy loss Exergy efficiency
Flame tube

Water-cooled front of reversing chamber

Gas pipes II pass

Gas pipes III pass

Total cycle

Energy analysis for each component of the hot water boiler and defined quantities can be represented (Table 4).

6
Table 4. Energy analysis of the components of the hot water boiler
Component Energy loss Energy efficiency
Flame tube

Water-cooled front of reversing chamber

Gas pipes II pass

Gas pipes III pass

Total cycle

RESULTS AND DISCUSSION

When the analytical results are achieved, the following assumptions are adopted:
1. Water and gas temperature field are uniform on the control boundaries of each component of the hot
water boiler.
2. The air temperature on the control surface is equal to the ambient temperature, so its exergy value is
zero.
3. The entire amount of thermal energy in flame tube and in the irradiated chamber is transferred by
radiation.
4. As the gas speed is quite large compared to the size of the front diverter chamber, it is considered that
there is no heat transfer between the gas and water in the front chamber.
The calculation results for each component are given in Table 5.
As observed, the total energy efficiency for complete hot water boiler is 89.9%, while the exergy efficiency is
only 60.97%, respectively in the entire system comes to the destruction of exergy of 39.03%, which is
irreversibly lost in the environment.

Table 5. The results of efficiency calculations for each of the components of hot water boiler
Energy efficiency Exergy efficiency Exergy destruction
Component
Flame tube 97.00 45.65 54.35
Water-cooled front of reversing chamber 99.00 65.57 34.43
Gas pipes II pass 99.00 84.16 15.84
Gas pipes III pass 99.00 93.52 6.48
Total 89.895 60.97 39.03

In order to make the representation of results more clear and to see what is going on in the system, it is
necessary to consider energy and exergy loss in relation to the system in general (Table 6).

Table 6. The results of calculations for each of the components of hot water boiler in relation to the entire system
Component Energy loss Energy loss Exergy loss Exergy loss
Flame tube 126.495 12.57 5497.087 73.06
Water-cooled front of reversing chamber 8.674 0.86 1212.384 16.11
Gas pipes II pass 34.993 3.48 440.308 5.85
Gas pipes III pass 5.456 0.54 151.565 2.01
Flue gases leaving the boiler 830.752 82.55 223.157 2.97

Based on the given results of energy and exergy analysis, it can be concluded that the greatest exergy loss
occurs in the flame tube of the hot water boiler, while the largest energy loss of the system takes place in the
area where the gases leave the boiler. The exergy loss through each of the components of the gas flow of
the hot water boiler can be traced. Exergy loss, obtained analytically, amounts from 73.06% in the flame
tube, 16.11% in water-cooled front of the reversing chamber, 5.85% in gas pipes II pass, 2.01% in gas pipes
III pass, while exergy loss of flue gases leaving the boiler through the chimney is 2.97%. The major exergy
loss occurs in the flame tube, which was expected, because of the chemically irreversible process of fuel
combustion. Exergy of exhaust gasses depends on the temperature at which the combustion heat is
transferred to the flue gases. Due to the large temperature difference of combustion products in the flame
pipe at high temperatures (1350 - 1450 C) and hot water heated to 90 - 100C, and heat transfer done by
radiation, the greatest exergy destruction can be related to the flame pipe.
It may be also noted that, following the gas flow of the boiler, the reduction of exergy losses and increasing
of exergy efficiency takes place. Based on the obtained results it can be seen that the trend of increasing

7
exergy efficiency is not present in the water-cooled front, which can be explained by different type of heat
exchangers. Namely, II and III pass gas pipes are associated to cross-section exchangers, where water
flows around the pipes, while in the water-cooled front of reversing chamber water flows along the pipes,
while exhaust gasses flow around the pipes, directed by diverter chamber. Analytic analysis of energy and
exergy balance shows that the greatest potential for increasing overall efficiency of the plant is actually in the
flame tube of the hot water boiler. Fig. 6 graphically represents the distribution of energy and exergy loss for
each of the components of the hot water boiler.

Energy and exergy analysis

90,00%
80,00%
70,00%
60,00%
50,00%
40,00% Energy loss [%]
30,00% Exergy loss [%]
20,00%
10,00%
0,00%
Flame pipe Water-cooled front Gas pipes II pass Gas pipes III pass Leaving flue gases
hot gas reversing
chamber
Fig. 6. Comparative graphical review of energy and exergy analysis of hot water boiler

PROPOSED IMPROVEMENT
The water-cooled pipes on the back of the flue gas diverter chamber is exposed to high temperature strain,
which results in damage of this part of the diverter chamber (Fig. 7) and drop outs of the boiler. This fact
reduces availability and reliability of the boiler. In this paper, affect of modification of this part of the
construction of the boiler was investigated and its impact on the overall efficiency of the boiler by means
exergy and energy analysis. The proposed solution analyzed in this paper is retrofit design of the diverter
chamber, by omitting the piping from the walls of the chamber. In this case, the complete volume flow of the
water would flow over the flame pipe, and II and III pass gas pipes in the main body of the boiler. It is
assumed that an equal amount of delivered energy and exergy is delivered with the fuel and air to the boiler.
In the case of modified contraction, due to lack of the water-cooled pipes on the walls of the diverter
chamber, available heat is transferred completely to the main body of the water, and the flue gasses after
leaving the III pass gas pipes, are redirected through the deflection chamber through the chimney. Although,
the same amount of energy and exergy is assumed to be transferred to the water body, in the modified case
scenario, the exhaust gasses leave the boiler at higher temperature level, since their heat is not rejected to
the water flowing through the reflection chamber wall piping. This results in slightly higher energy and exergy
losses for the analyzed retrofit boiler model. Fig. 7 shows the current state of damaged wall of water-cooled
pipes of the diverting chamber. Table 7 shows a breakdown of energy and exergy efficiencies with respect to
proposed model components. Fig. 8 shows energy and exergy ratio of losses to the total amount of losses
associated to each of the boiler model components.

8
Fig. 7. Damaged water-cooled pipes of the diverting chamber of the hot water boiler (left and right)

Table 7. The results of efficiency calculations for each of the components of hot water boiler without water-cooled pipes
Energy efficiency Exergy efficiency Exergy
Component
destruction
Flame tube 97.00 41.63 58.37
Gas pipes II pass 99.00 67.13 32.87
Gas pipes III pass 99.00 98.99 1.01
Total 89.93 57.27 42.73

Table 8. The results of calculations for each of the components of hot water boiler without water-cooled pipes in relation to the entire system
Component Energy loss Energy loss Exergy loss Exergy loss
Flame tube 145.438 14.93 5740.507 80.44
Gas pipes II pass 34.993 3.59 1345.974 18.86
Gas pipes III pass 5.456 0.56 27.751 0.39
Flue gases leaving the boiler 788.549 80.92 21.75 0.30

Energy and exergy analysis

90,00%
80,00%
70,00%
60,00%
50,00%
40,00% Energy loss [%]
30,00% Exergy loss [%]
20,00%
10,00%
0,00%
Flame pipe Gas pipes II pass Gas pipes III pass Leaving flue gases

Fig. 8. Comparative graphical review of energy and exergy analysis of hot water boiler without water-cooled pipes of reversing chamber

DISCUSION
The results presented in Table 6 and Table 8 are summarized in Fig. 6 and Fig. 8, respectively. By
comparing these results, it is clear that the distribution of component specific energy loss and exergy
destruction remains the same in the proposed boiler construction compared to the actual state. However,
energy and exergy leaving the boiler through the chimney has now increased, since the temperature of the
flue gasses is higher in the proposed model than that of the actual state model. Energy efficiency of the
proposed boiler model has slightly changed, from 89.895% for the actual state mode to 89.93% for the
proposed boiler model. As per Table 5 and Table 7, energy efficiencies of specific boiler components

9
remained the same. Total boiler exergy destruction has changed from 39.03% for the actual state model, to
42.73% for the proposed model. The greatest difference of exergy loss between the two models is found in
the flame pipe, which has changed from 73.06% for the actual state model to 80.44% for the proposed state
model, compared to the total exergy loss of the system. Additionally, the output of the boiler has changed for
the proposed model as a consequence of reduced heat transfer to the water in the piping on the back of the
flue gas diverter chamber by flue gasses flowing along the piping after leaving the III pass gas pipes on the
way to the chimney. Hence, the consequence of the proposed boiler modification is twofold: energy loss has
decreased in the flame pipe, and in the chimney which can be associated to the flue gasses leaving the
boiler. The exergy loss of the proposed modified model is slightly higher in the flame pipe and in the chimney
flue gasses, while exergy losses in the II and III pas gas pipes are redistributed in a new way.

CONCLUSION
According to the performed analysis, it is shown that energy and exergy analysis can be used to pinpoint
energy loss, exergy loss and exergy destruction in the structure of hot water boiler plants. In order to pinpoint
locations of these losses, it is necessary to create a physical model of the boiler, by applying decomposition
of the boiler with respect to function of each of its components. The greatest exergy loss occurs in the flame
tube of the hot water boiler, while the largest energy loss of the system takes place in the area where flue
gases leave the boiler. The major exergy loss in the flame tube is the consequence of irreversible chemical
combustion process on one side, and large temperature gradient between combustion products and hot
water heated by radiation. Significant energy losses occur in the chimney, since the flue gasses leave the
boiler at a high temperature level.
The obtained results are used to investigate possibilities of design modification to address reliability and
availability issues, caused by regular damage of water-cooled pipes in the gas diverter chamber. The
proposed case model was analyzed by means exergy and energy analysis, to determine the impact of
omitting this part from the diverter chamber contraction to the boiler efficiencies. It is found that the
modification affected the following: energy loss has decreased in the flame pipe, and in the chimney
associated to the flue gasses leaving the boiler. The exergy loss is slightly higher in the flame pipe and in the
chimney flue gasses, while exergy losses in the II and III pas gas pipes are redistributed in a new way.
Overall efficiency of the boiler decreased slightly, but it is anticipated that such construction will provide
better reliability and availability of the boiler. However, exergo-economy optimization should be addressed to
these issues to further investigate viability and profitability of the proposed solution and help in decision
making.

NOMENCLATURE
specific exergy, kJ/s
total energy rate, kJ/s
2
Earth's standard acceleration due to gravity, m/s
specific enthalpy, J/kg
exergy loss rate, kJ/s
mass flow rate, kg/s
pressure, Pa
heat transfer rate to the system, kJ/kg
specific entropy, J/kgK
temperature, K
thermal energy, W
3
specific volume, m
velocity, m/s
work rate or power done by system, W
distance, m
Greek Letters
efficiency, -
total exergy rate, kJ/kg
Subscripts
dead state conditions
chemical exergy
combustion product
exergy factor

10
 fuel
flame pipe
gas pipes with convection heating surface
energy factor
inlet
kinetic exergy
losses in the process
outlet
physical exergy
potential exergy
water-cooled front hot gas reversing chamber
total
Superscripts
combustion process
gas
water

REFERENCES
Hasanuzzaman, M., R. Saidur and N. A. Rahim. 2011. Energy, exergy and economic analysis of an
annealing furnace. International Journal of the Physical Sciences 6:1257-1266.
Saidur, R., J. U. Ahamed and H. H. Masjuki. 2010. Energy, exergy and economic analysis of industrial
boilers. Energy Policy, Elsevier 38:2188-2197.
Katavić, B and B. Jegdić. 2007. Analysis of demages on water boiler shield pipes. Zavarivanje i zavarene
konstrukcije 4:123-130.
Bejan, A., G. Tsatsaroni and M. Moran. 1996. Thermal design and optimisation, John Wiley & Sons, Inc.,
New York.
Kaushik, S. C., V. Siva Reddy and S. K. Tyagi. 2011. Energy and exergy analyses of thermal power plants: A
review. Renewable and Sustainable Energy Reviews, Elsevier 15:1857-1872.
Hasan, H. E., V. A. Ali, Burhanettin, D. Ahmet, H. S. Suleyman, S. Bahri, T. Ismail, G. Cengiz and A. Selcuk.
2009. Comparative energetic and exergetic performance analyses for coal-fired thermal power plants in
Turkey. International Journal of Thermal Sciences 48:2179–2186.
Aljundi Islam, H. 2009. Energy and exergy analysis of a steam power plant in Jordan. Applied Thermal
Engineering 29:324–328.
Datta, A., S. Sengupta and S. Duttagupta. 2007. Exergy analysis of a coal-based 210MW thermal power
plant. International Journal of Energy Research 31:14– 28.
Naterer, G. F., P. Regulagadda and I. Dincer. 2010. Exergy analysis of a thermal power plant with measured
boiler and turbine losses. Applied Thermal Engineering 30:970–976.
Reddy, B.V. and K. Mohamed. 2007. Exergy analysis of natural gas fired combined cycle power generation
unit. International Journal of Exergy 4:180–196.
Can Gulen, S. and R. W. Smith. 2010. Second law efficiency of the rankine bottoming cycle of a combined
cycle power plant. International Journal of Engineering for Gas Turbines and Power 132:1–10.
Kanoglu, M., I. Dincer, and M. A. Rosen. 2007. Understanding energy and exergy efficiencies for improved
energy management in power plants. Energy Policy 35:3967–3978.
Kanoglu, M., S. K. Isik and A. Abusoglu. 2005. Performance characteristics of a Diesel engine power plant.
Energy Conversion and Management 46:1692–1702.
Baehr, H. D., VDI-Gesellschaft Enegietechnik. 1965. Energie und Exergie; die Anwendung des
Exergiebegriffs in der Energietechnik. Düsseldorf.
Bejan, A. 1998. Advanced engineering thermodynamics. New York: Wiley Interscience.
Erlach, B., G. Tsatsaronis and F. Cziesla. 2001. A New Approach for Assigning Costs and Fuels to
Cogeneration Products. Internal Journal of Applied Thermodynamics 4:145-156.
Rant, Z. 1961. Zur Bestimmung der spezifischen Exergie von Brennstoffen, Allgem. Wärmetechn 10:172-
176.
Szargut, J. and T. Styrylska. 1964. Approximate evaluation of the exergy of fuels. Brennstoff-Wärme-Kraft
(BWK) 16:589–596.
Baehr, H. D. 1961. Ein Exergie-Entropie-Diagramm für Luft. Chemie Ingenieur Technik 33:335-338.
Malić, D., B. Đorđević and V. Valent. 1970. Termodinamika strujnih procesa. Građevinska knjiga, Belgrade.
Kanoglu, M., I. Dincer and M. Rosen. 2007. Understanding energy and exergy efficiencies for improved
energy management in power plants. Energy Policy 35:3967–3978.

11
 ENERGY AND EXERGY ANALYSIS OF A DIRECT STEAM GENERAT ION SOLAR
THERMAL-POWER PLANT IN CYPRUS
*1, 2
Armita Hamid Ugur Atikol
Mechanical Engineering Department, Eastern Mediterranean University, Gazimagusa, North Cyprus, Via
Mersin 10 Turkey
1*
amrita.hamidi@cc.emu.edu.tr
2
ugur.atikol@ emu.edu.tr

ABST RACT
The objective of this study is to investigate the energetic and exergetic feasibility of utilizing a solar thermal
power plant in Cyprus. The analysis carried out is two-fold. First, the efficiency of each component of an
existing steam power plant (Teknecik) is estimated using energy and exergy analyses. The results show that
the boiler of this power plant has the highest irreversibility rate due to the combustion process of Fuel oil
No.6. In the second step, it is proposed to change the conventional power plant into a direct steam
generation solar power plant. In this regards, parabolic trough collectors are used to generate superheated
steam at 87 bars and 510˚C. Moreover, the energy and exergetic efficiency of each component of the new
design has been estimated and compared with the results that obtained in the first part. It was found that the
highest exergy losses occurred in the collectors while the receiver subsystem has the maximum energy
losses. Furthermore, the average exergy efficiency of the solar field is 24% while the boiler average exergy
efficiency is 50%. The outcome clarifies that a DSG power plant has poor energetic and exergetic feasibility
due to the high rate of losses. Nevertheless, the exergy analysis identifies the cause and the location of the
losses which are worthwhile for optimizing and improving the solar field design and can make the solar
thermal power generation competitive with conventional technologies for producing electricity in large
scales.

INTRODUCTION
Due to recent increase in the price of carbon-based fuels and the concerns about environmental degradation
caused by carbon emissions, it is expected that renewable energy sources become more viable. Exploiting
solar energy for producing power has already shown to work in various parts of the world. In this regard,
employing solar energy for producing electricity in large capacities not only reduce emissions, but also by
applying economic optimization and optimization it can be cost effective. Cyprus as the third large
Mediterranean island geographically is located on a favorable position for exploiting solar energy. Regarding
to metrological data, average hourly sun beam intensity varies from 250 to 800 (Fig.1)

Currently solar energy is used only for heating water in approximately 90% of residences and 50% of hotels
(Cyprus Institute of Energy, 2009). This makes Cyprus top place around the world with installed flat plate
solar collectors per inhabitant.
However, electricity production is still relied on imported fuels. In north Cyprus, 70 million USD$ is spent
annually for importing fossil fuel for power generation. (KIB-TEK, 2009). Oil import in South Cyprus receives
about 62% of the export earnings (Koroneos et al., 2005). Therefore, utilizing solar thermal power plants for
producing electricity will also have positive impacts on the balance of trade and security

of supply that can result in economic growth. Worldwide experience shows that, setting up solar thermal
technology in one square kilometer of land is enough to generate approximately 110 GWh of electricity per
year which is comparable to the annual production of a 50 MW fossil-fired mid-load power unit (Hepbasli,
2008). Since solar thermal power plants (STPPs) are spreading all over the world, there is a number of
different technologies developed recently for performing this purpose. However, there is still room for
improving their design and performance.
The first concentrated solar power plants were established in the 1980s in the Mojave Desert of California,
later grown into the largest solar power plant in the world with a capacity of 354 MW. It is possible to design
a parabolic trough STPP based on the Direct Steam Generation (DSG). Other alternatives require a Heat
Transfer Fluid (HTF) and a Heat Recovery Steam Generator (HRSG) between the solar field and the power
unit, which produces heat losses additionally. Although DSG can be said to be a new technology, it is a
promising option to improve the efficiency of the whole system. Generally, the evaluation of the thermal
power plant performance is based on the energy analysis originated from the first law of thermodynamics.
This can demonstrate the electrical power and thermal efficiency. However the second law analysis was

* Corresponding author
12
 Fig.1. Hourly average direct normal irradiation (DNI) for each month
found to be more helpful in the design, optimization, improvement, and evaluation of thermal power plants
(Kaushika et al., 2011). In 2001, Dincer et al. (2001) analyzed a Rankine cycle steam power plant with
reheater. They have performed the energy and exergy analysis at different operating conditions with varying
boiler pressure, boiler temperature, mass fraction ratio and work output from the cycle. Since then several
studies have been done on different kinds of power plants by different scientists such as Fischer (1996),
Ameri (2008), Cihan (2006) and Rosen (2006). While developing and replacing alternative energy sources,
the goal of efficient use of these kinds of energy resources is also promoted. In this regard, Singh (Singh et
al., 1994) and Bannister (1998) presented simple models for solar thermal power collector systems based
on the second law of thermodynamics in order to estimate the thermodynamic losses. However, their
methods are complicated and not easily repeatable and hence, could not be further developed any further
other systems. Later, in 2009, Singh et. Al. (2009) added the analysis based on the second law (exergy
concept) to a solar thermal power system using a parabolic trough collector system which connected to a
Rankine heat engine cycle for power generation to evaluate the actual available exergy and second law
efficiency. More recently, Gupta and Kaushik (2010) performed exergy analysis for 5 MW DSG power plant
with various feed water heaters to minimize the exergy losses and improving the efficiency.

The aim of the present study is to utilize the energy and exergy techniques to evaluate the feasibility of
converting the conventional power plant in Cyprus into a solar thermal power plant. Since the parabolic
trough mirrors are proven all over the world, the study will consider this technology as the heating medium
for the working fluid. The design of the solar field in this study will be based on DSG technology using
parabolic solar trough collectors to produce 50 MW capacity. Energy and exergy analyses which are based
on the first and second laws of thermodynamics respectively are to be carried out with the aim of
determining the location and magnitude of the process irreversibilities in order to shed light to optimization of
the system.

ENERGET IC AND EXERGET IC ANALYSES

The first law of thermodynamics or energy balance for the steady flow process of an open system is given
by:

∑ ̇ ∑ ̇( ) ∑ ̇ ( ) ̇ (1)

The energetic efficiency of system is defined as:

(2)

13
The exergy of heat transfer from the control surface at temperature T is governed from maximum thermal
energy converts to work:

( ⁄ ) (3)

is the ambient temperature.

The exergy flow for steady flow process of an open system is given by:
∑ ̇ ∑ ̇ ̇ ∑ ̇ ̇ (4)

Where exergy is expressed as:

[ ] (5)

denotes the total energy in the system which is:

(6)

The exergy destroyed is proportional to the entropy generated:

̇ ̇ (7)

The exergy or second law efficiency is defined as:

(8)

For the evaluation of the fuel exergy, which is the corresponding ratio of simplified exergy is defined as
follows:

⁄ (9)

For gaseous fuel with , the following empirical equation is used to calculate (Kotas, 1985):

(10)

For the solar field, energy received which is received by the collector system is estimated by:

(11)

Where B is the width of the aperture of the collector, is the beam radiation falling on horizontal surface,
is the tilt factor, is the number of collector rows and is the number of collectors in each row. The tilt
factor is evaluated from:

(12)

is the angle of incidence and is zenith angle.

⁄
[ ] (13)

(14)

14
Declination is also calculated by:

⁄ (15)

The total exergy received by the collector system is computed by:

[ ( ) ( ) ] (16)

is apparent black body temperature of the sun which is approximately 5600K.The energy and exergy
which is absorbed by the receiver/absorber of solar collector field is as following:

(17)

Where is the optical efficiency at the design point.

⁄ (18)

Which is the mean temperature of the absorber.

(19)

is the heat loss from the collector which Takes place in the absorber. For reducing the heat loss, absorber
tube is enveloped with vacuum glass tube. Depending on the thermal resistances between the absorber
tube surface and the surroundings, the heat loss coefficient is calculated iteratively (Duffie et. al., 1991)
by solving the following equations (Gupta et. al., 2010):

́ (20)

́ ́ (21)

́ ́ (22)

́ ́ [ ( )] (23)

is the wind heat transfer coefficient and is the emissivity of glass cover. The emissivity of the
coating of receiver in terms of is taken equal to:

(24)

The correlation for in terms of is obtained as:

(25)

The useful heat which is transferred to water flowing through receiver tube is calculated by:

(27)

The useful exergy is given by:

[ ] (28)

15
 SYST EM DESCRIPTION
Teknecik power plant has two steam units each having installed capacity of 60 MW. The present study
considers a 50 MW power plant with similar characteristic to Teknecik power plant for comparison puposes.
For this reason the technical data available for the Teknecik power plant will be useful in our analysis. The
boiler consumes 15 tons/ hour of fuel oil No.6 for producing 50 MW power. The steam enters the turbine at
510ºC. Table 1 shows the thermodynamic properties of each point of the steam power cycle (defined in Fig.
2). Cooling water enters the condenser at 28.5ºC and mass flow rate of 2305 kg/s. The analysis has been
performed for a randomly selected day in each month during a year in order to observe the results for
different seasons and temperatures.
Generator

Heat source
(boiler or
solar field)

Fig. 2. Schematic of 50 MW steam cycle power plant

Table 1.Stream data for 50 MW steam power cycle

Point* m˙(kg/s) T(ºC) P(bar) h(kJ/kg) s(kJ/kg.K)

P-1 63 510 87 3415.1 6.709

P-2 41.892 40.81 0.077 2262.57 7.246
P-3 44.968 40.81 0.07706 170.95 0.5831
P-4 44.996 41.22 13 173.67 0.588
P-5 44.996 42.81 13 180.34 0.6091
P-6 50.511 69.41 13 291.54 0.947
P-7 3.323 84.41 13 353.5 1.127
P-8 50.511 104 13 436.81 1.351
P-9 63 136.61 3.28 574.51 1.704
P-10 63 138.22 105 588.2 1.711
P-11 9.695 148.22 0.0045 624.77 1.8231
P-12 63 174.23 105 742.9 2.071
P-13 5.802 184.23 0.001 782.47 2.1805
P-14 63 222 105 955.23 2.524
P-15 1.922 73.72 0.365 2410.89 7.062
P-16 3.323 108.89 1.38 2575.19 6.952
P-17 1.31 139.52 3.565 2718.82 6.9
P-18 6.463 234.26 9.9 2907.93 6.861
P-19 6.252 338.54 24.89 3102.06 6.799
P-20 2430 28.5 0.03229 119.52 0.416
P-21 2430 37.57 0.0653 157.4175 0.5398

16
 In North Cyprus, Fuel oil No.6 is used for electricity generation. Table 2 gives the fuel properties and its
composition

In the second part, the solar thermal field is studied as an option for replacing the boiler. A 50MW STPP is
designed based on the DSG technology. The maximum temperature of the solar field coincides with the
steam-cycle temperature and steam is directly sent to the power unit. Therefore, no heat exchanger or boiler
is considered in this design. In the present design, the solar collector field consists of 76 loops of ET-100
collectors. There are 10 collectors in each row which are arranged as following: 3 preheating collectors, 5
evaporating collectors and 2 superheating collectors. The total collector field area is estimated 421956.86
m². The parameters of the ET-100 parabolic-trough collector and their design-point parameters used in this
study are given in Table 3. The produced steam enters the turbine of the Rankine cycle at 510ºC and 87
bars. Location coordinates of Cyprus is 35.16° N and 33.35° E. The wind heat transfer coefficient is
assumed to be 29 W/m².K.

Table 2. Fuel Properties and components

Fuel type fuel oil No.6
Lower calorific value (LHV) 41 MJ/kg
Higher calorific value (HHV) 43 MJ/kg
Density 15 ton/hour
Carbon 87.87%(molar mass)
Hydrogen 10.33%(molar mass)
Sulphur 1.16%(molar mass)
Nitrogen 0.14%(molar mass)
Oxygen 0.5%(molar mass)
Fuel exergy 42.078 MJ/kg

Table 3. Design-point parameters of the ET-100 parabolic-trough collectors and their field arrangement
Number of collectors in each row Nc 10
Number of modules per collector 12
Number of collector rows in the solar field Nr 76
Gross length of each module 12.27 (m)
Overall length of a single collector L 98.5 (m)
Inner/outer diameter of steel absorber pipe Di/Do 0.055/0.07 (m)
Inner/outer diameter of glass cover Dci/Dco 0.125/0.130 (m)
Aperture width 5.76 (m)
Collector aperture 548.35 (m²)
Optical efficiency ηₒ 0.75
Intercept factor 0.92
Mirror reflectivity 0.92
Glass transmisitivity 0.945
Solar absortivity 0.94
Thermal emissivity 0.04795+0.0002331*T(ºC)

RESULTS AND DISCUSSION

It is evident from the first and second law analysis that the condenser and boiler have the lowest efficiency
among the other components. Moreover, the boiler is the major exergy destructor due to the chemical
reaction between air and fuel in the combustion process. Fig. 4. reveals that only 2% of exergy loss happens
in the condenser. As obtained from the second law analysis, exergy efficiency is almost constant at 50%
only slightly changing during the year, but the irreversibility rate has a linear increase from 86 MW to 91 MW
when the temperature rises from 17ºC to 37ºC. In other words, for each degree centigrade rise in the
temperature, the irreversibilities in the boiler increase by 0.5 MW.

Contrary to the second law analysis, this demonstrates that substituting the boiler system with a more
efficient heat source will provide more chances in enhancement of the overall efficiency of the plant.
Therefore, the solar thermal field is studied as an option for replacing the boiler. The 50MW STPP is
redesigned based on DSG. DSG technology avoids the use of a boiler in the power section since steam is
directly generated in the solar field and the maximum temperature of the solar field coincides with the steam-
cycle temperature. The heat loss coefficient ⁄ which is correlated with ranging from 350 to
800 K is calculated by solving one dimensional model using EES software. This value is slightly lower than

17
 0.387 0.9
Boiler 0.8
% 5% 0.7
2%
0.6
9% Turbine 0.5
0.4
0.3
Condenser 0.2
0.1
0
84% Pumps

Feed water
(a) heaters (b)

Fig. 4. (a) Percentage of irreversibility share of each component of 50 MW STPP

(b) exergy efficiency of each component of the fossil-fuel power plant

the true value since in the present study the conduction losses of some other components of the solar field,
for instance receiver support brackets, is not considered.Fig. 5(a). has shown that, except the two first
months of the year, the average energy losses of the receiver are higher than the collector subsystem.
However, fig. 5(b). indicates that the exergy losses of the collectors are quite higher than the receivers. The
solar field energy and exergy losses both occur their maximum and minimum values in June and February,
respectively. The energy efficiency of the receiver is lower than the collector except for the two first months
of the year. Unlike the energy efficiency, exergy efficiencies of the receivers are discernibly greater than that
of the collectors for the first two and last three months of the year while it fluctuates around the 45% value
during the rest of the year.

The changes in energy and exergy losses during a day are presented in fig.7 and fig.8, respectively. For the
colder months which include January, October, November and December, it is observed that between 11
am to 2 pm the rate of losses drop since the DNI is increased. On the other hand, this value rises between
the mentioned hours for the warmer months which are June, July and August. It is notable to mention that
the energy and exergy losses have a direct relation with the rate of radiation which received by the solar
field which is not just affected by DNI. The radiation received by the solar field depends on the tilt factor and
the declination δ which are the functions of and n.

400,000 450000
350,000 400000

300,000 350000
300000
Exergy loss(kw)

250,000
Energy loss(Kw)

250000
200,000 collector
200000
150,000 150000 receiver

100,000 100000 collector-

50000 receiver
50,000
0
September…

December…
November…
January 24th

August 16th
February 17th

June 11th
May 15th
March 26th

October 29th
April 15th

July 17th

0
September…

November…
December…
February…

August 16th
June 11th
January 24th

May 15th
March 26th
April 15th

October 29th
July 17th

(a) (b)

Fig. 5. (a) Average energy losses of the solar field (b) Average exergy losses of the solar field
18
 80%
60%
40%
20% energy efficiency

0% Exergy efficiency

DNI(W/m²K)
Fig. 6. Variation of energy and exergy efficiency of the solar field with DNI

800000
October 29th
700000
600000 November 14th
500000
Energy loss (kW)

December 24th
400000
January 24th
300000
200000 February 17th
100000 March 26th
0
8 9 10 11 12 13 14 15 16
Time
(a)
Energy loss (kW)

Time
(b)
Fig. 7. Variation of solar field energy losses on hourly basis for selected days in (a) colder months (b)
warmer months

19
 800000
750000
700000
650000 January 24th
600000
550000 February 17th
500000
Exergy loss (kW)

450000
400000 March 26th
350000
300000 October 29th
250000
200000 November 14th
150000
100000 December 24th
50000
0
8 9 10 11 12 13 14 15 16
Time
800000
750000 (a)
700000
650000
600000 April 15th
550000
Exergy loss (kW)

500000 May 15th

450000
400000 June 11th
350000
300000 July 17th
250000
200000 August 16th
150000
100000 September 28th
50000
0
8 9 10 11 12 13 14 15 16
Time
(b)
Fig. 8. Variation of solar field exergy losses on hourly basis for selected days in (a) colder months (b)
warmer months
60%
50%
Exergy efficiency

40%
30%
20% Boiler
10%
solar-field
0%

Fig. 9. Solar- field exergy efficiency versus boiler exergy efficiency during one year

Since a simple design of the solar field is considered, there is a high rate of the losses. Fig. 9. compares the
exergy efficiency of the solar field and the boiler. In contrast to the boiler, the rate of variation of exergy
efficiency of the solar field during the year is very high. This fact improves an urge of an optimization in the
solar field design and components. The maximum exergy efficiency of the solar field which is the closest to
the boiler occurs in February with 47%.

20
CONCLUSIONS AND SUGGESTIONS FOR OPTIMIZATION
The results demonstrate that the main exergetic loss occurs in the boiler followed by the condenser.
Therefore, the main reasons of the irreversibilities are the combustion of the fossil fuel and the heat transfer
processes. Once the boiler is replaced by the solar field, the greater energy and exergetic power losses
occur in the collector-receiver assembly. Unlike the energy losses, exergy losses of the collector are greater
than that of the receiver. Furthermore, the energy and exergy efficiency of the STPP decrease while DNI
rises due to the effect of incident angle. While the incident angle is related to the position and design
parameters besides the tilt of the solar beam radiation, it demonstrates that significant improvement chances
exist in the solar system rather than the boiler. Following are several suggestions for optimizing the energy
and exergy efficiency of the STPP conducting from results of this study:
 Collectors in the solar field have high energy losses. Increasing the number and changing the
arrangement of collectors can improve the energy efficiency. To reduce exergy losses in collector,
material constraints play an important role and hence, extensive work in this direction is to be carried out
to make STPP a real success.
 Receivers have the most irreversibility due to the heat transfer. Therefore, changing the material or the
length and diameter of the receiver tube can enhance the exergy efficiency of the receiver.
 Exergy of a system is carried out directly from the properties of the input and output stream flow of the
system. Thus studying the effects of changing one or more properties e.g. pressure, will be effective.
Moreover, the temperature of water at inlet to row of parabolic-trough collector must be optimum.
 Employing the boiler in the existing steam power plant as an auxiliary heater
 Increasing the temperature inlet water to the solar field by increasing the number of feed water heaters

NOMENCLATURE
Absorber area, m2
Receiver area, m2
Aperture width, m
Concentration ratio
Specific heat, kJ/kg.K
Receiver outside diameter, m
Receiver inside diameter, m
Glass cover inside diameter, m
Glass cover outside diameter, m
̇ Exergy rate, kW
Gravitational acceleration
Solar beam radiation (DNI), W/m2
̇ Irreversibility, kW
Thermal conductivity of glass cover, W/m K
Enthalpy, kJ/kg
Length of the collector, m
̇ Mass flow rate, kg/s
Number of collectors
Number of rows
̇ Heat transfer rate, kW
Solar power input to parabolic trough, kW
Solar power absorbed by receiver/absorber, kW
Useful thermal power gain by thermic fluid, kW
Tilt factor
Entropy, kJ/kg.K
Receiver temperature, K
Sun temperature, K
Ambient temperature, K
Loss coefficient of the solar field, W/m2.K
̇ Work rate, kW
Width of parabolic trough collector, m
Velocity, m/s
Elevation, m

21
REFRENCES
Ameri M., Ahmadi P., Khanmohammadi S., 2008. Exergy analysis of a 420MW combined cycle power plant.
International Journal of Energy Research 32.175–183.
Bannister P., 1998. A basic model of a high temperature solar thermal power collector system. In:
Proceedings of the Annual Conference of the Australian and New Zealand Solar Energy Society,
University of Melbourne, 17-19 November.
Cihan A., Hac hafzoglu O., Kahveci K., 2006. Energy–exergy analysis and modernization suggestions for a
combinedcycle power plant. International Journal of Energy Research 30.115-126.
Cyprus Institute of Energy 2009, http://www.cie.org.cy
Dincer I., Al-Muslim H., 2001. Thermodynamic analysis of reheats cycle steam power plants. International
Journal of Energy Research 25. 727–739
Duffie John A., Beckman William A., 1991. Solar engineering of thermal processes. Second edition. New
York: John Wiley and Sons, Inc.
Fischer DW., 1996. Searching for steam system efficiency. Plant Engineering 50. 64–68.
Gupta M.K., Kaushik S.C., 2010. Exergyanalysis and investigation for various feed water heaters of direct
steam generation solar–thermal power plant, Renewable Energy 35. 1228–1235.
Hepbasli A. 2008. A key review on exergetic analysis and assessment of renewable energy resources for a
sustainable future. Renewable and Sustainable Energy Reviews 12. 593-661.
Kaushika S.C., Reddya V., Tyagib S.K, 2011. Energy and exergy analyses of thermal power plants: A
review. Renewable and Sustainable Energy Reviews 15. 1857–1872
http://www.kibtek.com
Koroneos C., Fokaidis P., Moussiopoulos N.. 2005. Cyprus nergy system and the use of renewable energy
sources. International Journal of Energy 30:1889-1901.
Kotas Tj. 1985. The Exergy Method of Thermal Plant Analysis. Butterworths: London.
Rosen M., Tang R., 2006. Effect of altering combustion air flow on a steam power plant: energy and exergy
analysis. International Journal of Energy Research 31. 219-231
Singh N., Kaushik SC., 1994. Technology Assessment and Economic Evaluation of Solar Thermal Power
Generation: A State of Art Report. IIT, Delhi.
Singha N.,*, Kaushikb S.C., Misrac R.D., 2009. Exergetic analysis of a solar thermal power system.
Renewable Energy 19.135-143.

22
 ENERGY AND EXERGY ANALYSES OF A NEW INTEGRATED SOLAR ENERGY SYSTEM
FOR HYDROGEN PRODUCTION

C. Zamfirescu and I. Dincer

University of Ontario Institute of Technology, 2000 Simcoe St. N, Oshawa ON L1H 7K4 Canada
E-mails: Calin.Zamfirescu@uoit.ca and Ibrahim.Dincer@uoit.ca

Abstract
In this paper, a novel integrated system for solar radiation conversion to hydrogen is assessed using energy
and exergy analysis methods. The system combines photocatalysis, photovoltaics and thermal engine for
better solar energy harvesting and higher efficiency. The solar light is split into three spectra using optics and
dielectric mirror, namely: (i) the upper spectrum, consisting of high energy photons with wavelength shorter
than 500 nm, is used for photocatalytic water splitting to generate hydrogen, (ii) the middle spectrum with
wavelength between 500 nm and ~800 nm is used to generate electricity with photovoltaic (PV) arrays, and (iii)
the lower spectrum with wavelength longer than 800 nm is used to generate electricity with a thermally driven
organic Rankine cycle (ORC) heat engine. The electricity generated by PV and ORC is employed to generate
additional hydrogen by electrolysis and to recycle intermediate chemical species used within the photocatalytic
process. The energy and exergy balances of the integrated system and its auxiliaries are presented. A model is
developed based on conservation equations and transport equations applied for each essential component of
the system. The model allows for assessment of system performance via energy and exergy efficiency and the
comparison with other solar hydrogen production systems with respect to cost.

Keywords: Solar hydrogen production, system integration, photocatalysis, organic Rankine cycle, exergy
analysis.

INTRODUCTION
Hydrogen production with solar energy is intensely studied in recent years. New potential methods emerged
such as photo-catalytic and photo-electrochemical as reviewed in Zamfirescu et al. (2011a). According to
Kromer et a. (2011), the US Department of Energy (DoE) has identified solar tower technology combined with
thermochemical water splitting as the most promising one. The DoE established the hydrogen production goal
with concentrated to $6/kg for 2015 and $2-3/kg for 2025. In a review of solar tower system Romero et al.
(2002) show that this technology is on the verge of commercialization as is proven feasible for sites with annual
insolation of around 2 MWh/m2.yr. At least 14 remarkable prototype plants covering a range of thermal power
from 2 to 30 MW were successfully demonstrated in the last 30 years. A system lifetime of 45 years,
comparable to nuclear and conventional fossil fuel power plants, seems to be achievable.

The major cost item of concentrated solar tower systems is represented by a heliostat field which, according to
Vogel and Kalb (2010), takes approx. 40 % of the investment. With a simple calculation one can estimate
roughly the required cost margin for square meter of heliostat in order to achieve the goal of DOE of $2/kg of
hydrogen. Thus, with 2 MWh/m2.yr incident radiation one calculates the total solar energy per square meter of
heliostat for the entire system lifetime to 324 GJ/m2. In terms of HHV, this energy is equivalent to that of 2,285
kg of hydrogen, which is further equivalent to a total worth of $4,570. With an interest rate of 4% (which is
common for large scale projects of hundreds of MW range) the present worth – corresponding to investment for
45 year time span – is 4,570/1.0445 = $782. In tis regard, the heliostat cost must be cheaper than
per m2 for the ideal case (all incident energy converted to hydrogen).

The DoE’s projected solar-to-hydrogen conversion efficiency, according Komer et al. (2011), becomes in the
range of 30%–40% for 2025. Consequently the cost of the heliostat must be reduced below $90/m2 for the
DoE’s goal to be achieved. It is arguably justified in Vogel and Kalb (2010) that such low costs can be obtained
well by mass production of heliostat units. A parallel strategy to speed-up the large scale solar hydrogen
technology commercialization is by finding alternative routes of improving the efficiency while keeping similar
investment cost. An increase of hydrogen generation efficiency to 45% will allow for heliostat cost of $120/m2.

According to Dincer and Zamfirescu (2012), it appears that the decisive factor for success of large-scale solar
hydrogen production is the harmonious integration of key technologies and multiple valuable products
generation while keeping the system simplicity. Previously, the authors studied various multigeneration
systems with concentrated solar radiation. A potential cogeneration of power and heating is presented in
Zamfirescu et al. (2012a). A thermodynamic analysis of fuels reforming to hydrogen which integrates thermo-

23
catalytic reformers, solar concentrators, solar receivers and turbomachines is given in Wagar et al. (2011).
Photochemical water splitting with combined reactors for hydrogen and oxygen evolution is discussed in
Zamfirescu et al. (2011b). A photo-electro-chemical chlorination system which integrates concentrated solar
power with spectral splitters and an original electrochemical cell that generated H 2 and CuCl2 is analyzed in
Zamfirescu et al. (2012b). The feasibility of a dual photo-electrochemical solar collector for hydrogen, oxygen
and heating generation is studied in Zamfirescu et al. (2012c). A review of potential photo-catalytic and photo-
electrochemical light-driven hydrogen production methods that can be integrated with other system operating
with concentrated solar radiation can be found in Zamfirescu et al. (2011b).

In this paper, an integrated system that combines the photocatalysis, photovoltaics and Rankine cycle for better
solar energy harvesting and higher efficiency is presented. Potentially, the solar spectrum is better harvested is
instead of converting the concentrated radiation in high temperature heat one generates hydrogen using high
energy photons of upper spectrum, one generates electricity with photovoltaic arrays using middle spectrum
photons and one converts only the high temperature heat associated to lower spectrum photons to electricity
using Rankine cycle. In addition, the proposed system has a very high capacity factor due to its ability of
seasonal storage of solar energy in form of convertible chemical energy. This is achieved by an intrinsic
integration of chemical energy storage with a Rankine cycle. Furthermore, the system provides a valuable high
purity sulfur as byproduct, resulting in more cost effectiveness. A feasibility study based on energy and exergy
analyses is presented.

SYSTEM DESCRIPTION
A block diagram is presented in Figure 1 which describes the proposed system. Solar radiation harvested by
the heliostat field (#1) is focused on the aperture of a hot mirror placed at the top of solar tower. The hot mirror,
acting as spectral splitter consisting of on-face coated dielectric mirror which diverts (by reflection) the lower
spectrum portion (#4) of the concentrated radiation (#2) toward a volumetric solar receiver placed at the tower
foot. The transmitted radiation (#3) is dispersed and partially scattered in an enclosed room at the back of the
hot mirror. The room is paved at inside with PV arrays covered with band-pass coating.

8 9 10 16

17
PV-array
Photocatalytic Alkalyne
5 6 (Middle
Scattered middle
reactor system Upper/Middle electrolyzer
spectrum
(Upper spectrum Spectral Splitter (high pressure
absorption)
(PV-array)
absorption) & temperature)

7
11 15
3

2
Hot mirror Water Electric power
Heliostat field 20
spectral splitter management conditioning
18 system system

19
4
1 14

Volumetric
Integrated
receiver Chemical energy
12 13 Rankine power
(Lower spectrum storage system
plant
absorption)

Fig. 1: Conceptual block diagram of the integrated system for solar hydrogen production.

24
Owing to the coating, the PV arrays reflect back the upper spectrum while absorbing only in the range of ~ 500
– 800 nm. The upper spectrum radiation (#5) is intercepted by an assembly of photoreactors placed in the
same enclosure (room). The photoreactors consist of cylindrical tubes in glass, covered with high band pass
coating which reflects wavelengths longer than ~500 nm. The reactors comprise photocatalysits with bandgap
at 500 nm and generated hydrogen from water. Whence, the optical arrangement inside the enclosure is such
that it acts as a upper/middle spectral splitter which divides the incoming upper/middle spectrum light in (#3)
into a upper spectrum (#5) and middle spectrum (#6).

The photocatalytic reactor system consumes sulfides, sulfites and hydrogen sulfide in aqueous form, provided
in #7 (see Figure 1). Note that these materials are aqueous wastes from chemical and petro-chemical
industries. From these aqueous wastes the photocatalysis system generates two valuable products: hydrogen
(#8) and high purity sulfur (#9). The residual stream (#10) is free of H2S, and comprises only lower value
thiosulfates.

The electricity generated by the PV-array (#11) is directed toward the electric power conditioning system
placed on the ground. This unit has the role to supply electric power at right parameters to all power consuming
units of the overall system. The unit has two power inputs, namely from PV generator (available during
daylight) and a Rankine plant generator (24 h).

The lower spectrum radiation (#4) is focused on a volumetric receiver which generates a high-temperature,
high-power heat flux (#12) that is transferred to the chemical energy storage system. The chemical energy
storage system is simple and cheap: it operates based on reversible thermo-catalytic reaction
. The storage system operates with a solar multiplicity of at least three, meaning that it stores two times
more energy that it delivers to the Rankine plant in (#13). This allows for continuous operation of the Rankine
plant which delivers power to the conditioning unit in (#14). The conditioning unit supplies with power a large
scale alkaline electrolyser station which continuously generates hydrogen (#16) and oxygen (#17). A water
management system processes fresh water (#18) and provides it to the process (#19, #20).

ANALYSIS
Energy and exergy balance equations must be written for all system components to eventually form a closed
system and determine the exergy destructions, the exergy corresponding to produced hydrogen and the overall
system efficiency. Of particular interest is the light harvesting and processing subsystem consisting of heliostat
field, hot mirror spectral splitter, and upper/middle spectrum splitter.

The exergy input from solar light will be determined first based on the effective photon temperature derived
in a recent work by Chen et al. (2008). Accordingly, and the correspondent Carnot factor of a monochromatic
photonic radiation with wavelength is

(1)

where , taken here 298.15 K is the reference temperature for exergy calculations.

The solar radiation is polychromatic consisting of a broad spectrum of wavelengths. In conjunction with this, a
reference spectrum must be assumed – for example, AM 1.5 solar spectrum given in ASTM G 173-03 (2012) –
in order to obtain the solar radiation entropy and exergy. Based on the assumed reference spectrum ( ), the
entropy rate of solar radiation per unit of surface, can be determined by integration as follows:

(2)

where the notion of temperature of polychromatic radiation ( ) is introduced, which depends only on the
specified spectrum ( ). In addition, the temperature ( ), and rates of energy ( ), entropy ( ) and exergy ( )
of solar radiation that enters our system (Figure 1) are given by

25
where is the total area of the reflective surface formed of heliostats each having a reflective
area , and is the direct normal spectral irradiation intensity, which can be derived from AM 1.5
spectrum.

The energy and exergy fluxes reflected by the heliostat field depend on the optical properties of the mirrors.
The heliostat reflects majority of incident radiation according to its spectral reflectance, . Conventional
heliostats use “second surface float glass wet silvered mirrors” which are very common in many harsh
environment applications. They consist of a glass sheet back coated with 700 silver layer, a 300 copper
layer and 25 black paint. The heliostat is made of many flat silver-on-glass mirrors as such, fixed on a
metallic structure to for a quasi-parabolic surface with a focal point at solar tower. The use of silver is justified
by its high reflectance in visible range which is of 95%, respectively.

A more cost effective option is of using aluminum-based stretched metallic mirrors consisting of an aluminum
foil stretched over a circular metallic frame made of a rectangular steel profile (see Kolb et al., 2007). The
aluminum foil is over-coated with a protective layer such SiO2 or TiO2 (see Palik, 1997) which confers extended
lifetime while not degrading the spectral reflectance. The spectral reflectance is calculated based on the
extinction coefficient of the metal and the refraction index according to (Palik, 1997):

(3)

Table 1: Extinction coefficient and refraction index of aluminum and silver.

Metal Wavelength
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 2.0 4.0 10.0
Aluminum 2.3 3.61 4.86 6.08 7.26 8.31 8.45 8.30 9.58 20.7 39.8 89.8
0.12 0.28 0.49 0.77 1.2 1.83 2.80 2.06 1.35 2.15 6.43 25.3
0.93 0.92 0.92 0.92 0.92 0.91 0.87 0.89 0.94 0.98 0.98 0.99
Silver 1.24 0.96 1.95 2.92 3.73 4.52 5.29 6.04 6.76 12.2 24.3 54
1.07 1.51 0.17 0.13 0.12 0.14 0.16 0.18 0.21 0.65 2.3 13.3
0.26 0.16 0.87 0.95 0.97 0.97 0.98 0.98 0.98 0.98 0.98 0.98
Source: Palik (1997)

The reflectance of silver mirrors drops sharply in UV region whereas the reflectance of aluminum remains fairly
high for all spectrums, except a small drop in the region of 800 to 900 nm. Table 1 gives the values of extinction
and refraction coefficients and reflectance for aluminum and silver. Using the reflectance data for heliostat
mirror ( ) one can obtain the entropy and the equivalent temperature of reflected radiation as follows

(4)

The energy of incident radiation of spectral splitter system (flux #2 in Figure 1) is smaller than due to a
multitude of types of optical losses which vary depending on the heliostat position in the field and on the sun
angle. The optical losses of the heliostat field can be incorporated into a compounded optical field loss factor
defined by

where the factors at left hand side refer to losses due to momentarily incident angle of the sun with respect to
reflecting surfaces, shading of surfaces, partial blocking or reflected light, atmospheric attenuation, optical
aberration and light interception by the spectral splitter, and the incident angle of concentrated radiation on the
spectral splitter surface plane, respectively.

The factor is an averaged one for all heliostats and can be determined by detailed solar field modeling
based on known heliostat positions and optical characteristics and known geographical location. Using the
compounded factor for heliostat field losses, the energy of the concentrated radiation beam at the entrance in
solar splitter module can be calculated as follows:

26
It follows that the energy and exergy balance equations for heliostat field subsystem are
. The determination of energy losses and exergy destruction for solar field allows for subsequent
calculation of efficiencies according to the following equations:

The field efficiencies can be calculated for each day of the year and for each hour of daylight. Furthermore,
average values can be determined for each month and for the year.

In this work it is assumed that the dielectric mirror that consists of a selective reflecting layer coated on the top
of a selected borosilicate glass. The mirror is planar, whence several mirror modules must be supported by a
rigid (metallic) structure designed such that a spherical reflector is formed, which acts as a downward pointing
secondary concentrator for the lower spectrum portion of solar light.

There are many types of dielectric coatings that may form reflective surfaces to IR while being well transparent
to upper/middle spectrum. Although they are more expensive, dielectric coatings are more durable than metal
coatings. The entropy and temperature of radiation (#4) reflected by the dielectric coating can be calculated
based on the specified spectral reflectance of the dielectric coating. Thus, one writes

where is the intercept factor of the secondary concentrator system.

The transmitted UV/V light is attenuated by glass substrate which has an assumed spectral reflectance (or
transmittance ). Consequently, the entropy, temperature, energy and exergy of radiation in #3 are

The energy and exergy balance equation for the hot mirror are written as follows: and
. Consequently, the energy and exergy efficiencies of the spectral splitter are:
.

The transmitted and dispersed radiation from (#3) is intercepted mostly by PV arrays installed at the inner side
of the enclosure. They will absorb and reflect radiation in accordance to the spectral reflectance of their
low-pass coating. Whence for (#6) one writes

The photocatalytic reactor intercepts radiation scattered by the surrounding PV arrays and absorbs it in
accordance to the spectral reflectance of its high-pass coating. Therefore:

The balance equations and the efficiency equations can be written for the upper/middle spectral splitter
subsystem similarly as above. These are: for energy and exergy
balance and for efficiencies of the upper/middle spectral filter splitter
(subscript fs).

27
The general system absorbs light radiation at three points (#4, #5, #6), in form of high temperature heat (#4),
excited electrons at the conduction band energy level of photocatalyst (#5, 2.48 eV) and excited electrons at
the conduction band energy of the PV-cell semiconductor (#6, ~1.6 eV). Thus, the total absorbed exergy of the
system is

where represents the thermal energy efficiency of the receiver, is the receiver surface temperature, is
the rate of photons and is the band-gap energy for photocatalyst and PV cell, respectively. The rate of
photons in a spectrum with spectral irradinace is obtained by integration according to equation:
, where is Planck’s constant and is light speed constant.

The molar rate of photocatalytically-generated hydrogen ( ) is proportional to rate of photons according to

quantum efficiency: , where quantum efficiency of the process is assumed independent of
wavelength. The authors previously investigated the use of engineered catalysts to conduct water splitting in
aqueous solution with poly-sulfides. In Zamfirescu et al. (2013) it is shown that quantum efficiency of over 60 %
can be obtained when catalysts based on CdS, ZnS and PdS are engineered. Actually, quantum efficiencies
higher than 93% were demonstrated for upper visible and near ultraviolet range (Yan et al., 2009). Assuming a
process and similar to that described in Zamfirescu et al. (2013), hydrogen can be generated from an aqueous
solution containing sulfites wastes (SO32- and ) according to:

The other approximation of the same photocatalytic process is given in Banica et al. (2012) and can be
described with the equation:

Assuming that these two reaction paths run at same rate, then the resulting overall process is described by
reaction:

The sulfur extraction requires some hydrogen sulfide feedstock. The process is detailed in Banica et al.(2012)
and is described by

Consequently, the summation of the two above equation describe the overall process:
. The (hydrosulfide ion) can be reduced photo-
catalytically to hydrogen according to a process disclosed by Linkous andMuradov (2003), described by
equation: . We obtain thus a reasonable description of the process in our
photocatalytic reactor system:

Based on this equation the molar consumption rate is determined with respect to the mole of hydrogen
generated. We have that generation of 1 mole of hydrogen and 0.22 moles of high purity sulfur requires a
feedstock of 0.889 moles of liquid water, 0.44 moles of , 0.667 moles of sulfide, 0.44 moles of sulfite and it
produces a low value aqueous waste stream comprising 0.22 moles of hydroxyl, 0.33 moles of disulfide, 0.22
moles of thiosulfate and 0.22 moles of sulfate. Note that the process works well with a large palette of
feedstock provided that sulfide ion is present; however, the amount of sulfur extraction is in direct relation to the
amount of hydrogen sulfide in the feedstock.

28
The electric power generated by PV array (#11) depends on the spectral response of the selected PV-cell and
on their basic characteristics: open circuit voltage ( ), short circuit current density ( ), filling factor ( ), band
gap energy ( ), saturation current density ( ) and internal resistance ( ). The energy and exergy in point #11
on system diagram (Figure 1) is the maximum generated power by the PV modules given by

where is the current density of the photocurrent and the spectral quantum efficiency. The open circuit
voltage is related to photocurrent according to the following equation:

In the above equations the empirical correlation for saturation current density is taken from Landsberg (2003).
The filling factor is given by the equation

Due to solar radiation intermittency and fluctuation there is a mismatch between power demand and power
generation. The role of energy storage unit is to compensate for the mismatch. The system is designed such
that allows for a continuous operation of the Rankine plant. There are known two main options for thermal
energy storage in solar tower systems: storage in a thermocline (e.g., molten salt) or as hot air compressed in
porous beds (see Vogel and Kalb 2010). The molten salt poses difficulties of flow hydraulics and raises some
stringent materials issues since ionic species can dissociate and be corrosive. Air, although intrinsically safe to
materials requires large flow rates to maintain reduced temperature differences in conditions when the heat flux
density at solar receiver can reach 1 MW/m2.

The alternative proposed here is to use ammonia decomposition reaction to harvest thermal energy from
concentrated solar radiation atop of the tower. This process is sufficiently well understood and its
thermodynamic very favorable up to pressure of 200 bar. The receiver temperature is maintained high since the
reaction evolves at constant temperature.

The use of ammonia for concentrated solar energy storage with solar dish systems has been first proposed by
Lovegrove (2004) bases on the fact that the reaction is reversible and there are no side products. The forward
reaction receives thermal energy when sun is available and the reverse reaction
delivers it on demand. The authors investigated in a past paper – Wagar et al. (2011) – the
reformation of ammonia to hydrogen with solar energy and proposed a process diagram which uses turbine for
additional power generation.

Solar heat Thermal energy

input (#12) output (#13)

b c
Thermal to chemical Chemical energy Chemical to thermal
energy conversion storage in form of energy conversion
NH3 ® 1.5 H2 + 0.5 N2 a 1.5 H2 + 0.5 N2 d 1.5 H2 + 0.5 N2 ® NH3
NH3 (l)

Fig. 2: Explanatory diagram for the thermochemical energy storage in ammonia.

In Figure 2 we give a description of the chemical energy storage sub-system integrated to our overall system.
Ammonia in liquid form (a) from the bottom of a storage tank is drawn and directed to the solar receiver. We
choose to conduct the forward reaction at 20 bar because more compact volumetric receiver are applicable at

29
that pressure (VogelandKlab, 2010). The reaction temperature is around 773 K – 873 K. For better kinetics the
conversion is maintained at 75% of the equilibrium. Thus the returning stream in (b) comprises H2, N2 and NH3.
The backward reaction proceeds with a stoichiometric amount (c) drawn from the gases atop of liquid enclosed
in a pressurized tank for storage with assumed pressures of 50 bar (empty) and 250 bar (full). The reaction is
facilitated by high pressures and will be conducted to 300 bar and 573 K – 673 K.

The energy and exergy balances of the energy storage subsystem are written as

Since the energy storage is in chemical forms the storage temperature is about the same as the surroundings
temperature; whence there are no essential thermal losses. The main losses of the system are of thermo-
mechanical nature and are due to the streams conditioning: the products of forward reaction must be
recompressed from (pressure at which the forward reaction is conducted, 20 bar) to (the pressure in
storage tank); the feedstock for the backward reaction must be compressed up to is the pressure at which
the backward reaction is conducted (~300 bar). With a assumed isentropic efficiency ( ) of compression
processes the energy losses due to a necessary compression process are estimated with

where is the molar enthalpy of the stream, is the number of moles of product gas returned from the
chemical reactor to the storage in b, is the number of moles of equilibrium mixture of H2, N2, NH3 extracted
from the storage in c (Figure 2) and is the molar entropy. Also in the above equation heat-to-work conversion
factors are assumed as the Carnot factors for temperatures and for charging and discharging of the
storage, respectively.

For the remaining sub-systems the analysis is relatively simple because it is reasonable to assume some
practical values for their efficiencies. Thus, for the Rankine cycle which operates at a source temperature it
can be assumed an energy efficiency . The energy and exergy rates in (#14) are
. The exergy efficiency is . For the electric power conditioning subsystem one can
assume efficiency higher to , which is reasonable. Therefore, . For the
electrolyzer, the typical efficiencies are in the range of , whence

RESULTS AND DISCUSSION

In order to comparatively assess the energy and exergy efficiencies of the integrated system we selected a
location from Alberta where there is a massive demand of hydrogen for oil sand processing. We selected
Calgary as a representative location for southern Alberta. The first step of study is represented by optimization
of solar field of heliostats at the selected geographical location such that the irradiance of concentrated
radiation is maximized throughout the year. Computer codes were developed in the past for performing optical
system design of heliostat fields. We used in this work DELSO3 (1986) of Sandia National Laboratory which is
included in System Advisor program package provided by National Renewable Energy Laboratory of US
Department of Energy; this program has been used in this work.

In Calgary, the annual averaged amount of beam normal irradiance is considered as 1,819 kWh/m2. A solar
receiver design with a rectangular aperture is assumed, which is exposed to concentrated radiation from a
solar field placed at north side of the tower. The heliostats consists of solar dishes with a diameter of 20 m
each, made with stretched membrane technology with aluminum mirror coated on exposed face, having an
average reflectance of 0.9 for all spectrum and optical error of 1.5 mrad. The parameters and the layout of the
optimized field for Calgary location are given in Figure 3. It is assumed that wind stow speed is 15 m/s and field
startup delay time is 12 minutes. The values of the compounded optical loss factor of solar field are
computed.

30
 350° 0° 10°
340° 20°
330° 30° Number of heliostats on each sector
320° 40°
r (m) 0° 10.9° 21.8° 32.7° 43.6° 54.5°
310° 50° 289.9 7 7 7 7 7 7
300° 60° 486.6 20 20 20 20 20 20
683.4 24 24 24 24 24 24
290° 70° 880.2 26 26 26 26 26 26
1,077 27 27 27 27 27 27

2,600
2,400
2,200
2,000
280° 80°

1,800
1,274 27 27 27 27 27 27

1,600
1,400
1,200
1,000
800
1,471 27 27 27 27 27 27

600
400
200
270° 90°
r (m) 1,667 26 26 26 26 26 26
260° Heliostat field layout 100° 1,864 26 26 26 26 26 26
2,061 26 26 26 26 26 26
250° Heliostat radius: 20 m, Heliostats number: 2,997 110° 2,258 25 25 25 25 25 25
Land : 5,526 acre, Max. radius: 2,553 m, Span angle: 130o 2,455 12 12 12 12 12 9
Reflective area: 913,289 m2, Tower height: 255 m
240° 120°
Note: r = radius to sector centre; angle = span
Focal point aperture: 29 m width x 25 m height
230° 130° angle of sector; the field is symmetric
220° 140°
Fig. 3: Optimized layout of heliostat field at Calgary for 500 MWh light intensity on aperture.
210° 150°
200° 160°
1 190°180° 170°

0.8

0.6
Ffield

0.4

0.2

0
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000
t (h)
Fig. 4: Annual variation of the compounded optical loss factor for the heliostat field.

In Figure 4 we show the variation of for each hour during the year for the optimized heliostat field at
Calgary. Using and an assumed spectral distribution of the solar light the energy and exergy of the
concentrated radiation can be computed for each hour of the year. This can be achieved by first calculating the
radiation entropy and temperature for AM 1.5 spectrum along the optical path of the solar tower system (points
#1 – #5). The calculated parameters of AM 1.5 spectrum light along the optical path are given in Table 2.

31
Table 2: Light parameters along optical path for AM 1.5 spectrum.
Point Location T (K) I (W/m2) S (W/m2K) N (photons/s)
1 Incident normal radiation 6064 903.6 0.149 3.998E+21
2 Incident on hot mirror 5979 838.7 0.1403 3.764E+21
3 Transmitted through hot mirror 9294 387.6 0.0417 1.119E+21
4 Reflected by hot mirror 4486 434 0.09673 2.595E+21
5 Upper spectrum (photo-reactor) 9089 146.6 0.01613 3.443E+20
6 Middle spectrum (PV array) 8163 216.6 0.02653 7.120E+20

The spectral reflectances of heliostat field were calculated for aluminum mirrors according to Eq. (3) and data
from Table 1. Aluminum is used due to its high reflectance in UV spectrum, needed for photocatalytic process.
For hot mirror we selected two dielectric coatings that are commercially available at ZC&R Coatings, namely
HM VS 1150 which has a transmittance better than 93 % for wavelengths shorter than 675 nm and a
reflectance better than 95% for wavelengths longer than 750 nm, and HM VS 3000 for extended IR reflection.
The compounded effect of the two coatings is estimated by superposition of their reflectance and is given in
Figure 5. In addition, the figure gives the reflectance of a selected borosilicate BK7 glass of 8 mm on which the
hot mirror coating is deposited. The reflectance spectra of the coatings selected for upper/middle spectral
splitting are also given. These are taken from ZR&C coatings catalogue: the UV cold mirror coating that reflects
90-95% at wavelengths sorter than 500 nm is used is used for PV-array coating whereas a standard
subtractive cyan filter that reflects longer wavelengths shorter than 550 nm and has a transmittance better than
85% in higher visible and UV (see Figure 5) is coated around the photocatalytic reactor glassing. The
reflectance spectra are used to estimate energy and exergy rates along the light path using the equation
detailed in the analysis section.

The concentration ratio of the secondary concentrator (hot mirror) is chosen such that the maximum irradiance
on the volumetric receiver is limited to 1MW/m2. With a reasonable intercept factor of 0.95 it is found that the
secondary concentration factor is 35; the primary concentration factor result from solar field optimization and it
is 800. Note that only about half of the radiation incident on the mirror is re-concentrated on the thermal
receiver.

The improvement factor of sunlight exergy absorption is introduced next as the ratio between the absorption
exergy efficiency of the system and the sunlight absorption exergy efficiency of a thermal only reference
system. The sunlight absorption efficiency is defined based absorption exergy rate introduced in Analysis
Section and the exergy incident on solar field . The reference system consists of a conventional solar
tower system that uses a cavity receiver of similar aperture as the new system; the solar field and geographical
location are also the same. Whence, . The improvement factor is calculated for each
operational our throughout the year and reported in Figure 7. The average value of the improvement factor is
1.102, which illustrates that at least 10 % sunlight absorption exergy efficiency is obtained with the novel
system with respect to the conventional system. The annual average exergy efficiency of light absorption is
0.45 with a standard deviation of or 29 %.

32
 1

Borosilicate BK 7 glass
HM 1150/3000 compounded
dielectric coating
ZR&C UV cold mirror coating
ZR&C standard substractive
l

cyan filter coating

R

0.1

0.009
280 500 1,000 2,000 3,000
l (nm)
Figure 5: Reflectance spectra of selected coatings for optical surfaces spectral splitting system.

1 9,000

0.9 8,000
0.8 7,000
0.7
6,000
0.6 hyear (h)
5,000
h

0.5
4,000
0.4
3,000
0.3

0.2 2,000

0.1 1,000

0 0
0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000
hoper (h)
Figure 6: Optical efficiency (energy based) of the system and of the optical subsystems for each hour of
operation ( , legend: dashed lines = efficiency of hot mirror and spectral filter splitter, filled
small circles = , large filled squares = , thick continuous line = hour from beginning of year at 1st of
January).

The temperature of radiation along the optical pathway is shown in Figure 8. The three receivers are exposed
to concentrated light with temperatures of K (thermal absorber), K (photocatalytic
reactor) and K (PV-array). The light is absorbed at temperatures of K (photocatalyst
bandgap at 500 nm), K (PV-cell bandgap at 770 nm), K (thermal receiver).

Due to better light absorption and more direct energy conversion to hydrogen by photocatalysis, the novel
system has an improved exergy efficiency of hydrogen generation with respect to a thermal only soar tower
which runs a large scale water electrolyzer.

33
 1.6

1.5
IFabs = yabs / yabsref

1.4

1.3

1.2

1.1

0.9 IFabs = 1.102 ± 0.04195

yabs = 0.45 ± 0.13
0.8
0 500 1,000 1,500 2,000 2,500 3,000 3,500
toper (h)
Fig. 7: Improvement factor of light absorption exergy efficiency vs operation time.

11,000

8,800

6,600
T (K)

4,400

2,200

0
T1 T2 T3 T4 T5 T6 TPC TPV TTH
Inci- Refl- Transm- Received Absorbed
dent ected mitted
Figure 8: Temperature of photonic radiation along the optical path ( ) and at absorption ( for
photocatalyst bandgap, for photovoltaic panel bandgap, for thermal reeiver).

34
 2

1.9

1.8
IF = y / yref

1.7

1.6

1.5

1.4
IF = 1.443 ± 0.05 y = 0.20 ± 0.059
1.3
0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000
t oper (h)
Fig. 9: Improved factor of exergy efficiency of hydrogen system with respect to a reference system based on thermal only solar energy
absorption.

It is assumed that the Rankine cycle runs with an organic fluid at 35% energy efficiency. The improvement
factor is calculated for each operational hour and plotted in Figure 9. Its average value is and the
average hydrogen production exergy efficiency determined to 20 % with standard deviation of %.

CONCLUDING REMARKS
In this paper, a novel integrated system for solar radiation conversion to hydrogen that combines
photocatalysis, photovoltaics and thermal engine for better solar energy harvesting and higher efficiency is
assessed. The integration of two additional technologies beside the thermal receiver will not modify the
investment capital in the system since the most important capital fraction is due to the solar field.

The improvement factor of light absorption with the novel system is 10% (annual average) with respect to the
reference (conventional) system. Furthermore, the overall exergy efficiency of hydrogen production increases
with 40% (annual average) with respect to a conventional tower system that run an electrolyzer. The annual
average of exergy efficiency of hydrogen generation is 20%.

Conventional solar field are typically designed for 100 MW thermal. From this paper, it appears that solar fields
of 500 MW thermal can be devised for better economics. The solar tower designed for Alberta (Calgary)
requires 5,526 acres of land, whence it generates 90 kW thermal equivalent per acre of land. If hydrogen is
produced by conversion of thermal energy into electricity with Rankine cycle followed by water electrolysis,
then the production is 34.5 t per day, whereas with the integrated system it becomes 41.4 t/day, based on the
annual averages.

NOMENCLATURE
area, m2
speed of light constant, m/s
photon constant, m/s
electron charge, C
energy rate, W
energy rate, W
factor
filling factor
Planck’s constant

35
 current density, A/m2
extinction coefficient
Botltzman constant
refraction index
photon rate, s-1
reflectance
entropy rate, W/K
voltage, V
temperature, K

Greek letters
energy efficiency
exergy efficiency

Subscript
abs absorbed
cat catalysis
wawelength
loss energy losses
oc open circuit
PC photocatalysis
PV photovoltaic
sc short circuit
TH thermal

REFERENCES
ASTM G 173-03 2012. Standard Tables for Reference Solar Spectral Irradiances: Direct Normal and
Hemispherical on 37° Tilted Surface. Internet source http://rredc.nrel.gov/solar/spectra/ am1.5/ accessed in
May 2013.
Banica R., Nyari T., Sasca V. 2012. Zn2x(CuIn)1-xS2 photocatalysts synthesis by a hydrothermal process using
H4EDTA as complexing agent. International Journal of Hydrogen Energy 37:16489-16497.
Chen Z.S., Mo S.P., Hu P. 2008. Recent progress in thermodynamics of radiation – exergy of radiation,
effective temperature of photon and entropy constant of photon. Science in China Series E:Technological
Sciences 51, 1096–1109.
Dincer I., Zamfirescu C. 2012. Renewable-Energy-Based Multigeneration Systems. International Journal of
Energy Research 36:1403-1415.
Kolb G.J, Jones S.A., Donnley M.W., Gorman D., Thomas R., Davenport R., Lumia R. 2007. Heliostat Cost
Reduction Study. Sandia National Laboratory, Report 2393.
Kromer M., Roth K., Takata R., Chin P. 2011. Support for Cost Analyses on Solar-Driven High Temperature
Thermochemical Water-Splitting Cycles. Tiax: Lexington, MA.
Landsberg P.T. 2003. Ideal Efficiencies. In: Practical Handbook of Photovoltaics: Fundamentals and
Applications. Markvart T., Cartañer L., Eds., Elsevier: Oxford.
Linkous C.A., Muradov N.Z. 2003. Close Cycle Photocatalytic Process for Decomposition of Hydrogen Sulfide
to its Constituent Elements. US Patent 6,572,829 B2.
Lovegrove K. 2004. Developing ammonia based thermochemical energy storage for dish power plants. Solar
Energy 76:331-337.
Palik E.D. (Eds.) 1997. Handbook of Optical Constants in Solids. Elsevier Inc.
Romero M., Buck R., Pacheco J.E. 2002. An Update on Solar Central Receiver Systems, Projects, and
Technologies. ASME Journal of Solar Energy Engineering 124:98-108.
Vogel W., Kalb H. 2010. Large-Scale Solar Thermal Power. Technologies, Costs and Development. Wiley-
VCH: Weinheim

36
Wagar W.R., Zamfirescu C., Dincer I. 2011. Thermodynamic Analysis of Solar Energy Use for Reforming Fuels
to Hydrogen. International Journal of Hydrogen Energy 36:7002-7011.
WagarW.R., Zamfirescu C., Dincer I. 2011. Thermodynamic Analysis of Solar Energy Use for Reforming Fuels
to Hydrogen. International Journal of Hydrogen Energy 36:7002-7011.
Yan H., Yang J., Ma G., Wu G., Zong X., Lei Z., Shi J. 2009.Visible-light-driven hydrogen production with
extremely high quantum efficiency on Pt–PdS/CdS photocatalyst. Journal of Catalysis 266:165–168.
Zamfirescu C., Naterer G.F., Dincer I. 2011a. Solar light-based hydrogen production systems. In: Encyclopedia
of Energy Engineering and Technology. Anwar S. eds. Taylor Francis.
Zamfirescu C., Dincer I., Naterer G.F. 2011b. Analysis of a Photochemical Water Splitting Reactor with
Supramolecular Catalysts and a Proton Exchange Membrane. International Journal of Hydrogen Energy
36:11273-11281.
Zamfirescu C., Dincer I., Stern M., Wagar W.R. 2012a. Exergetic, Environmental and Economic Analyses of
Small-capacity Concentrated Solar-driven Heat Engines for Power and Heat Cogeneration. International
Journal of Energy Research 36:397-408.
Zamfirescu C., Naterer G.F., Dincer I. 2012b. Photo-electro-chemical chlorination of cuprous chloride with
hydrochloric acid for hydrogen production. International Journal of Hydrogen Energy 37:9529-9536.
Zamfirescu C., Naterer G.F., Dincer I. 2012 Water Splitting with a Dual photo-electrochemical Cell and Hybrid
Catalysis for Enhanced Solar Energy Utilization. International Journal of Energy Research (early view, DOI:
10.1002/er.2910).
Zamfirescu C., Dincer I., Naterer G.F., Banica R. 2013. Quantum Efficiency Modeling and System Scaling-Up
Analysis of Water Splitting with Cd1-xZnxS solid-solution photocatalyst. Cemical Engineering Science 97:235-
255.

37
 EXERGY ANALYSIS OF A HYBRID SOLAR-WIND-BIOMASS SYSTEM WITH THERMAL AND
ELECTRICAL ENERGY STORAGE FOR A COMMUNITY
1 1 1
K.Hacatoglu , I.Dincer , and M. A. Rosen
1
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology
2000 Simcoe Street North, Oshawa, Ontario, Canada L1H 7K4
Email: Kevork.Hacatoglu@uoit.ca, Ibrahim.Dincer@uoit.ca, Marc.Rosen@uoit.ca

ABSTRACT
Sustainable community energy systems often require integration of renewable energy systems with energy
storages. Here, a hybrid solar-wind-biomass system with thermal and electrical storages is proposed to supply
continuous and reliable energy to a typical Ontario community. Exergy analysis is applied, and indicates that an
1800 m2 parabolic trough collector with a wind turbine (16 m rotor radius) and anaerobic digestion can meet the
energy needs of a community of 50 households when thermal and electrical storages are utilized, and that the
steam turbine and space heating components offer the best potential for improving the overall efficiency.

INTRODUCTION
Although energy fuels the development and prosperity of billions of people around the world, concerns exist
regarding access to adequate, affordable, and reliable energy supplies and accelerated rates of climate change
caused by greenhouse gas emissions from fossil fuel combustion. The twin threats of energy security and climate
change are driving interest in renewable energy sources such as solar, wind, biomass, and geothermal.

A key technical challenge associated with renewable energy is intermittency. The availability of solar and wind
energy fluctuates but integration with storage technologies diminishes variability and improves the reliability of
supply. Another method of improving the reliability of renewable energy systems is through hybridization.
Combining several technologies into a hybrid system improves the chances that at least one of the energy sources
will be available at a given time.

Unlike photovoltaic-based technologies, which directly convert sunlight to electricity, solar thermal approaches
heat a working fluid that drives a power conversion unit similar to conventional power plants. A number of
commercial solar thermal power plants were built from 1984 to 1991 in the Mojave Desert in California (Kreith and
Kreider, 2011). The most commercialized solar thermal technology is parabolic trough collectors that concentrate
direct radiation 30-80 times (Romero-Alvarez and Zarza, 2007). Other promising technologies include central
receiver or dish/Stirling systems.

Thermal storage technologies such as two- and single-tank systems can improve the reliability of solar thermal
systems. In the two-tank system, fluid is divided into hot and cold tanks. The direct system uses the same fluid in
the solar circuit and storage whereas the two-tank indirect system uses different fluids. Single-tank systems have
a temperature gradient from hot to cold that can cool a hot fluid (charging) or heat a cold fluid (discharging).

Efficient use of energy resources requires a clear understanding of not just quantity but also quality of energy.
Traditional energy analysis based on the first law of thermodynamics should be augmented by a second law-
based exergy approach that also considers quality (Rosen and Dincer, 1997). Linkages between exergy and
increasingly important areas such as the environment and sustainable development have expanded the potential
benefits of exergy analysis (Dincer and Rosen, 2012).

Sustainable communities of the future will likely need to be powered by adequate and reliable flows of renewable
energy. The objective of this research is to conduct an exergy analysis and efficiency assessment of a hybrid
renewable energy system using actual energy-use data and environment conditions, and to investigate the
sources of exergy destruction and the dependence of the system on its parameters. This system utilizes solar
thermal, wind, and bioenergy technologies, along with thermal and electrical energy storages.

SYSTEM DESCRIPTION AND COMPONENT EXERGY DESTRUCTION RATES

Many components comprise the considered solar-wind-biomass system (Figure 1). Solar and wind energy are
converted to heat, cold, and electricity to meet the energy demands of the community. Municipal solid waste
(MSW) from local households is converted to a methane-rich biogas through anaerobic digestion. Variability is
managed using thermal and electrical storage technologies. The system will be either in charging or discharging
modes depending on the balance between demand and supply. Excess supply from the solar thermal system will
charge thermal storage tanks while surplus electricity will charge electrical storages.

38
  x ) for each component.
Figure 1. Schematic of the hybrid solar-wind-biomass community energy system with exergy destruction rates ( E
D

Solar thermal subsystem

The solar thermal component is responsible for meeting space heating, cooling, and some of the electricity
demand (Figure 2). Integrating the solar thermal system with thermal storage ensures reliability and continued
production during periods of solar unavailability.

Figure 2. Solar and thermal storage subsystems of the overall community energy system with exergy destruction rates.

Direct solar radiation is reflected by parabolic trough collectors to a heat transfer fluid (HTF) circulating through a
receiver. The HTF is modeled as Therminol, which is also used in the aforementioned solar plants in the Mojave
Desert. The HTF in a solar thermal system should be kept in a liquid state and pressurization is required for
temperatures above Therminol’s normal boiling point of 257°C. Concentration of radiation allows for the
temperature of the HTF to increase up to approximately 350°C. The hot HTF can then be used in four ways: 1)
space heating, 2) absorption refrigeration, 3) electricity generation, or 4) storage.

The solar thermal system is the primary source of space heating and cooling for the community. Hot HTF from the
receiver or the hot storage tank passes through heat exchangers to directly provide domestic space heating in the
winter or to vaporize and separate the refrigerant (ammonia or NH3) from the transport medium (water or H2O) in
the generator of an absorption refrigeration cycle in the summer. The NH3-rich vapour then condenses and flashes
via a thermal expansion valve and then evaporates due to the heat collected from households in the community.
The weak solution (low in NH3) leaving the generator preheats the incoming strong solution and flashes via a
thermal expansion valve. The NH3-rich vapour and weak solution recombine to form a strong solution that is
pumped and preheated in a heat exchanger.

39
A power generating Rankine cycle is only activated when the supply of electricity is insufficient. Hot HTF enters
the subsystem through a steam-generating heat exchanger. The superheated steam drives a turbine that produces
mechanical work followed by electricity via an electric generator. The low-pressure steam is condensed and
pumped back into the steam generator.

A two-tank direct thermal storage system manages the effects of fluctuations in solar availability. When solar
thermal energy is plentiful, a portion of the circulating HTF is diverted to charge a high-temperature storage tank
maintained at 350°C. Charging the hot tank simultaneous discharges the low-temperature tank (250°C). During
periods of high demand or low solar and wind availability, hot HTF is discharged from storage, which in turn
charges the cold tank.

Wind and anaerobic digestion subsystem

The wind and anaerobic digestion components of the hybrid system are responsible for meeting most of the
electricity demand of the community (Figure 3). Integrating these components with electrical storage improves
reliability by providing a back-up during periods of low wind speeds.

Figure 3. Wind, anaerobic digestion, and electrical energy storage subsystems of the overall community energy system with exergy
destruction rates.

The kinetic energy in wind is converted to rotational mechanical work by a wind turbine. We assume a rotor radius
of 16 m and cut-in and cut-out speeds of 5 and 23 m/s, respectively. Surplus power production charges a battery
that can be discharged during periods of high demand or low wind availability.

Waste biomass is consistently available throughout the year locally as the average person in Ontario produces
approximately 1 kg day-1 of MSW (Jackson, 1999). Anaerobic digestion of MSW generates a methane-rich biogas
and digestate that can be converted to a marketable high-quality humus material via aerobic composting
(Kayhanian et al., 2007). The biogas can generate power in an open Brayton cycle with air/exhaust gas as the
working fluid. Combustion of biogas generates hot exhaust gases that expand in a gas turbine producing
mechanical work. The temperature of the gas at the outlet of the turbine is high enough to preheat fresh incoming
compressed air in a regenerator.

ANALYSIS
A thermodynamic model of the hybrid system was developed using the Engineering Equation Solver (EES)
software. The model was run for 365 days with five different input variables: 1) electricity demand per household
(excluding air conditioning), 2) air conditioning demand per household, 3) direct solar irradiance, 4) wind speed,
and 5) ambient temperature.

Electricity demand is separated from air conditioning because the proposed model supplies cooling through an
absorption refrigeration cycle driven by thermal energy. We can calculate the required heat removal rate from air
conditioning data (Saldanha and Beausoleil-Morrison, 2012) using an average coefficient of performance (COP) of
4 for central air conditioners. The daily electricity needs of a typical household in Ontario are illustrated in Figure 4.

40
Figure 4. Daily power needs over one year for a typical household in Ontario, excluding air conditioning (A/C) and only for air conditioning
(day “0” corresponds to August 1, 2009).

Solar and wind are intermittent energy sources that exhibit variable availability over the year. The direct solar
irradiance available to Toronto, Ontario, Canada, can be approximated by a semi-empirical model presented in

(Kreith and Kreider, 2011). The direct solar irradiance ( Q Solar ) on a tilted plane is a function of the extraterrestrial
 ), clearness number ( C ), local extinction coefficient ( k ), elevation angle (  ), and angle
solar irradiance ( QETR n C
of incidence ( i ).

Q Solar  Q ETR Cn exp( k c sin  ) cos i (1)

The angle of incidence is a function of the elevation angle, solar azimuth angle ( a s ), collector azimuth angle ( a w ),
and collector tilt angle (  ).

cos i  cos  cosas  aw sin   sin  cos  (2)

Elevation and solar azimuth angles can be determined from the day of the year, latitude, and longitude of the
location. The parameters used to calculate direct solar irradiance and the average monthly ambient temperature
( T0 ) in Toronto can be found in Table 1.

The wind speed ( WS ) profile of an average site can be modelled based on a Weibull probability density function
with a shape parameter ( k ) ranging between 1.5 and 2.5 (dimensionless) and a scale parameter ( c ) between 5
and 10 m s-1 (Zini and Tartarini, 2010).

k  WS 
k 1
  WS  k 
f WS     exp     (3)
c c    c  

Table 1. Direct solar irradiance parameters kC (dimensionless), Cn (dimensionless),  (degrees), and aw (degrees) and T0 (°C) for
Toronto, Ontario, Canada.
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
kC 0.142 0.144 0.156 0.180 0.196 0.205 0.207 0.201 0.177 0.160 0.149 0.142
Cn 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85
 70 60 50 30 20 20 20 30 40 60 70 70
aw 0 0 0 0 0 0 0 0 0 0 0 0
T0 -6 -5 0 6 12 17 20 19 15 9 3 -3

Solar irradiance and wind speed profiles for Toronto (Figure 5) were constructed based on the solar parameters in
Table 1 and shape and scale parameters of 2.0 and 8 m s-1, respectively.

41
Figure 5. Average daily direct solar irradiance and wind speeds over a year for Toronto, Ontario, Canada (day “0” corresponds to August 1,
2009).

The thermal exergy rate ( E  x ) of a thermal energy flow ( Q ) at a temperature

Q
T is usually defined as the heat
flow multiplied by the Carnot factor:

 T 
E x Q  Q 1  0  (4)
 T 

x
An alternative method of calculating the exergy rate of solar radiation ( E
Q
Solar ) is by multiplying the product of

Q Solar and solar collector area ( AC ) by the Petela-Landsberg-Press factor, which is considered to be a more
accurate measure of the exergy of blackbody radiation (Badescu, 2008):

 4 T 1  T0  
3
  
Ex Solar  QSolar AC 1 
Q 0
   (5)
 3 TSun 3  TSun  
 
The kinetic energy of wind is directly proportional to the swept area of the turbine blades and to the cube of wind
speed:

R 2 WS 3
1
KEWind  (6)
2

where  is the density of air and R  )

is the rotor radius. The actual power generated by the wind turbine ( PWind

depends on the power coefficient ( C P ), mechanical efficiency (  Mech ), and the efficiency of the electric generator
( Gen ):

PWind  C P Mech Gen R 2 WS 

1 3
(7)
2

Most of the material in MSW cannot be converted to biogas. Only the organic fraction of MSW ( OFMSW ) can
be converted, exclusive of moisture content ( MC ). Furthermore, it is the biodegradable volatile solids ( BVS )
component of the volatile solids fraction ( VS ) of the dry organic material that can be converted to biogas. A biogas
 BOM ):
conversion factor ( BGCF ) represents the fraction of biodegradable organic material that forms biogas ( m

m  MSW  OFMSW  1  MC   VS  BVS  BGCF

 BOM  m (8)

42
Biodegradable organic material can be represented chemically by C 60H95O38N, which can be used to calculate the
amount of methane and carbon dioxide produced by anaerobic digestion:

C60H95O38N + 18H2O → 32CH4 + 28CO2 + NH3 (9)

The above equations combined with mass, energy, and entropy balances determine mass, heat, and work flow
rates through all the system components. An exergy analysis determines the exergy destruction within each
component and calculates the system exergy efficiency. The overall specific exergy ( ex ) of a substance is the
ph ch
sum of its physical ( ex ) and chemical ( ex ) specific exergies, while the specific physical exergy of a flowing
material is based on its specific enthalpy and specific entropy, and the reference environment conditions:

ex  ex ph  ex ch (10)
ex ph  h  h0  T0 s  s0  (11)

The chemical exergy of a substance needs to be considered when calculating exergy changes of chemical
reactions. The chemical exergy for various molecules is available in (Morris and Szargut, 1986). The exergy
reference environment of 25°C and 101.325 kPa presented in (Morris and Szargut, 1986) is used in the analysis.
System exergy efficiency (  ) is the ratio of total exergy outputs to total exergy inputs:

W Load  E x Evap
Q
 E x SH
Q
 m InHT ex HT  m InCT exCT
 (12)
E x Solar
Q
 E xWind  m Out
HT
ex HT  m Out
CT
ex CT  m MSW ex MSW

where W Load denotes the community electrical power demand, E xEvap

Q
the rate thermal exergy needs to be

removed from community households, E xSH

Q
the thermal exergy rate of space heating, m InHT the mass flow rate
into the hot tank, ex HT the specific physical exergy of high-temperature heat transfer fluid, m InCT the mass flow
rate into the cold tank, and exCT the specific physical exergy of low-temperature heat transfer fluid.
Values for several parameters are defined for the model (see Table 2):

Table 2. Hybrid solar-wind-biomass model parameters.

Parameter Value Units

Area of solar collector 1800 m2

Biodegradable volatile solids (BVS) content 0.75 -
Biogas conversion factor (BGCF) 0.90 -
Concentration ratio 30 -
Efficiency of electric generator 0.92 -
Liveable floor space 195 m2
Moisture content of MSW (MC) 0.20 -
Mechanical efficiency (wind turbine) 0.60 -
MSW production rate 1 kg person-1 day-1
Number of households 50 -
Number of people per household 4 -
Organic fraction of MSW (OFMSW) 0.78 -
Power coefficient (wind turbine) 0.45 -
Rotor radius 16 m
Space temperature of a household 20 °C
Temperature of cold tank 250 °C
Temperature of hot tank 350 °C
Volatile solids fraction of dry organic solids (VS) 0.835 -

RESULTS AND DISCUSSION

The solar-wind-biomass model was run for 365 days with input variables of household electricity demand, air
conditioning demand, direct solar irradiance, wind speed, and ambient temperature. The daily change in the
amount of HTF in the hot tank is a function of all of the above input variables and is illustrated in Figure 6.

The highest demand for hot HTF (411 kg) occurs on day 299 (May 26, 2010). The electrical demand per
household that day is also very high (3.07 kW) as is the demand for air conditioning (0.93 kW or 3.72 kW of heat
removal based on an average COP of 4). The hot tank is charged with the most HTF (339 kg) on day 294 (May 21,
2010), when demand per household is only 0.85 kW with no air conditioning and high wind speeds (12 m/s).

43
However, the size of the storage tank is based on the maximum HTF accumulation (positive or negative) (see
Figure 7).

On day 25 (August 25, 2009) the hot tank has accumulated -4,320 kg of hot HTF whereas on day 296 (May 23,
2010) the accumulation is 6,080 kg. At year end, the net accumulation is greater than zero (1,108 kg). Conversely,
the cold tank accumulates -6,080 kg by day 296. To ensure a reliable supply of heat, cold, and electricity, each
tank requires a capacity of at least 6,080 kg or 7.5 m3 (based on a Therminol specific volume of 0.001232 m3 kg-1).

Figure 6. Change in the mass of HTF in the high-temperature storage tank over time (day “0” corresponds to August 1, 2009).

Figure 7. Overall accumulation of HTF in each storage tank over time (day “0” corresponds to August 1, 2009).

Component exergy destruction rates are shown in Figures 1-3, and the total exergy destruction rate of the system
over one year is found to be 192,000 kW. Of the various subsystems shown in Figure 1, the solar thermal system
has the most exergy destruction (70,000 kW or 36% of the total) followed by the wind turbine (50,900 kW or 26%).
Of the various individual components, the steam turbine exhibited a high destruction rate (17,300 kW or 9.0%).
Multi-stage expansion with reheating could significantly reduce exergy destruction and improve the overall
efficiency of the system. Another large source of exergy destruction is space heating (14,400 kW or 7.5%), which
requires low-grade thermal energy. The exergy efficiency of the system could be improved by using HTF from the
cold tank instead of the hot tank to provide space heating.

The accumulation of electrical energy in the battery occurs rapidly and builds until year end (Figure 8). This
becomes a significant source of exergy destruction as battery charging or discharging is approximately 80%
efficient (Zini and Tartarini, 2010). The battery reaches its maximum charge of 146,000 kWh on day 259 (April 16,
2010) and has a net charge of 133,000 kWh after one year. A combined energy and economic analysis could
reveal the ideal size of the battery to ensure continuous and reliable operation while minimizing cost.

44
Figure 8. Overall accumulation of electrical energy in the battery over time (day “0” corresponds to August 1, 2009)

A parametric study was conducted to ascertain the impact on overall exergy efficiency of several key parameters:
1) ambient temperature (Figure 9), 2) collector area (Figure 10), and 3) hot tank temperature (Figure 11). The
parametric study is performed for input variable values of 1.39 kW, 0.29 kW, 288 W m-2, and 19°C for household
electrical power demand, air conditioning demand, direct irradiance, and ambient temperature, respectively.

Figure 9. Effect of ambient temperature on system exergy efficiency at various wind speeds.

The effect of increasing the ambient temperature is a decrease in the exergy efficiency of the system for all wind
speeds, although the descent is steepest at 8 m/s. The thermal exergy of heat flows and the physical exergy of a
substance both decrease as ambient temperature increases.

Increasing the solar collector area gradually pushes the system towards a permanent thermal storage charging
state depending on the wind speed (faster wind speeds enter a charging state at a lower collector area). The
exergy efficiency increases because the system is charging the hot and discharging the cold storage tank.

Figure 10. Effect of solar collector area on system exergy efficiency at various wind speeds.

45
Figure 11. Effect of temperature of hot storage tank on system exergy efficiency at various wind speeds.

Increasing the upper limit of the hot HTF is observed in Figure 11 to increase exergy efficiency. As the temperature
of the HTF rises, the temperature of the working fluid of the Rankine cycle also rises, which raises the electrical
power that can be generated. However, beyond a certain temperature, which is dependent on wind speed, the rate
of heat loss from receiver tubes and the exergy destruction in components operating at lower temperatures (e.g.
space heater and the generator) exceeds the gains from higher Rankine cycle efficiencies.

CONCLUSIONS
The hybrid solar-wind-biomass system is able to meet the energy needs of a typical community in Ontario of 50
households. Approximately 1,800 m2 of parabolic trough collectors, a wind turbine with a rotor radius of 16 m, and
anaerobic digestion of MSW generates a net amount of hot HTF after 365 days. The size of each tank should be
at least 7.5 m3 to accommodate solar and wind fluctuations and balance energy demand with supply. The exergy
analysis revealed that there are significant efficiency improvements to be made especially with respect to the
steam turbine and space heating, which can both be enhanced through better design.

ACKNOWLEDGEMENTS
The authors gratefully acknowledge the support provided by the Natural Sciences and Engineering Research
Council of Canada.

NOMENCLATURE

AC Solar collector area, m2

as Solar azimuth angle, degrees
aw Collector azimuth angle, degrees
c Scale parameter, m s-1
Cn Clearness number, dimensionless
CP Power coefficient, dimensionless
ex Specific exergy, kJ kg-1
E x Exergy rate, kW
h Specific enthalpy, kJ kg-1
i Incidence angle, degrees
k Shape parameter, dimensionless
kC Local extinction coefficient, dimensionless
KE Kinetic energy, kW
m Mass flow rate, kW
Q Thermal energy rate, kW
R Rotor radius, m
s Specific entropy, kJ kg-1 K-1
T Temperature, K
W Work rate, kW

46
WS Wind speed, m s-1

Greek Letters
 Elevation angle, degrees
 Collector tilt angle, degrees
 Exergy efficiency, %
 Energy efficiency, %
 Density, kg m-3

Subscripts
0 Ambient conditions
BOM Biodegradable organic matter
CT Cold tank
ETR Extraterrestrial radiation
Evap Evaporator
Gen Generator
HT Hot tank
Mech Mechanical
MSW Municipal solid waste
SH Space heating

Superscripts
ch Chemical
D Destruction
ph Physical
Q Thermal

REFERENCES

Badescu,V. 2008. Exact and approximate statistical approaches for the exergy of blackbody radiation Central
European Journal of Physics 6:344-350.
Dincer, I. and M.A. Rosen. 2012. Exergy: Energy, Environment and Sustainable Development 2nd Edition. Boston:
Elsevier.
Jackson, J. 1999. Resources – not garbage. The Environmental Agenda for Ontario Project.
Kayhanian, M., G. Tchobanoglous, and R.C. Brown. 2007. Biomass conversion processes for energy recovery. In
Handbook of Energy Efficiency and Renewable Energy, edited by F. Kreith and D.Y. Goswami. Boca Raton:
CRC Press.
Kreith, F. and J.F. Kreider. 2011. Principles of Sustainable Energy. Boca Raton: CRC Press.
Morris, D.R. and J. Szargut. 1986. Standard chemical exergy of some elements and compounds on the planet
Earth. Energy 11:733-755.
Romero-Alvarez, M. and E. Zarza. 2007. Concentrating solar thermal power. In Handbook of Energy Efficiency and
Renewable Energy, edited by F. Kreith and D.Y. Goswami. Boca Raton: CRC Press.
Rosen, M.A. and I. Dincer. 1997. On exergy and environmental impact. International Journal of Energy Resources
21:643-654.
Saldanha, N. and I. Beausoleil-Morrison. 2012. Measured end-use electric load profiles for 12 Canadian houses at
high temporal resolution. Energy and Buildings 49: 519-530.
Zini, G. and P. Tartarini. 2010. Wind-hydrogen energy stand-alone system with carbon storage: modeling and
simulation. Renewable Energy 35:2461-2467.

47
 ENERGY AND EXERGY ANALYSES OF AN INTEGRATED SOLAR ENERGY-BASED
HYDROGEN PRODUCTION AND LIQUEFACTION SYSTEM
1 1
T. A. H. Ratlamwala , I. Dincer
1
University of Ontario Institute of Technology
1
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology,
2000 Simcoe Street North, Oshawa, Ontario L1H 7K4, Canada.
Emails:tahir.ratlamwala@uoit.ca, ibrahim.dincer@uoit.ca

ABSTRACT
The present paper concernsstudies on energy and exergy analyses of a new integrated system with heliostat field,
Cu-Cl cycle, Isobutane cycle and Linde-Hampson system. The present system is capable of producing liquefied
hydrogen for easier storage than hydrogen gas. A parametric study is conducted to investigate the effects of
variation in solar light intensity, ambient temperature and flow rate of makeup water required by the Cu-Cl cycle on
hydrogen production rate, hydrogen liquefaction rate, and overall energy and exergy efficiencies. The results show
that an increase in solar light intensity has positive effect on hydrogen production rate and hydrogen liquefaction
rate as they increase from 205 L/s to 492.5 L/s, and 43 L/s to 103 L/s, respectively. The overall energy and exergy
efficiencies are observed to be increasing from 4.1% to 7.3%, and 4.9% to 8.2%, respectively with increase in
solar light intensity from 600W/m2 to 750 W/m2. The rise in ambient temperature from 290K to 330 K affects the
performance of the system in the positive manner. The increase in supplied rate of makeup water to the Cu-Cl
cycle from 0.05 L/s to 0.15 L/s results in increase in overall exergy efficiency of the integrated system from 8.2% to
9.6%.

INTRODUCTION
Increased emissions of harmful greenhouse gasses due to extensive use of fossil fuels have made researchers and
organizations around the world re-think about potential solutions for a major energy carrier. These fossil fuels have
caused some critical local and global issues through increased CO2, NOx, and SO2 emissions. In this regard, the
world has started alarming ‘‘accelerated global warming’’, while the need for the clean and renewable energy has
become inevitable (Ratlamwala et al., 2012). One alternative to present energy carrier is hydrogen. Hydrogen offers
great advantages as a fuel and working fuel over the conventional ones. In the future, the role of hydrogen may
become more important, as some researchers suggest that the world’s energy systems may undergo a transition to
an era in which the main energy carrier will be hydrogen (Muradov and Veziroglu, 2008; Midilli and Dincer, 2009;
Ratlamwala et al., 2010; Dincer, 2006).At present the use of hydrogen as fuel is limited to the transportation due to
lack of proper infrastructure for supply or production of hydrogen. Dufour et al, (2009) mentioned that presently 96%
of the hydrogen is produced using the SMR technology. The other major technologies which are being currently used
or developed for production of hydrogen are(a) the water electrolysis, (b) the coal gasification and (c) the
thermochemical water splitting cycles (TWSCs). Among all the above mentioned technologies which are being
studied for hydrogen production; thermochemical water splitting cycles have an edge.
The Study conducted by Giaconia et al. (2011) concluded that TWSCs represent an appealing carbon-free
option for hydrogen production powered by alternative (carbon-free) energy sources. Aghahosseini et al., (2011)
explained that a thermochemical cycle is a process consisting of a closed loop of thermally driven chemical reactions,
where all intermediate compounds are recycled, while water is decomposed into the hydrogen and the oxygen.
Among the TWSCs, Cu-Cl cycle holds an edge because of its lower operating temperatures (around 530oC) which
results in reduced material and maintenance cost as mentioned by (Dincer and Balta, 2011; Orhan et al., 2009).
Another benefit associated with Cu-Cl cycle is that it has one of the highest energy efficiencies among other TWSCs
as studied by (Ratlamwala and Dincer, 2012; Lewis et al., 2010; Zamfirescu et al., 2010).
The heat required by TWSCs can be supplied by the use of advanced solar energy systems. The use of
solar energy systems will ensure that the overall hydrogen production process is environmentally benign as energy is
generated using renewable energy source and hydrogen is produced using TWSCs which are environmentally
benign systems, as they recycle all the chemicals within the process. The use of high temperature solar technologies
is governed by the temperature which they can generate. In a study, Kalogirou (2004) presented the working
temperature range of different solar thermal technologies which are parabolic trough collector (60 oC - 300 oC), solar
dish (100oC - 500oC), and heliostat field (150 oC - 2000 oC). Studies conducted by several researchers (Barigozzi et
al., 2012; Huang et al., 2012; Xu et al., 2011) concluded that using high temperature solar thermal collectors are
beneficial from both heat production and efficiency perspectives. Within the high temperature solar technologies, the
heliostat field system has an advantage of catering to wide temperature range which can be useful for operating the
integrated system which requires multiple operating temperatures.
In this paper, we study a new integrated system with heliostat field, Cu-Cl cycle, Isobutane cycle and Linde-
Hampson cycle. The system presented is capable of producing hydrogen and later liquefying it for better storage and
handling. A parametric study is carried out to investigate the effects of varying the operating conditions such as solar
light intensity, ambient temperature and flow rate of makeup water required by the Cu-Cl cycle on hydrogen
production rate, hydrogen liquefaction rate, and overall energy and exergy efficiencies.

48
SYSTEM DESCRIPTION
The integrated system presented in this paper is shown in Fig. 1. The system utilizes the solar energy to produce
hydrogen using the Cu-Cl cycle (Fig. 2a) and later liquefy it using the Linde-Hampson cycle (LH). The solar energy is
harnessed with the help of Heliostat field system. The molten salt passes through the receiver of the Heliostat system
to carry energy extracted from the solar light in the form of heat. The high temperature molten salt then enters the
high temperature heat exchanger where it releases heat to the water coming from the Cu-Cl cycle and isobutane
liquid coming from the binary power plant. The heat carried by the water coming from the Cu-Cl cycle is supplied to
the heat exchanger network of the Cu-Cl cycle as shown in Fig. 2b. The water stream leaving the heat exchanger
network of the Cu-Cl cycle is later supplied to the Kalina cycle to recover excess heat in form of power. The power
produced is supplied to the Cu-Cl cycle in order to decrease the energy demand from the grid. The high temperature
isobutane steam leaving the high temperature heat exchanger is supplied to the turbine of the isobutane cycle (Fig.
2c) in order to produce power. The power produced by the isobutane cycle is supplied to the Linde-Hampson cycle
(Fig. 2d) for liquefying hydrogen produced by the Cu-Cl cycle. In the LH system,makeup gas is mixed with the
hydrogen gas coming from the Cu-Cl cycle, and the mixture at state 1 is compressed to state 2 from 101 kPa to 10
MPa. Heat is rejected from compressed gas to a coolant. The high pressure gas is cooled to state 3 in a regenerative
counter flow heat exchanger (I) by the uncondensed gas, and is cooled further by flowing through two nitrogen baths
(II and IV) and two regenerative heat exchanger (III and V) before being throttle to state 8, where it is a saturate
liquid-vapor mixture. The liquid is collected as the desired output for further usage and the vapor is routed through the
bottom half of the cycle.The detailed system description of the Cu-Cl cycle can be found elsewhere (Ratlamwala and
Dincer, 2012).

Fig. 1. Schematic of an integrated system.

49
Fig. 2a. Schematic of the Cu-Cl cycle.

Fig. 2b. Schematic of the heat exchanger network for the Cu-Cl cycle.

50
Fig. 2c. Schematic of theisobutane cycle.
LN2 GN2 LN2 GN2
Makeup
gas Q I II III IV
1 2 3 4 5 6
m

10 9
W 11
V
m  m f

g 8

Fig. 2d. Schematic of the Linde-Hampson cycle.

f
m f Liquid
ENERGY AND EXERGY ANALYSES
This section provides energy and exergy analyses of the integrated system studied in this paper.

Heliostat system
The heliostat field energy model is taken from the study conducted by Xu et al. (2011). The rate of heat received
by the solar light intensity is calculated as

(1)

The total rate of heat received by the receiver is defined as

(2)
where

51
The rate of heat absorbed by the molten salt passing through the receiver is found using

(3)

The exergy rate carried by the solar light is calculated as

(4)

Cu-Cl cycle
The specific enthalpy at any given state in the Cu-Cl cycle is calculated as

(5)

The specific entropy at any given state in the Cu-Cl cycle is calculated as

(6)

The exergy rate at any given state in the Cu-Cl cycle is

(7)

The thermal exergy rate in each heat exchanger is defined as

(8)

The heat supplied to the generator of the Kalina cycle is defined as

(9)

The net power output that can be obtained from the Kalina cycle is

where

Isobutane cycle
The net power that can be obtained from the binary isobutane plant is expressed as

(10)

where

Linde-Hampson cycle
The ideal specific work required to compress the hydrogen from 101 kPa to 10 MPa is given as

(11)

The actual specific work input to the liquefaction cycle per unit mass of the hydrogen is stated as

52
 (12)

The total specific power required by the LH cycle is calculate as

(13)
where

(14)
where represents the fraction of liquefied hydrogen which is taken out of the cycle.
The amount of hydrogen liquefied is then calculated using

(15)

Efficiencies
The overall energy and exergy efficiencies are defined as

(16)

where .

(17)

where

RESULTS AND DISCUSSION

This paper presents an integrated system capable of producing and liquefying hydrogen using solar energy
source. The integrated system studied consists of Heliostat field system, Cu-Cl cycle, Kalina cycle, Isobutane
cycle, and Linde-Hampson cycle. The heliostat field system modelled introduced by Xu et al. (2011) is used in
present study. The energy efficiency of the heliostat field system obtained by Xu et al. (2011) was 75% as
compared to 80% in present study. The efficiency of the Cu-Cl cycle in this study is found to be 52% as compared
to 55% reported by Lewis et al. (2010).

Effect of solar light intensity

The continuous variation in solar light intensity throughout the day makes it very important to study the effect of
variation in solar light intensity on the performance of the system. The effect of rise in solar light intensity on the
hydrogen production and liquefaction rate is studied and is shown in Fig. 3. The hydrogen production rate and
hydrogen liquefaction rate is observed to be increasing from 205.3 L/s to 492.5 L/s, and 42.84 L/s to 102.8 L/s,
respectively with rise in solar light intensity from 600 W/m2 to 750 W/m2. Such behavior is observed because
increase in solar light intensity results in higher rate of heat production by the Heliostat field system. As the rate of
heat generated by the Heliostat system increases the hydrogen production capability of the Cu-Cl cycle and the
power generating capability of the isobutane system increases. The Cu-Cl cycle is mostly heat dependent and with
increase in the rate of heat supplied; the hydrogen production capacity of the Cu-Cl cycle increases for the
constant operating temperatures. The increase in the rate of heat supplied to the isobutane system results in
higher temperature of isobutane fluid entering the turbine. As the temperature of the fluid entering the turbine
increase, the power produced by the turbine increases. The increase in hydrogen liquefaction is directly related to
the increase in power supplied to the LH cycleand as a result of increase in power produced by the isobutane
system due to rise in solar light intensity, the hydrogen liquefaction rate increases. The overall energy and exergy
efficiencies are found to be increasing from 4.1% to 7.3% and 4.9% to 8.2%, respectively with rise in solar light
intensity as shown in Fig. 4. The increase in overall energy and exergy efficiencies are directly associated with
increase in the hydrogen liquefaction capability of the integrated system. These results show that the performance
of the integrated system studied in this paper enhances with rise in solar light intensity for the given operating
conditions.

53
Fig. 3. Effect of rise in solar light intensity on rate of hydrogen production Fig. 4. Effect of rise in solar light intensity on overall energy and exergy
and liquefaction. efficiencies.

Effect of makeup water supplied to the Cu-Cl cycle

The makeup water supplied to the Cu-Cl is an important parameter to study because the supplied makeup water
helps in achieving the desired concentration of different chemicals circulating in the Cu-Cl cycle. The hydrogen
production and liquefaction rate is observed to be increasing from 492.1 L/s to 503 L/s and 102.7 L/s to 105 L/s,
respectively with increase in water flow rate at state cc[1] from 0.05 L/s to 0.15 L/s as shown in Fig. 5. The
increase in makeup water helps achieving the desired concentration of aqueous HCl based on the chemical
balance. This aqueous HCl then enters the electrolyzer where it reacts with Cu-Cl to produce hydrogen and CuCl2.
As the availability of the hydrogen ions in the HCl solution entering the electrolyzer increases with increase in the
makeup water supplied, the hydrogen production rate also increases. The overall energy and exergy efficiencies of
the integrated system is found to be increasing from 7.2% to 7.3% and 8.2% to 9.6%, respectively with increase in
supplied makeup water rate at state cc[1] as shown in Fig. 6. It is noticed that overall energy efficiency hardly
changes with increase in makeup water flow rate but the overall exergy efficiency increases by almost 1.5%. Such
behavior is noticed because energy analysis doesn’t take in consideration the gains or losses occurring in the
system due to the interactions between the chemicals whereas, exergy analysis does take in consideration all the
gains or losses happening due to the interactions of the chemical compounds. This difference between energy
analysis and exergy analysis makes it essential to study system from exergy perspectives as it provides with better
picture with regards to real life operation of the system in comparison to the energy analysis.

Fig. 5. Effect of increase in makeup water flow rate on rate of hydrogen Fig. 6. Effect of increase in makeup water flow rate on overall energy and
production and liquefaction. exergy efficiencies.

Effect of ambient temperature

The ambient temperature plays a major role in performance of any system which interacts with the environment
due to the temperature difference between the system and the surroundings. The temperature difference between
the system and the surrounding can either lead to better performance of the system or degrading performance of
the system based on the system being used. For the systems which require high temperatures to achieve its
output, the increase in ambient temperature is beneficial as it results in lower heat loss from the system to the
environment. The hydrogen production and liquefaction rate is seen to be increasing from 473.3 L/s to 571.8 L/s,
and 98.7 L/s to 119.3 L/s, respectively with rise in ambient temperature from 290 K to 330 K as shown in Fig. 7.
The integrated system studied in this paper is highly dependent on the temperature for the production of hydrogen
and power. The rise in ambient temperature helps facilitate the performance of the system because of the lower
temperature differences between the system boundary and the environment. This decrease in the temperature

54
difference leads to lower losses from the system to the surrounding in the form of heat. The effect of rise in
ambient temperature on overall energy and exergy efficiencies is displayed in Fig. 8. The overall energy and
exergy efficiencies are observed to be increasing from 7.2% to 7.5% and 8.1% to 8.6%, respectively with rise in
ambient temperature. This increase in energy and exergy efficiencies is directly related to the higher rate of
hydrogen liquefaction and lower heat loses due to the lower temperature difference between the system boundary
and surrounding with rising ambient temperature.

Fig. 5. Effect of rise in ambient temperature on rate of hydrogen production Fig. 6. Effect of rise in ambient temperature on overall energy and
and liquefaction. exergy efficiencies.

CONCLUSIONS
In this paper, we have presented an integrated system consisting of heliostat field system, Cu-Cl cycle, Kalina
cycle, Isobutane cycle, and Linde-Hampson cycle for hydrogen production and liquefaction. Parametric studies are
carried out to see the effect of variation in solar light intensity, makeup water flow rate and ambient temperature on
hydrogen production rate, hydrogen liquefaction rate and overall energy and exergy efficiencies. The results
obtained show that the hydrogen production and liquefaction rate are found to be increasing from 205 L/s to 492.5
L/s, and 43 L/s to 103 L/s, respectively with increase in solar light intensityfrom 600W/m2 to 750 W/m2. The overall
energy and exergy efficiencies are observed to be increasing from 4.1% to 7.3%, and 4.9% to 8.2%, respectively
with increase in solar light intensity. The increase in supplied rate of makeup water to the Cu-Cl cycle from 0.05
L/s to 0.15 L/s results in increase in overall exergy efficiency of the integrated system from 8.2% to 9.6%. Also, the
overall exergy efficiency increases from 8.1% to 8.6%, respectively, with rise in ambient temperature.

Acknowledgement
The authors acknowledge gratefully the financial support provided by the Ontario Research Excellence Fund and
the Turkish Academy of Sciences.

NOMENCLATURE
A Area
C Concentration ratio
d Diameter, m
Exergy destruction rate, kW
f Fraction
Fr View factor
h Specific enthalpy, kJ/kg; Heat transfer coefficient, W/m2 K
hw Hot water
HHV Higher heating value
I Solar light intensity, W/m2
Mass flow rate, kg/s
M Molecular weight, kg/mol
mf Mass fraction
P Pressure, kPa
Heat flow rate, kW
T Temperature, K
s Specific entropy, kJ/kg K
w Specific work, kJ/kg
Work rate, kW

55
Greek Letters
Efficiency
ε Receiver surface emissivity
Stefan-Boltzmann constant, W/m2 K4
ρ Density, kg/m3
∂ Thickness, m
λ Thermal conductivity, W/m K
Subscripts
abs Absorbed
avg Average
ch Chemical
cond Conduction
conv Convection
convr Conversion
comp Compressor
elec Electrolyzer
em Emissive
en Energy
ex Exergy
fc Forced convection
H Heliostat
H2 Hydrogen
HE Heat exchanger
i Inner
iso Isobutane
insi Inner side of receiver
insu Insulation
i ith state
k kth state
liq Liquid
m mth state
ms Molten salt
nc Natural convection
o Outer
ov Overall
p Pump
ph Physical
rec Receiver
ref Reflection
S Solar
surf Surface
sys System
th Thermal
turb Turbine
w Wall surface
0 Ambient state

REFERENCES
Aghahosseinin, S., I. Dincer and G. Naterer. 2011. Integrated gasification and Cu-Cl cycle for trigeneration of
hydrogen, steam and electricity. International Journal of Hydrogen Energy 36, 2845-2854.
Barigozzi, G., G. Bonetti., F. A. Perdichizzi and S. Ravelli. 2012. Thermal performance prediction of a solar hybrid
gas turbine. Solar Energy 86, 2116-2127.
Dincer, I. and M. T. Balta. 2011. Potential thermochemical and hybrid cycles for nuclear-based hydrogen
production. International Journal of Energy Research 35: 123-137.
Dincer, I. 2006. Environmental and sustainability aspects of hydrogen and fuel cell systems. International Journal
of Energy Research 31: 29-55.
Dufour, J., D. P. Serrano, J. L. Galvez, J. Moreno and C. Garcia. 2009. Life cycle assessment of processes for
hydrogen production: environmental feasibility and reduction of greenhouse gases emissions. International
Journal of Hydrogen Energy 34, 1370-1376.
Giaconia, A., S. Sau, C. Felici, P. Tarquini, G. Karaginnakis, C. Pagkoura, et al. 2011. Hydrogen production via
sulfur-based thermochemical cycles: Part 2: Performance evaluation of Fe2O3-based catalysts for the sulfuric
acid decomposition step. International Journal of Hydrogen Energy 36, 6496-6509.

56
Huang, W., P. Hu and Z. Chen. 2012. Performance simulation of a parabolic trough solar collector. Solar Energy
86, 746-755.
Kalogirou, S. A. 2004. Solar thermal collectors and applications. Progress in Energy and Combustion Science 30,
231-295.
Lewis, M. A., J. G. Masin and R. B. Vilim. 2010. Development of the Low Temperature Cu-Cl Thermochemical
Cycle. International Congress on Advances in Nuclear Power Plants.
Muradov, N. Z. and T. N. Veziroglu. 2008. ‘‘Green’’ path from fossil-based to hydrogen economy: An overview of
carbon-neutral technologies. International Journal of Hydrogen Energy 33: 6804-6839.
Midilli, A. and I. Dincer. 2009. Development of some exergetic parameters for PEM fuel cells for measuring
environmental impact and sustainability. International Journal of Hydrogen Energy 34: 3858-3872.
Orhan, M. F., I. Dincer, and M. A. Rosen. 2009. Energy and exergy analyses of the drying step of a copper-
chlorine thermochemical cycle for hydrogen production. International Journal of Exergy 6, 793-808.
Ratlamwala, T. A. H., A. H. El-Sinawi, M. A. Gadalla and A. Ahmad. 2012. Performance analysis of a new
designed PEM fuel cell. International Journal of Energy Research 36: 1121-1132.
Ratlamwala, T. A. H. and I. Dincer. 2012. Energy and exergy analyses of a Cu–Cl cycle based integrated system
for hydrogen production. Chemical Engineering Science 84:564-573.
Ratlamwala, T. A. H., M. A. Gadalla and I. Dincer. 2010. Energy and exergy analysis of an integrated fuel cell and
absorption cooling system. International Journal of Exergy 7: 731-754.
Xu, C., Z. Wang, X. Li and F. Sun. 2011. Energy and exergy analysis of solar power tower plants. Applied Thermal
Engineering 31, 3904-3913.
Zamfirescu, C., I. Dincer and G, Naterer. 2010. Thermophysical properties of copper compounds in copper–
chlorine thermochemical water splitting cycles. International Journal of Hydrogen Energy 35, 4839-4852.

57
 ENERGY AND EXERGY ANALYSES OF A ZERO EMISSION POWER PLANT FOR CO-
PRODUCTION OF ELECTRICITY AND METHANOL
C.Acar and I. Dincer
Faculty of Engineering and Applied Science
University of Ontario Institute of Technology
2000 Simcoe Street North, Oshawa, Ontario, L1H 7K4, Canada
Emails: Canan.Acar@uoit.ca, Ibrahim.Dincer@uoit.ca

ABSTRACT
In this study, we study and evaluate a zero-emission integrated system, as taken from the literature, for
co-production of electricity and methanol. The investigated integrated system has three sub-systems:
water electrolysis, Matiant power plant (oxy-fuel combustion of pure methane) and methanol production
unit. The system and its components are analyzed energetically and exergetically. The rates of exergy
destructions, relative irreversibilities and sustainability indexes of each sub-unit of each sub-system, as
well as the overall system are analyzed to identify the greatest exergy losses, and possible future
research directions. The total rate of exergy destruction of the overall system is calculated to be around
280 MW. The greatest rate of exergy destruction, therefore the greatest irreversibility, occurs within the
power plant unit (about 60% of the total rate of exergy destruction). The energy efficiencies of electrolysis,
power plant, and methanol synthesis unit are found to be 30%, 76%, and 41%, respectively. The exergy
efficiencies of electrolysis, power plant, and methanol synthesis unit are found to be 30%, 64%, and 41%,
respectively. Depending on the utilization of the heat rejected from the different units of each sub-system,
the overall system could have energy and exergy efficiencies up to 68% and 47%, respectively.

Keywords: Methanol, Carbon Capture, Co-generation, Electricity, energy, exergy, efficiency.

1. INTRODUCTION
During the twentieth and the beginning of the twenty-first century, world’s energy consumption has
increased steadily due to global rise in population and standards of living. The increasing trend in world’s
energy need is expected to continue in the future. As a result, a growth in energy generation capacity will
be needed (IEEE, 2012).One of the main sources of CO2 emissions is power generation (Davison, 2007).
Therefore, reduction of fossil fuel utilization by increasing efficiency and decreasing emissions by converting CO 2
into valuable products offer somepotential solutions to world’s current energy-related problems.

There have been several studies focusing on co-production of fossil fuels with CO2 utilization. A
comparative study of electricity and hydrogen production systems by CO 2 capture and storage has been
presented by (Damen et al, 2006). (Minutillo and Perna, 2010) studied a tri-reforming process in fossil
fired power plants to generate a synthesis gas for production of chemicals including methanol. Energy
analysis, economy and policy scenarios of electricity/methanol synthesis co-production systems’ are
taken into account in (Guang-jian et al, 2010; Chen et al, 2011)’s work. (Katayama and Tamaura, 2005;
Takeuchi et al, 2001) considered methanol synthesis via CO 2 recovery and renewable hydrogen from
water electrolysis unit byrenewable energy sources as a zero emission integrated system.(Mignard et al,
2003) proposed anintegrated system for co-production of electricity and methanol in small scale/partial
carbon capture in common fossil fuel power plants. The utilization of renewable hydrogen for methanol
production is investigated by Cifre and Badr (2007).

The present paper aims to study and evaluate a zero-emission integrated system for co-production of
electricity and methanol as originally proposed and studied by Soltanieh et al. (2012). The characteristics
and performance of electricity-methanol cogeneration system with water electrolysis, Matiant power cycle
and methanol synthesis unit is energetically and exergetically evaluated. Energy, exergy, and
sustainability analyses of every component of the systemas well asthe overall system are performed. The
performance of the system is assessed for various environmental parameters such as relative
irreversibility and sustainability index. In the end, recommendations for better performance in terms of
environmental and sustainability issues are provided.

58
2. SYSTEM DESCRIPTION

2.1 Overall System

The integrated system, as originally developed and published by Soltanieh et al. (2012), is presented in
Figure 1. It has three sub-systems: water electrolysis for H2 and O2 production, the matiant cycle power
plant to produce electricity via oxy-combustion, and methanol synthesis unit fed by the CO 2 from power
plant and renewable H2. In this study, all processes are assumed to be steady-state and steady-flow
(SSSF) where potential and kinetic energy changes are neglected. It is also assumed that no chemical
reactions occur except the water electrolysis modules, combustion chambers, and methanol reactor. Heat
loss and pressure drop in tubing, valves, and connections are ignored. All equipment except the heat
exchangers is assumed to be adiabatic. Compressors and turbines are assumed to operate at an
isentropic efficiency of 85% and the efficiency of heat exchangers is assumed to be 90%.

Fig. 1 Integrated system block diagram (modified from Soltanieh et al., 2012).

2.2 Water Electrolysis Unit

Figure 2 shows the water electrolysis unit. The electrolysis unit uses wind (renewable) energy to provide
the necessary energy for the electrolysis reaction is: (ΔH = 286 kJ/mol H2) (Perry
and Green, 2000). The technical criteria used for the unit is selected based on Soltanieh et al. (2012)’s
study. The water feed (stream 1) is assumed to undergo complete electrolysis. The product H2 (streams
2-10) is either stored or sent to the methanol synthesis unit. The product O 2 (streams 11-19) is either
stored or sent to the matiant power plant. The inlet molar flow rate of water to the electrolysis unit is 1.4
kmol/s at 80°C and 1 bar. 0.75 kmol/s of this amount comes from the power plant and methanol synthesis
units while 0.65 kmol/s is the fresh feed. H2 and O2 are desired to be stored at 140°C and 10 bar. O 2 is
sent to the power plant at 160°C and 40 bar. Then, H2 is sent to the methanol synthesis at 180°C and 50
bar.

Fig. 2 Process flow diagram of the water electrolysis unit (modified from Soltanieh et al., 2012).

59
2.3 Matiant-Cycle Power Plant

The Matiant cycle power plant is presented in Figure 3. In this sub-system, the fuel (pure methane) is
burned with pure oxygen from the electrolysis unit and undergoes complete combustion. The combustion
reaction is: (ΔH = -800 kJ/mol CH4) (Perry and Green, 2000). The technical
criteria used for the plant is selected based on Soltanieh et al. (2012)’s study. The products of two
combustion chambers (CO2 and H2O) are separated, the H2O is sent back to the electrolysis unit and
about 92% of the CO2 is used as a thermodynamic fluid within the Matiant cycle while the remaining part
is sent to the methanol synthesis unit. 70% of the O2 from the electrolysis unit is sent to the combustion
chamber H, and the remaining is sent to the combustion chamber K. The CH4 inlet temperature and
pressure are 30°C and 2 bar, respectively. The output temperature and pressure of water from the flash
drum are 80°C and 2 bar. CO2 is sent to the methanol synthesis unit at 180°C and 50 bar.

Fig. 3 Process flow diagram of the matiant cycle power plant (modified from Soltanieh et al., 2012).

2.4 Methanol Synthesis Unit

Figure 4 presents the methanol synthesis unit. The methanol reactor converts CO2 to CH3OH according
to the following reaction: (ΔH = -49.51 kJ/mol CH3OH). The technical criteria
used for the unit selected based on Soltanieh et al. (2012)’s study. 0.25 kmol/s CO2 is fed to the unit from
the power plant. The CO2 to CH3OH conversion ratio in the methanol reactor is 20%. Methanol and water
are leaving the distillation column at 1 bar and 40°C and 80°C, respectively.

Fig. 4 Process flow diagram of the methanol synthesis unit (modified from Soltanieh et al., 2012).

3. ANALYSIS

The operating conditions and production capacities of each sub-system as well as the overall system are
selected based on the literature work (Soltanieh et al., 2012).The general mass, energy, entropy and
exergy balance equations can be summarized as:

Mass Balance Equation (MBE): (1)

60
Energy Balance Equation(EBE): (2)

Entropy Balance Equation (EnBE): (3)

Exergy Balance Equation (ExBE): (4)

Specific exergy of a stream is comprised of two components, namely physical and chemical exergy, and
defined as follows (Dincer and Rosen, 2007; Kotas, 1995):

(5)

(6)

(7)

(8)

where T0 is the ambient temperature, R is the specific gas constant; h/h0 and s/s0 are the specific
enthalpy and entropy of a stream at a given state/at ambient state and xi is the mol fraction of component
i.Relative irreversibilities (RI) and sustainability indexes (SI) of each unit of each sub-system as well as
the sub-systems themselves and the overall system are calculated based on the following equations:

(9)

(10)

In order to study the system energetically and exergetically, each stream is defined by its state,
components, temperature and pressure. After the states were defined, the specific enthalpy, entropy,
exergy and energy and exergy flow rates of each stream is calculated based on the specified inlet data
given elsewhere (Soltanieh et al, 2012) mentioned in section 2 or via balance equations.Ambient state is
taken to be 20°C and 1 bar, respectively. The thermodynamic properties of H 2O, H2, O2, CH4, CO2, and
CH3OH are found using the Engineering Equation Solver (EES) software package.

4. RESULTS AND DISCUSSION

4.1 Water Electrolysis Unit

Overall, the electrolysis unit needs 1.4 kmol/s of water (in total 0.75 kmol/s of this water comes from the
matiant power cycle and methanol production units). After the complete electrolysis, 0.75 kmol/s of H2 is
sent to the methanol production unit while 0.65 kmol/s of H 2 is sent to the storage tank. Similarly, 0.35
kmol/s of O2 is sent to the matiant power cycle unit while 0.35 kmol/s of O2 is sent to the storage tank.

Table 1 summarizes the energy and exergy efficiencies, as well as the energy input/output (either work or
heat) of electrolysis modules and compressors and heat exchangers and the overall sub-system itself.
The black cells can be assumed to be 0 or “not applicable”.

Table 1 Summary of the major components and the overall electrolysis unit.

Description η Ψ Q (kW) W(kW)

Electrolysis Modules 51% 38% 0 7,514
Heat Exchangers 90% 64% 27,904
Compressors 85% 80% 0 38,652
Overall Electrolysis Unit 30% 30% 27,904 46,166

61
The electrolysis unit’s total work input is around 46 MW, this amount is required to drive the electrolysis
reaction and compress the products H2/O2. Heat exchangers’ load indicates the total amount of cooling
required following the compression processes. The system has an energy and exergy efficiency of 30%.
The overall sustainability index of the water electrolysis sub-system is 1.43. These energy and
exergyefficiencies are calculated by the following equations:

(11)

(12)

Compressors
26%
Electrolysis
Modules
49%
Heat
Exchangers
25%

Fig. 5 Relative irreversibilities of the major components of the electrolysis sub-system

Figure 5shows that 49% of the system irreversibility, therefore the exergy destruction, occurs in the
electrolysis modules. This means when the heat exchangers and compressors are combined together,
they have about the same exergy destruction rates with the electrolysis modules alone. When compared,
the total exergy destruction rate in compressors is very close to the one of heat exchangers in total.
However, on average, compressors are exergetically more efficient thanthe heat exchangers. This might
be due to the high temperature differences betweenthe inlet and outlet of heat exchangers.

Initially, the ambient state is taken to be 20°C and 1 bar, respectively. In order to find how the electrolysis
unit performs exergetically under different ambient conditions the system is run under various ambient
temperatures. The exergy efficiency and exergy destruction rate of the electrolysis modules,
compressors, and heat exchangers as well as the overall electrolysis unit at different ambient
temperatures are shown in Figure 6.

Figure 6 shows that, with increasing ambient temperature, a slight increase is observed in the exergy
destruction rate of heat exchangers and compressors. This increase is less than 6% when the ambient
temperature is increased from 2 to 28°C. However, there is a slight decrease in the exergy destruction
rate of electrolysis modules. This decrease is less than 2% when the ambient temperature is increased
from 2 to 28°C. The overall system’s exergy destruction rate increases with increasing ambient
temperature. This increase is less than 2% as the ambient temperature is increased from 2 to 28°C.

Furthermore, it can be seen from Figure 6 that an increase in the ambient temperature has a negligible
effect on the exergy efficiencies of the electrolysis modules and the compressors (1% change as the
ambient temperature increased from 2 to 28°C). However, the exergy efficiency of the heat exchangers
decreases with increasing ambient temperature (3%-decrease as the ambient temperature increased
from 2 to 28°C). And the exergy efficiency of the overall electrolysis unit increases with increasing
ambient temperature (2%-increase as the ambient temperature increased from 2 to 28°C). Taking the
small percentage of the change of rate of exergy destructions and the exergy efficiencies, it can be
concluded that the ambient temperature doesn’t have a dramatic effect on exergetic performance of the
system within the studied temperature interval.

62
 32 85% Electrolysis Exergy
Destruction Rate

27 75% Compressors Exergy

Rate of Exergy Destruction (MW)

Destruction Rate

HEX Exergy
22 65%
Destruction Rate

Exergy Efficiency
Overall Exergy
17 55% Destruction Rate

Electrolysis Exergy
12 45% Efficiency

Compressors Exergy
Efficiency
7 35%
HEX Exergy Efficiency

2 25%
Overall Exergy
0 5 10 15 20 25 30
Ambient Temperature (°C) Efficiency

Fig. 6The effect of ambient temperature on exergy destruction rates and exergy efficiencies of major components of the electrolysis unit.

4.2 Matiant-Cycle Power Plant

In power plant, 0.5 kmol/s of the O2from the electrolysis unit is used for complete combustion of 0.25
kmol/s of natural as (pure methane). As a result, 0.50 kmol/s of H2O is produced and sent back to the
electrolysis unit. About 0.25 kmol/s of the CO 2 is sent to the methanol synthesis unit while about 3 kmol/s
of CO2 is used as thermodynamic working fluid in the system.

Table 2 Summary of the major components and the overall matiant-cycle power plant

Description η Ψ Q (kW) W(kW)

Heat Exchangers 95% 78% 35,579 0

Expanders (Turbines) 85% 80% 0 132,588

Combustion Chambers 79% 69% 0 0

Compressors and Pumps 85% 75% 11,426 16,710

Flash Drum 95% 78% 0 0

Overall Unit 76% 64% 47,005 115,878

Table 2 summarizes the energy and exergy efficiencies, as well as the energy input/output (either work or
heat) of the major components and the overall system itself. The black cells can be assumed to be 0 or
“not applicable”. From Table 2, it can be seen that the matiant-cycle power plant has an overall energy
efficiency of 76% and an exergy efficiency of 64%. The sustainability index of the overall power plant is
around 2.76. The total heat removal (from the heat exchanger used before the flash drum and the
compressor stage) is about 47 MW, the net-work output (the work output of the turbines – the work
required by compressor and pumps) is around 115 MW. The total exergy destruction rate is about 168

63
MW. The energy and exergy efficiency of the matiant-cycle power plant unit is calculated based on the
following equations:

(13)

(14)

Compressors Flash Drum Heat

and Pumps 3% Exchangers Expanders
13% 11% (Turbines)
15%

Combustion
Chambers
58%

Fig. 7 Relative irreversibilities of the major components of the matiant-cycle power plant.

Figure 7 shows that the greatest irreversibilities occur at combustion chambers (combined together, 58%
of the exergy destruction happens at the combustion chambers). The temperature and pressure
difference in combustion chambers is significant; also the chemical exergy destruction has a considerable
effect on the irreversibility at the combustion chambers. Following the combustion chambers, turbines,
compressor and pumps, and heat exchangers have similar relative irreversibilities, between 11-15%.
Flash drum has a small effect on the system’s irreversibility. One of the reasons of this result might be
because almost no temperature/pressure change is observed between the inlet/outlet streams of the flash
drum.

From Figure 8, it can be seen that as ambient temperature is increased from 2-28 °C, the total exergy
destruction rate of the turbines increases by 2.7%. Combustion chambers’ total exergy destruction rate
also increases by 5.9%. Overall, the total exergy destruction rate of the power plant system increases by
1.6%. The different effect of the ambient temperature on exergy destruction rates of various major
components can be explained by the different temperature and pressures of inlet/outlet streams of each
component. Since each stream’s exergetic potential depends on how close/far they are to/from the
ambient (dead) state, some streams’ exergetic potential increases as the ambient temperature is
increases, while some of them decrease. A decrease in exergy efficiency with increasing ambient
temperature occurs in combustion chambers. Within the defined temperature range, the decrease in
exergy efficiency of combustion chambers is about 1%. The turbines’ exergy efficiency decreases by
about 0.4% as the temperature increased within the defined range. Overall, the power plant’s exergy
efficiency decreases by 2% within the defined ambient temperature range.

4.3 Methanol Synthesis Unit

The methanol synthesis unit uses the H2 from the water electrolysis unit and the CO2 from the power
plant. Overall, the methanol synthesis unit needs 0.25 kmol/s of CO 2 (from power plant unit) and 0.75
kmol/s H2 (from electrolysis unit). As a result, 0.25 kmol/s of H 2O is produced and sent back to the
electrolysis unit. The methanol synthesis unit produces 0.25 kmol/s of CH3OH.

64
 180 85%
170 Turbines Exergy
Destruction Rate
160
150 80%
Rate of Exergy Destruction (MW)

Combustion
140
Chambers Exergy

Exergy Efficiency
130 Destruction Rate
120 75% Overall Exergy
110 Destruction Rate
100
90 70% Turbines Exergy
80 Efficiency
70
60 65% Combustion
50 Chambers Exergy
Efficiency
40
Overall Exergy
30 60%
Efficiency
0 5 10 15 20 25 30
Ambient Temperature (°C)

Fig. 8The effect of ambient temperature on exergy destruction rates and exergy efficiencies of major components of the power plant

Table 3 summarizes the energy and exergy efficiencies, as well as the energy input/output (either work or
heat) ofthe major components and the overall system itself. The black cells can be assumed to be 0 or
“not applicable”.

Table 3 Summary of the major subunits and the overall methanol synthesis unit.

Description η Ψ Q (kW) W(kW)

Heat Exchangers 87% 78% 1,934

Methanol Reactor 68% 68%

Flash Drum 90% 82%

Compressor 85% 76% 2,284

Distillation Tower 89% 90%

Overall Synthesis Unit 41% 41% 1,934 2,284

Overall, the methanol synthesis unit requires about 2 MW of heat and about 2.2 MW of work input. At
given ambient conditions (20°C and 1 bar) the synthesis unit has an energy and exergy efficiency of
around 41%. As a result, the sustainability index of the methanol synthesis unit is 1.7. The energy and
exergy efficiency of methanol synthesis unit is calculated based on the following equations:

(15)

(16)

65
 Distillation
Compressor Tower Heat
14% 12% Exchangers
37%
Flash Drum
12%
Methanol
Reactor
25%

Figure 9 Relative irreversibilities of the major components of the methanol synthesis unit.

During the process, the rate of exergy destruction is about 80MW, and Figure 9 shows that 25% of the
irreversibility occurs at the methanol reactor. However, combined together, heat exchangers have a
significant effect on exergy destruction rate of the overall system, which means the greatest irreversibility
occurs at heat exchangers, combined. Following the heat exchangers, methanol reactor has the second
highest irreversibility. Compressors, distillation tower and the flash drum have similar relative
irreversibilities, around 12-14%. One of the reasons of the low irreversibility at flash drum and distillation
column is almost no temperature/pressure change is observed between the inlet/outlet streams of the
flash drum and distillation tower. The compressor’s pressure difference is not significant either. The
overall system requires additional heat, which is provided by the rejected heat and is not mentioned here.

Figure 10shows thatambient temperature has different effects on different major units of the methanol
synthesis unit. For instance, as ambient temperature is increased from 2-28 °C, the exergy destruction
rate of heat exchangers (combined) decreases by about 6%. The exergy destruction of the reactor
increases by 12%. Overall, the exergy destruction rate of the methanol synthesis unit increases by about
1% within the given ambient temperature range. The different effect of the ambient temperature on
exergy destruction rates of various major components can be explained by the different temperature and
pressures of inlet/outlet streams of each component. Since each stream’s exergetic potential depends on
how close/far they are to/from the ambient (dead) state, some streams’ exergetic potential increases as
the ambient temperature is increases, whereas some of them decrease. From Figure 10, it can be seen
that the decrease in exergy efficiency of heat exchangers is about 2%. The reactor’s exergy efficiency
decreases by about 0.11% as the temperature increased within the defined range. Overall, the methanol
synthesis unit’s exergy efficiency increases by about 0.2% within the defined ambient temperature range.

4.4 Overall System Analysis

Overall, the integrated system has 3 major units: the electrolysis, the matiant-cycle power plant, and
methanol synthesis units. In (Soltanieh et al, 2012)’s study wind energy is proposed as the required
power input required for water electrolysis. And the work required by compressors is provided by the
output of the power plant unit. The system has 10 compressors, , 9 heat exchangers that require heat
removal by a different process, 3 turbines, 2 heat exchangers that use different streams of the same unit
to provide heat/cold to one another, 2 combustion chambers, 2 flash drums, 2 storage tanks (1 for H 2 and
1 for O2), 1 electrolysis unit, 1 heat exchanger that require heat input from a different process, 1 methanol
reactor, 1 distillation tower and 1 pump. In this study, each unit is examined within its boundary limits, in
order to see where the greatest irreversibilities occur and how the ambient temperature affects the
system exergetically; the overall system is also investigated.

66
 80 80%
HEX Exergy
75
Destruction Rate
75%
70
65 70% Reactor Exergy
Rate of Exergy Destruction (MW)

60 Destruction Rate
55 65%

Exergy Efficiency
Overall Exergy
50
60% Destruction Rate
45
40 55% HEX Exergy
35 Efficiency
30 50%
Reactor Exergy
25
45% Efficiency
20
15 40% Overall Exergy
0 5 10 15 20 25 30 Efficiency
Ambient Temperature (°C)

Fig. 10The effect of ambient temperature on exergy destruction rates and exergy efficiencies of major components of the methanol synthesis
unit.

Methanol Electrolysis
29% 11%

Power Plant
60%

Figure 11 Relative irreversibilities of the electrolysis, power plant and methanol synthesis unit.

Figure 11 shows that the greatest irreversibility in the overall system occurs at the power plant, this is
mainly because of the effect of two combustion chambers in the power plant. The combustion chambers’
effect on exergy destruction rates is discussed earlier in section 4.2. Temperature difference has an
important effect on exergy destruction rates, therefore, any improvement to reduce the temperature
difference between inlet and outlet streams would improve the system’s overall performance by reducing
the exergy destruction rates, therefore irreversibilities.

The energy and exergy efficiencies of the system can be defined in several ways. Table 4 summarizes
the results based on the three different methods mentioned below:

1. Heat rejected from the system is not considered as a useful output, but some of this heat is used
to provide the necessary input required for the distillation column input of the methanol synthesis
unit. The work input required for the compressors and pumps are provided by the work output of
turbines. The work input required for the electrolysis modules is provided externally:
(17)

(18)

67
 2. Heat rejected from the system considered as a useful output, some of this output is used to
provide the necessary input required for the distillation column input of the methanol synthesis
unit. The work input required for the compressors and pumps are provided by the work output of
turbines. The work input required for the electrolysis modules is provided externally:
(19)

(20)

Case Study: (Soltanieh et al., 2012) propose to use the work output of the turbines for the work required
in compressors and pumps. The paper uses wind energy to run the electrolysis unit. If electrolysis unit
also used the work output of the turbines, the energy and exergy efficiencies become

(21)

(22)

Table 4Summary of different scenarios with corresponding energy and exergy efficiencies.

Case η Ψ

1 35% 42%

2 64% 46%

3 68% 47%

5. CONCLUSIONS

The integrated system, as modified from a literature work, electrolyzes water using wind energy, stores
46% of the produced H2 and 28% of the produced O2. The remaining O2 is sent from the electrolysis unit
to the Matiant-cycle power plant and goes into complete combustion with natural gas (pure methane).
The combustion product water is sent back to the electrolysis unit and CO 2 is sent to the methanol
synthesis unit. In the methanol synthesis unit, CO2 reacts with the H2 coming from the electrolysis unit,
produces methanol. The overall system does not release CO 2 to the environment. And the heating
requirement of the system is met by the rejected heat within the integrated system. Exergy efficiency of
the overall system increases by 2% if the rejected heat is used as a product, such as water or space
heating. This amount increases by 1% if the work requirements of not only the compressor and pumps
but also the electrolysis modules are met by the work output of the turbines. The integrated system’s
outputs are about 133 MW of electricity energy (60 MW, if the pumps, compressors, and electrolysis
modules’ input is provided) and 0.25 kmol/s methanol.By converting the CO2 to methanol, the methanol
synthesis unit avoids the release of 0.25 kmol/s CO2into the atmosphere. If the system is assumed to
work 8000 hours, the avoided CO2 emissions could reach around 160 tons.

REFERENCES

Chen, Y. Adams II, T.A. and Barton, P.I.“Optimal Design and Operation of Static Energy Polygeneration
Systems”Industrial and Engineering Chemistry Research 50 (2011): 5099–5113.

Cifre, P. G. and Badr, O.“Renewable Hydrogen Utilization for the Production of Methanol” Energy
Conversion and Management 48 (2007): 519–527.

68
Damen, K. Troost, T. Faaij, A. and Turkenburg, W. “A Comparison of Electricity and Hydrogen
Production Systems with CO2Capture and Storage, Part A: Review and Selection of Promising
Conversion and Capture Technologies” Progress in Energy Combustion Science 32 (2006): 215–246.

Davison, J. “Performance and Costs of Power Plants with Capture and Storage of CO2” Energy 32
(2007): 1163–1176.

Dincer, I. and Rosen, M. A. “Exergy: Energy, Environment and Sustainable Development”Oxford:

Elsevier (2007).

Guang-jian, L. Zheng, L. Ming-hua, W. and Wei-dou, N. “Energy Savings by Coproduction:

AMethanol/Electricity Case Study” Applied Energy87 (2010): 2854–2859.

International Energy Agency Technical Report, 2012 Key World Energy Statistics, Website:
http://www.iea.org/publications/freepublications/publication/kwes.pdf; 2012 [accessed 10.01.2013].

Katayama, Y. and Tamaura, Y.“Development of New Green-Fuel Production Technology by Combination

of Fossil Fuel and Renewable Energy”Energy 30 (2005): 2179–2185.

Kotas, T.J. “The Exergy Method of Thermal Plant Analysis”London-Boston: Krieger Publishing Company
(1995).

Mignard, D. Sahibzada, M. Duthie, J. M.and Whittington, H. W.“Methanol Synthesisfrom Flue-Gas CO2

and Renewable Electricity: AFeasibility Study” International Journal of Hydrogen Energy 28 (2003) 455–
464.

Minutillo, M. and Perna, A. “A Novel Approach for Treatment of CO2from Fossil Fired Power Plants, Part
B: The Energy Suitability of Integrated Tri-Reforming Power Plants (ITRPPs) for Methanol Production”
International Journal of Hydrogen Energy 35 (2010): 7012–7020.

Perry, R.H.and Green, D. W.“Perry’s Chemical Engineering Handbook, Seventh Ed.” McGraw-Hill, USA
(2000).

Soltanieh, M. Azar, K. M. and Saber, M. “Development of a Zero Emission Integrated System for Co-
production of Electricity and Methanol through Renewable Hydrogen and CO 2 Capture” International
Journal of Greenhouse Gas Control 7 (2012): 145-152.

Takeuchi, M. Sakamoto, Y.and Niwa, S. et al“Study on CO2Global Recycling System”Science of the Total
Environment277 (2001): 15–19.

69
 INVESTIGATING THE EXERGETIC AND ENVIRONMENTAL EFFECTS OF SUB-COOLING
AND SUPERHEATING PROCESSES ON THE PERFORMANCE OF DIRECT EXPANSION
SYSTEMS
a b c d
M. Z. Sogut , H. Karakoc , Z. Oktay , A. Hepbasli
a
Technical and Computer Applications Center, Military Academy, Ankara, Turkey
b
School of Civil Aviation, Anadolu University, Eskisehir, Turkey
c
Mechanical Engineering Department, Faculty of Recep Tayyip Erdoğan University, Rize, Turkey
d
Department of Energy Systems Engineering, Faculty of Engineering, Yaşar University, 35100 Bornova, Izmir Turkey
mzsogut@yahoo.com, hkarakoc@ anadolu.edu.tr, arifhepbasli@gmail.com, zuhal.oktay@gmail.com

ABSTRACT
Recently, sub-cooling and superheating applications, which are proposed as an approach for the efficient use of
energy, have positive effect on the emission next to increasing the energy efficiency. This study examines
energetic, exergetic and environmental effects of sub-cooling and superheating processes in cooling applications
separately. In this study, direct expansion system (DX), which is used widely in supermarket cooling applications,
are taken into consideration. The evaporator temperature -25°C and 0°C, the refrigerants R-22, R-404A and R-
134a, which are preferred widely in DX system, and minimum and maximum leakage rates %3 and % 30 are taken
as reference for the analysis. According to results, while sub-cooling and sub heating processes increase
approximately 10% of COP and exergy efficiencies based on refrigerants and decrease CO2 emissions for all
refrigerants. The study was concluded with an emphasis on the importance of sub-cooling and superheating
applications, especially for DX systems and the importance of the preference refrigerant.

Key Words: DX systems, COP, Refrigerants, Exergy analyses, CO2 emission

INTRODUCTION

Negative impacts of the energy costs in recent years are an important problem before the sustainable economic
development for all countries suffering difficulties in energy supply. Efficient use of energy and energy saving
works are an obligation for all sectors that are subject to this impact. There are also increasing amount of works
about this obligation due to the heavy energy consumption in the air conditioning and cooling sector which take
around 9% of the world energy consumption(ECCJ, 2002,).

Since 1990s, refrigerant fluids have become prominent due to the environmental threats in particular due to the
ozone consumption and greenhouse gas emissions. Although the ozone consumption is minimized by the
processes limiting refrigerant fluid supply through the international activities of the international society like the
Kyoto Protocol and London Conference, these fluids still have important effect on the climate change based on
global warming. The selection of appropriate system, conducting a configuration with high efficiency and especially
the choice of appropriate fluid for all sectors using these fluids are important due to energy costs and
environmental effects. However, excessive cooling and excessive heating have been a catalyst on the evaporator
and condenser units for especially the DX systems in recent years.

Among the cooling applications, market applications in particular have come to the fore with their heavy use of
refrigerant fluid and heavy energy consumption. Among them, supermarkets with a sales area of 3700-5600 sqm
are structures of heavy energy consumption in commercial applications with an annual average energy
consumption of 2-3 million kWh (Baxter, 2003). Energy consumptions of these systems used for the protection of
food in supermarkets can reach approximately 50% of the total energy consumption of a market. This ratio goes
up to 65% in low capacity market applications (Baxter et all., 2003, IPCC/TEAP;2005, Arthur, 2002). Around 60% -
70% of the energy consumption in these systems is carried out in compressors and condensers (IEA,2003).These
systems have gained importance because of their negative impacts on total consumption costs as well as the
environmental threats caused by refrigerants and energy consumption. Work to develop refrigerant systems is
based on the reduction of energy use requirements in the systems to reduce these effects and on the choice of
environment friendly refrigerant fluids with low GWP and ozone depletion potential in refrigerant systems.

With respect to the storage areas in market applications, it was observed that many refrigerant systems were used
based on the commercial requirements and that the Direct Expansion (DX) system was remarkable among them
with a ratio of up to 60%. This system is followed by secondary lop systems, distributed systems, low-charge

70
multiplex systems and advanced self-contained systems. Supermarket systems with high energy consumption as
well as high amounts of refrigerant fluids have important environmental effects in addition to the direct or indirect
effects of refrigerant fluids. In addition the leakage losses arising from the repair and maintenance processes,
setbacks in operation conditions and installation errors in this type of structure with high amounts of refrigerant
fluid charge have significant impacts on their energy consumption, the efficiency performance of the systems and
the environmental effects they cause. In this type of refrigerating application, refrigerant fluid choices with less
environmental effects, system choice, installation and operation are important, in addition to the capacity of
efficient energy use.

In current systems, reducing energy costs is possible only by the efficient use of energy or use of energy saving
technologies. Saving works in refrigerant systems are examined in many applications from energy consuming fan
and compressor choices to ambient conditions and from the flow properties of system elements like the evaporator
to fluid and system control. Among them, the sub cooling and superheating applications which directly affect
compressor efficiency due to its impact on the cooling capacity, and thus they have an indirect effect on energy
consumption and provide significant savings in the systems. The sub-cooling application reduces the amount of
energy required by minimizing the cooling requirement in the system. So it reduces operation costs in the cooling
system. A temperature fall of 2 °C in sub-cooling applications corresponds to around 1% savings in the operation
costs of the systems. Efficiency is affected adversely by the mass loss up to 30% because of high refrigerant fluid
charge or excessive leakage for providing high pressure in the system of refrigerant applications. Superheating
application ensures optimum use of energy by ensuring the refrigerant fluid control depending on the needs of the
systems. Superheating can be defined as turning a refrigerant fluid which cannot be fully gasified into gas and
preventing entry of the liquid refrigerant to the compressor (Energetic,2012).

The energy saving potential in refrigerant systems is calculated by making energy analyses in traditional
applications. In this type of analysis, the coefficient of performance (COP) is calculated according to the first law of
thermodynamics and the energy consumption potential caused by the systems is evaluated in comparison. These
evaluations ensure the examination of the properties of the fluids used. Energy analyses in the cooling systems
are not sufficient in the evaluations of the irreversibilities and saving potentials based on irreversibility, unlike other
thermodynamic applications. Energy analyses are only the numeric results of energy use. Therefore analyses for
defining efficiency in performance evaluations based on the quality of the system and determining the real
irreversibilities should be carried out by exergy analyses according to the second law of thermodynamics.

Exergy, which is defined as the maximum work which can be produced by a process as it comes to equilibrium
with a reference environment, is a measurement of the quality of the energy stream. For industrial processes,
exergy analysis is a powerful concept in the modern thermodynamic method, which is used as an advanced
tool(Dincer and Rosen, 2005; Schijnel et al., 1998). For refrigeration applications, exergy analysis is also a
powerful concept and the modern thermodynamic method used as an advanced tool. . By use of an exergy
analysis, the process can be optimized into more sustainable processes. The exergy analyses depend mainly on
the effects of environmental parameters such as temperature and pressure. Thus, environmental parameters
affect efficiency directly. Exergy is exhausted or destroyed due to irreversibility in real processes. The basic aim of
exergy analyses is determined as real potential of losses exhausted or destroyed in thermodynamic processes.
Consequently, exergy analyses are an important criterion for all of the thermodynamic evaluations. Today, many
science studies concerning the exergy concept for thermal processes are being made. For example, (Dincer and
Rosen, 2005; Kanoğlu et all. 2007; Hepbasli,2010; Korenes et all., 2011; Wall, 2003 and Sogut,2012) have
performed most extensive studies in the exergy field.

Today, improving environmental awareness requires that the CO2 emission potential caused by this type of system
be kept under control. Many scientific approaches can be seen in these scientific studies. One of them, the
concept of Total Equivalent Warming Impact (TEWI), has been developed to combine the global warning effects of
CO2 released into the atmosphere due to energy used over the lifetime of a system (indirect effect) with the effects
resulting from refrigerant and blowing agent emissions system (direct effect). Contrary to Life Cycle Climate
Performance (LCCP) analysis, which also involves emissions from energy expenditures for production of the
system and further greenhouse gas emissions caused during the production of the refrigeration system, the focus
of a TEWI analyses consists of the emissions following the operation of the system (Rhiemeier et all, 2009)

In this study, the energy and exergy analyses based on different refrigerant fluids were carried out for the Direct
Expansion Systems commonly used in supermarket applications. In addition, the performance effects of the sub-
cooling and superheating applications in the range of 0-5 K were examined for using efficiency of energy and
exergy in DX systems and for reducing the environmental effects caused by refrigerant fluid.

71
SUB-COOLING AND SUPERHEATING OF DIRECT EXPANSION SYSTEM (DX)

Direct Expansion System (DX)

Supermarkets with increasing importance in meeting daily food requirements are commercial structures of heavy
energy consumption with an annual average of 1000 kWh/m2. The cooling techniques used in this type of structure
spend around 50% of the total energy consumption with the cooling and air conditioning systems of different
properties. In small capacities, cooling systems spend approximately 65% of total energy consumption. In these
structures, where cooling applications of various types with large volumes are used, the charge amounts of the
refrigerant fluids used in the system reached up to 2500 kg. In the supermarket applications of cooling systems,
different cooling temperatures are needed to keep different products fresh. In applications of commercial
refrigeration, three different kinds of application are used: supermarkets refrigeration direct expansion system,
indirect system and decentralized system. The low and medium evaporating temperatures of these types of typical
refrigeration systems are around -35 ºC and -10 ºC respectively (SERDP, 1999).

Refrigeration systems used widely in supermarket applications are determined as direct systems or direct
expansion (DX) systems. In these systems, the refrigerant is used to directly transport heat from the areas to be
cooled to the outside areas where heat is released. DX systems have four main components: evaporator,
compressor, condenser and expansion valve, and the main work principle is they are run as a vapour expansion
cooling cycle. The flow schema of a DX system is given in Fig.1.
Condenser

Desuperheating
Condenser
Desuperheating

Condensing
Condensing

Subcooling
Power
Heat exchanger
Subcooling
Power
Expansion valf
Compressor

Compressor
Evaporating Superheating

Evaporating
Superheating

Evaporator Evaporator

a. Normal system b. System with sub-cooler and superheating

Fig.1. Direct expansion system (SERDP, 1999)

In these applications, a separate cooling machine with compressor is often used for each evaporator and cooling
unit. In other cases, several or all evaporators and cooling units in a system are connected to one or more
compressors placed in a central machine room. Condensers may be in an air-cooled machine room, remote
outdoors or indoors. Heat recovery from the condensers may be utilized for space or water heating. The refrigerant
changes from liquid to vapour or gas in the evaporator that is placed in the space to be cooled. The pressure (and
temperature) of the refrigerant gas is increased in the compressor and the gas then passes to the condenser
where the refrigerant is cooled and condenses to liquid. The liquid then passes to the expansion valve, a throttling
device where the pressure of the refrigerant is reduced, and then back to the evaporator and the process starts all
over again (See Figure 1). The refrigerant is then often transported through long tubes, requiring a relatively large
refrigerant charge (Sogut et al.,2012).

Direct expansion systems are the dominant technology for supermarkets world-wide. Refrigerants used include
HCFC R-22 and HFCs such as R-507 and R-404A and to a smaller extent R-134a for medium temperature
applications. But R-22, R-134a and R-404A refrigerants are used commonly in DX systems. The thermodynamic
and environmental parameters of these refrigerants are given in Table. 1.

72
Table. 1 Thermodynamics and environmental parameters of refrigerants

Thermodynamics parameters

Refrigerants R-22 R-134a R-404A

CHF2CF3/CH3CF3/
Chemical Formula CHC1F2 CH2FCF3
CH2FCF3
Molecular weight (kg/kmol) 86.5 102 97.6
Boiling point at (1.013 bar) -40.8 -26.1 -46.6
Enthalpy of evaporation (kJ/kg) 182.5 177.5 140.1
Critical temp. (ºC) 96.1 101.1 72.1
Critical Pressure (bar) 49.9 40.6 37.4
Cplikid (kJ/kgK) 1.26 1.425 1.541
Cpvapor (kJ/kgK) 0.864 1.011 1.2
Environmental Parameters
ODP 0.04 0 0
GWP 1810 1430 3900
Atmospheric life (year) 12 14 16

Sub-cooling of Refrigerants

Excessive cooling of the refrigerant fluids in cooling systems provides an indirect saving by increasing the cooling
capacity of the system. It works best where year-round outdoor temperatures are high and constant. The
effectiveness of the application depends mainly on the cooling fluid. Sub-cooling which causes increases in the
cooling capacity in the system uses an external heat surface. It is not very easy to determine the sub-cooling
temperature in a system. However the general approach in practice is accepted as the difference of the Bubble-
point temperature and actual liquid temperature. Fig.2 illustrates the sub-cooling impact on systems.

Fig.2. Sub-cooling and superheating in cooling cycle (Aleph Zero,2012)

Excessive cooling is used in old or new type direct expansion vapour compression as the most usable system.
The sub-cooling system can be either a dedicated mechanical sub-cooling system or an integrated mechanical
sub-cooling system.

Ambient sub-cooling involves the use of an oversized condenser or an additional sub-cooling heat exchanger to
sub-cool the condensed high-pressure refrigerant. Sub-cooling reduces the enthalpy of the liquid refrigerant, which
is equal to the enthalpy of the two-phase stream of the refrigerant leaving the expansion valve and entering the
evaporator. The specific capacity of the refrigerant in Btu/lb is increased, hence reducing the required mass flow
rate of refrigerant to be compressed and the required compressor electric load. Ambient sub-cooling is effective
only when the head pressure control prevents further reduction in head pressure. Otherwise, reduction in head
pressure is more efficient than simply reducing the liquid temperature. Hence, the savings for ambient sub-cooling

73
are generated during times of low ambient temperature, when head pressure is being maintained at a high level.
This will represent a large percentage of operating hours in cooler climates.

As with ambient sub-cooling, mechanical sub-cooling involves further reduction in enthalpy of the condensed liquid
refrigerant. Mechanical sub-cooling is provided by the expansion of part of the refrigerant liquid in a sub-cooling
heat exchanger, as shown in the Figure below. The expanded refrigerant is compressed from an intermediate
pressure to the common discharge pressure. The specific refrigerant capacity of the main stream of liquid is
increased, thus reducing the compressor electric load for this stream of refrigerant. The additional electric load of
the sub-cooling compressor must be subtracted from the potential savings. However, since the sub-cooling
compressor has a higher suction pressure, its specific work requirement in kWh/kg of refrigerant is less. The result
is an overall savings in electricity usage.

Superheating of Refrigerants

In a cooling cycle, a fluid entering into the evaporator units is theoretically expected to evaporate and enter into the
compressor after receiving heat. However, this cannot be always achieved in evaporator units with different
working zones especially in market applications. In order to prevent liquid entrance to the compressor, usually heat
collectors or driers are added. This causes performance to be affected negatively since the flow process is
affected. In recent years, the superheating process in the evaporator exit in these systems increases the heat
withdrawal effect from the environment to be cooled and removed these adverse effects (see Figure xx). In
addition, the superheating application prevents cooling refrigerant entrance to the compressor and limits the
working range and has positive effect on the system energy consumption. However, it is important to control the
evaporator exit temperature in superheating application. If the temperature is lower than the expected values, this
may cause the entrance of the fluid to the compressor in the evaporator exit. If this value is high, it causes
insufficiency of the fluid amount and excessive work of the compressor for the desired compression work.
Therefore, it is necessary to take the overheating temperature with recommended values determined through
experimental data. On the other hand, superheating temperature for a system can be defined as the difference
between the actual steam temperature of the fluid and dew-point temperature.

EXERGY CONCEPT IN COOLING SYSTEMS

As a measure of the machine’s working activities in cooling systems, the thermal efficiency and the coefficient of
performance (COP) have been defined. The coefficient of performance (COP) for a cooling machine is;
QL
COP  (1)
Wnet
where QL indicates the temperature moving away from the cooled environment and W net indicates the net effort
input into the compressor (Sogut,2011). Cooling machines are systems that work based on the cycle principle and
the energy saving for the cycle may be expressed as follows:
Wnet  QH  QL (2)
where QH is the heat output from the machine. In cooling systems, COP is expressed as efficiency according to
the first law of thermodynamics. In these systems, the working ability of a machine is explained by the second law
of thermodynamics and this is defined as exergy. In the cooling system, exergy efficiency is compared to the
possible highest COPtr of the real COP in the same work parameters. In a cooling machine, COP tr can be
expressed as;
1
COPtr  (3)
TH
TL 1
(Akpınar et al., 2007; Çengel, 2001). Where TH is the outside temperature and TL is the environmental temperature
removed. In this case, the exergy efficiency of the cooling machine can be expressed as;
COP
 II  (4)
COPtr
If equation (1) and equation (3) are united (Hepbasli, 2008);
 QL
TH
QL
 II 
TL
(5)
Wnet
The maximum improvement in exergy efficiency for a process or system is most certainly achieved at the point
where the exergy loss or irreversibility is minimized. The exergy loss or irreversibility is based on the difference

74
between the exergy input and output (Van Gool ,1997). The improvement potential (IP) is developed for
irreversibility processes used in thermal applications (Hammond,2001).
IP  (1  II ).( E xin  E xout ) (6)

TEWI CONCEPT IN COOLING SYSTEMS

At the meeting in Copenhagen, a concept for evaluating the total sum of global warming caused by refrigeration
systems, namely through refrigerant emissions and also through CO2 emissions from the fossil fuel power plant
which produces the energy for the refrigeration system was presented; it is called the Total Equivalent Warming
Impact, TEWI, and was developed by the US Department of Energy and Oak Ridge National Laboratories
(Horst,2000).

An appropriate method of evaluating any refrigerant is the TEWI value as defined in EN 378:2000. This is an
internationally accepted formula that may be used to directly compare the relative effect of different systems. It
represents the environmental effects, based on the global warming potential of the refrigerant, and energy
consumption of the plant over its lifetime. The higher the TEWI value, the poorer the system in terms of its overall
impact on the environment:

TEWI  [ GWP . L . n ]  [GWP . m . (1   re cov ery )]  [ n . E annual .  ]

Leakagelosses Recov ery losses Energy consumption (kgCO2/year) (7)
Direct GWP Indirect GWP
Where TEWI is the total equivalent warming impact, GWP is the global warming potential, L is the amount of
leakage, n is the system operating time, m is the refrigerant charge, αrecovery is the recovery or recycling factor,
Eannual is the annual energy consumption, β is the CO2 emission produced by power generation (Moore,2005).

The global warming effects of different substances are compared with carbon dioxide using an index known as the
global warming potential or GWP. The GWP of a substance of the global warming impact due to the emission of 1
kg of that substance is divided by the global warming impact of 1 kg of carbon dioxide, both effects are measured
over a 100 year time span (Dupont,2005).

The entire climatically relevant impact by the refrigeration process is composed of direct and indirect emission of
greenhouse gases. The direct impact includes the emission of working substances by the greenhouse effect due
to leakages, maintenance and losses during disposal. The indirect impact is related to the CO 2 emissions,
particularly due to the generation of electrical energy in power stations. But also those indirect CO2 emissions
which are created by the energy consumption needed by supplying material, material processing, maintenance,
and waste disposal have been taken into consideration (Eurammon,1996). In TEWI analysis, the CO2 emission
per kWh of generated energy is taken into consideration for the calculation of indirect TEWI contribution. The CO 2
emission for the production of electricity is dependent upon the mixture of energy carriers used in power plants.
The burden on the Earth’s climate resulting from the greenhouse effect is a global problem. The study is based on
published data on CO2 emissions from power plants and utilizes the average value for the regions of North
America, Japan and Europe. In this study, the average value is 0.47 kg CO2/kWh (Hellmann et al., 1997).

RESULTS AND DISCUSSIONS

In this study, energetic, exergetic and environmental effects of the sub-cooling and superheating applications of
the system with different refrigerants and under adiabatic conditions by taking a DX system as the model have
been examined. The sub-cooling and superheating temperature in the study have been taken respectively as 0K
and 5K. The cooling capacities of DX systems vary between 20 kW and 1000 kW. The study takes as reference
836 kg charge amount in the cooling system of a market with a sales area of 3500 sqm for the evaluation of
environmental effects. When the market cooling is examined, we see that the evaporator temperature ranges is
between -35°C to 0°C. However, the most popular range -25°C to 0°C has been accepted as the evaporator
temperatures in the study. The study examines COPs to evaluate the energetic performances in the unit cooling
load depending on the refrigerants taken as reference. Fig.3 includes the COP breakdown of the system for the
sub-cooling and superheating application in the range of 0K and 5K.

75
 003

003

003
COP

002
R-22
002 R-134a
R-404A
002
0 1 2 3 4 5
Subcooling and super heating °C

Fig.3. COP distrubition of refrigerants based on sub-cooling and superheating temperatures

It has been observed that the subcooling and superheating application have direct effects on energetic
performance. However, the fluid has an important part in this effect. It was observed that the COP performance of
the R-22 gas restricted by the international society increased by 4.25% in average in the range of 0K to 5K. R134a
gas which is used commonly in many systems from domestic applications to vehicle air conditioning has better
COP performance by 6.17%. R-404A gas which very common in commercial applications and especially in market
applications has a significant increase of 10.37%. The exergetic performances of each gas per unit cooling load
have been individually examined and the results are given in Fig.4.

0,65

0,6
Exergy efficiency

0,55

0,5
R-22
R-134a
0,45
R-404A

0,4
0 1 2 3 4 5
Subcooling and superheating °C
Fig.4. Exergy efficiencies of refrigerants based on sub-cooling and superheating temperatures

Exergy analysis is an important tool in defining the irreversibilities in system as well as in defining the effects on
the irreversibilities depending on sub-cooling and superheating applications in this type of systems. Refrigerant
effects were also observed in the improvement ratios in the exergetic efficiency of the system for base cooling
capacity. In current applications, the average change ratio of exergetic efficiency in systems using R-22 gas is
4.27% for the superheating application. This effect was determined as 6.22% in the systems using R-134a.
However, the highest effect for the R-404A gas which is intensively used in market applications was found as
10.50%. The increase ratios for both COP and exergetic efficiency is defined with the energy consumption values
in the systems. Therefore, the effects of the sub-cooling and superheating values for each refrigerant were
examined and the results were given in Fig.5.

76
 0,49
R-22
Commpressor power (kW)

0,47 R-134a
0,45 R-404A
0,43

0,41

0,39

0,37

0,35
0 1 2 3 4 5
Subcooling and superheating °C
Fig.5. Exergy efficiencies of refrigerants based on sub-cooling and superheating temperatures

The change in the power consumptions in the systems has different effects for each fluid. It is seen that the
highest change is valid for the systems using R-404A gas. There is 10.07% decrease in average in the energy
consumption for the unit cooling capacity in the systems using this gas. In these systems, the change was found
linear for each temperature increase in the sub-cooling and superheating application. A decrease of 6.38% in
average was determined in the energy consumption for the systems using R-134a gas. This effect has a 3.36%
degree for systems using the least R-22 gas. However, even this effect is an important potential when the full
cooling loads are evaluated. In fact, the most important effect in the sub-cooling and superheating application can
be determined as providing the sustainability of the system thermodynamic parameters by creating flow control.
This affect causes a decrease in the flow rates. The change in mass flows for each refrigerant in the sub-cooling
and superheating application was examined and the results were given in Figure 6.
12
R-22
11 R-134a
R-404A
Mass flow (g/s)

10

6
0 1 2 3 4 5
Subcooling and superheating °C
Fig.6. Mass flow changes of refrigerants based on sub-cooling and superheating temperatures

For the maintenance and repair process of systems and increase of efficiency, the fluid is charged usually with the
pressure value taken as basis. However, due to the temperature based change of effectiveness, optimum point is
usually exceeded and high charge is perceived as an advantage. For some operation temperatures, this effect
causes excessive work of the compressor and high energy consumption. In the sub-cooling and superheating
application, it also affects the charge ratio in low flow systems due to temperature and because it controls the
vapour phase into the compressor. The highest decrease in this effect has been determined as 12.55% in systems
using R-404A gas. This effect has been found to be 8.44% for the R-134a gas and 6.57% for the R-22 gas.

Market applications are systems with intensive refrigerant charge. The most important environmental effect in
these systems is the direct effect based on the leakage ratio of the refrigerant. These effects are subject to indirect
emission releases caused by the energy consumptions of the system. However, the most important actual effect in
the system performances is the fluids. The GWP potentials of fluids have direct influence on the effect based on
the leakage ratio in the emission releases. It is seen that the lowest and highest leakage ratio are in the range of
3% - 30% in the market applications. In the study, the TEWI values have been calculated separately taking the

77
refrigerants as reference for both peak points. Fig.7 includes the change of TEWI values for the sub-cooling and
superheating applications in the 3% leakage ratios.
30
R-22 R-134a R-404A
25
TEWI MillionCO2/year

20

15

10

0
0 1 2 3 4 5
Subcooling and superheating °C
Fig.7. TEWI changes of refrigerants based on 3% leakage rate

When we consider the refrigerants, we noticed that the effects of the sub-cooling and superheating application are
very limited. This effect is limited by 5.88*10-2% for the R-22 gas, by 15.28*10-2% for the R-134a gas and by
11.1910-2% for the R-404A gas. This limited effect is caused by the indirect effect of the energy change on the
emission potential. In this type of systems, emission potential based on the refrigerant is important. Especially the
TEWI value of the R-404A gas which is heavily used in market applications has the highest degree as 26.17
Million CO2/year. It has a 47.50% higher emission potential than the R-22 gas which it replaces in this kind of
systems. Similar TEWI analyses have been conducted for the leakage ratio of 30% and the results are given in
Fig.8.
140
R-22 R-134a R-404A
120
TEWI Million CO2/year

100

80

60

40

20

0
0 1 2 3 4 5
Subcooling and superheating °C

Fig.8. TEWI changes of refrigerants based on 30% leakage rate

The increase in the leakage ratio increases the direct effect on emission. However the emission effect of the sub-
cooling and superheating application is a decreasing effect for all fluids. The study found the average increase for
the sub-cooling and superheating temperatures for the R-22 gas, R-134a gas and R404A gas to be respectively
.25*10-2%, 3.25*10-2% and 2.36*10-2 %. For the 30% leakage ration, the R404A gas has a higher potential of
47.78% than the R-22 gas. However, the emission potential of the R-134a gas has 30.6% less effect than the R-22
gas and 60.04% less effect than the R-404A gas.

CONCLUSIONS

The energetic, exergetic and environmental effects of the sub-cooling and superheating applications for different
refrigerants in the DX system taken as reference in this work have been examined separately. The results
obtained are given below.

78
 Depending on the refrigerants, the sub-cooling and superheating applications cause an increase of around
10% in the energetic and exergetic performances of the systems.
 Depending on the choice of refrigerants, an effective sub-cooling and superheating application causes a
saving of up to 10% in the energy consumption.
 The sub-cooling and superheating application controls the mass flows of the systems and increases the
compressor performance in particular. This prevents the unnecessarily high cooling charge.
 The sub-cooling and superheating application has a very limited effect in the total emission effect of the
system. However the choice of refrigerant has direct effect on the TEWI value.
In conclusion, an effective sub-cooling and superheating application has direct effect on the energetic and
exergetic performance of the system. The choice of refrigerant has also direct effect on the energetic, exergetic
and environmental performances. Therefore, choice of correct system as well as natural fluid choice together with
low GWP value will further increase these effects.

References

Akpinar E.K., Hepbasli A., 2007, A comparative study on exergetic assessment of two ground- source
(geothermal) heat pump systems for residential applications, Building and Environment, 42 (2007) 2004–2013

Aleph Zero - Refrigeration, A Simple Practical Cycle (2012)http://www.alephzero.co.uk/ref/practcyc.htm

Sogut Z. 2011. A study on the exergetic and environmental effects of commercial cooling systems, Int. J. Exergy,
Vol. 9, No. 4.

Arthur D. Little, Inc. (ADL)., Global comparative analysis of HFC and alternative technologies for refrigeration, air
conditioning, foam, solvent, aerosol propellant, and fire protection applications, Final Report to the Alliance for
Responsible Atmospheric Policy 2002. Available at http://www.arap.org/adlittle/HFCstudy3-22JD.pdf.

Baxter V.D. Advances in Supermarket Refrigeration Systems, Oak Ridge National Laboratory, TN 37831-
6070,2003, Oak Ridge www.arb.ca.gov/cc/commref/adv_supmkt_ref_syst.pdf
Baxter, Van D. and David H. Walker.,2003, Analysis of advanced low-charge refrigeration for supermarkets,
ASHRAE Transactions.

Cengel Y., Boles M.A. 2001. Thermodynamics: an engineering approach, 4th ed. New York: McGraw-Hill
HEPBASLI A. 2008. A key review on exergetic analysis and assessment of renewable energy resources for a
sustainable future, Renewable & Sustainable Energy Reviews 12(3):593-661.

Dincer, I. and Rosen, M.A. (2005) ‘Thermodynamic aspects of renewable and sustainable development’,
Renewable and Sustainable Energy Reviews, Vol. 9, pp.169–189.

Dincer I, Rosen M A, Exergy, Energy, Environment and Sustainable Development, Elsevier 2007.

DUPONT. 2005. Dupont refrigerants the science of cool, Du Pont de Nemours (Deutschland) GmbH, Germany,
www.refrigerants.dupont.com

Eurammon, 1996. Evaluation of the environmentally friendly refrigerant ammonia according to the TEWI Concept,
NH3 for ecologically friendly future, Frankfurt, Germany http://www.eurammon.com

ECCJ, 2002, Energy Conservation Center, Japon, http://www.eccj-or-jp/summary/local0303/eng/03-01.html,

http://www.eccj-or-jp/databook/2002-2003e/03-04.html, http://www.eccj-or-jp/databook/2002-2003e/03-05.html.

Energetic, The effect of sub-cooling on refrigeration processes, Energistics.biz, 1361 Shorebird Lane, Carlsbad,
California, USA, 2012, www.energistics.biz/pdf/SubCoolingEffect.pdf

Hammond A.J. 2001. Stapleton, Exergy analysis of the United Kingdom energy system. Proceedings of the
Institute of Mechanical Engineers 2001;215 (2): 141–62.

Hellmann J., and Barthélemy P. 1997. AFEAS-TEWI III study: results and evaluation of alternative refrigerants,
Solvay Fluor und Derivate GmbH Technical Service-product Refrigerants, Bulletin no. C/11.97/06/E, page 5.

Hepbasli A., “ A review on energetic, exergetic and exergoeconomic aspects of geothermal districit heating
systems(GDHSs), Energy Conversion and Management Volume 51, İssue 10, October 2010, Pages 2041-2061

79
Horst K. 2000. Refrigerant use in Europe, ASHRAE journal September 2000, www.ashraejournal.org

IEA (International Energy Agency), 2003, IEA Annex 26: Advanced süpermarket refrigeration/heat recovery
systems, Final Report Volume 1—Executive Summary,Compiled by Van D. Baxter, Oak Ridge National
Laboratory, April 2003.

IPCC/TEAP, Refrigeration, Draft of IPCC/TEAP Special Report on Ozone and Climate, Chapter 4,
Intergovernmental Panel on Climate Change/Technology and Economic Assessment Panel.Cambbridge
Üniversitesi, Newyork,USA, 2005.

Kanoglu M, Dincer I, Rosen M A, Understanding energy and exergy efficiencies for improved energy management
in power plants. Energy Policy, 2007; 35:3967–3978.
Koroneos C.J., Nanaki E.A., Xydis G.A., “Exergy analysis of energy use in Greece” Energy policy Volume 39,
İssue 5 May 2011, Pages 2475-2481

Rhiemeier J.M.,Kauffeld M., Leisewitz A., Comparative assessment of the climate relevance of supermarket
refrigeration systems and equipment, Environmental Research of the Federal Ministry of Environment Nature
Conservation and Nuclear Safety Research Report 206 44 300, UBA-FB 001180/e, March 2009, Federal
Environmental Agency,

Moore D. 2005. A comparative method for evaluating industrial refrigerant systems, Sabroe Ltd. (revA). Nov.
2005, www. Sabroe org.

Schijnel, P.P.A.J.V., Kasteren, J.M.N. and Janssen, F.J.J.G. (1998) ‘Exergy analysis–a tool for sustainable
technology – in Engineering Education’, Eindhoven University of Technology,The Netherlands.

SERDP, Refrigerant Subcooling, Technology for improving cooling system performance, U.S. Department of
Energy by the Pacific Northwest National Laboratory, 1999, www.eren.doe.gov/femp/

Sogut M.Z. “A research on exergy consumption and potential of total CO2 emission in the Turkish cement sector”,
Energy Conversion and Management 56 (2012) 37–45

Sogut M.Z., Oktay Z., Karakoc H., investigation of the effects of leakage rate on exergetic and environmental
performances of refrigerants in the direct expansion system, Global Conference on Global Warming 2012,
İstanbul.

Van Gool W. 1997 Energy policy: fairly tales and factualities, in: O.D.D. Soares, A. Martins da Cruz, G. Costa
Pereira, I.M.R.T. Soares, A.J.P.S. Reis (Eds.), Innovation and Technology—Strategies and Policies, Kluwer,
Dordrecht, 1997.

Wall G. Exergy tools. In: Proceedings of the Institution of Mechanical Engineers. Wilson Applied Science and
Technology Abstracts Plus Text; 2003. p. 125–36

80
 INVESTIGATE ENERGY EFFICIENCY BY MAKING EXERGY ANALYSIS
IN THE CEMENT SECTOR

H. G. KANDILCI and H.SARAC

Yıldız Technical University, Chemical and Metallurgical Faculty, Chemical Engineering Department
Davutpaşa - Esenler 34210 İstanbul –TURKEY
h.gizemkandilci@gmail.com, hsarac@yildiz.edu.tr

ABSTRACT
In Turkey, cement producing is one of the most important sector in terms of energy consuming and costs.
Owing to this reason, using energy in efficiently and decline in energy consumption have an importance in
this sector.
This study aims to make energy and exergy analysis of a cement factory in Adana and to realize the cost of
energy efficiency, by using the actual operational data. According to energy and exergy analysis, the energy
efficiency value for the raw mill are obtained to be 69% in this study, whereas the exergy efficiency value for
that are found to be 11%.
Key Words: Cement, energy and exergy analysis, efficiency.

Introduction
By the mid 1970s, the growth in energy resource consumption that occurred in each passing year was not a
source of general concern. Many researchers concern that unless corrective actions were undertaken,
difficulties would be encountered in providing energy for future needs (Moran, 1992).
The energy and mass balances are the fundamental method of a process analysis. They make the energy
analysis possible, points at the needs to develop the process are the key to optimization and is the main for
developing the exergy balance. Analysis of the energy balance results would define the efficiency of energy
utilization in particular parts of the process and enable to compare the efficiency and the process
parameters with the recently feasible values in the most modern installations. They will also form the
importance of the processes requiring consideration, either due to their excessive energy consumption or
due to their particularly low efficiency.
The exergy analysis is the modern thermodynamic method used as an advanced tool for engineering
process evaluation (Szargut et al., 1998). The exergy analysis is on the basis of both the first and the
second laws of thermodynamics. Both analyses make use of also the material balance for the considered
system.
Exergy analysis also bring about a better concept of the influence of thermodynamic phenomena on the
process effectiveness, comparison of the priority of different thermodynamic factors, and the determination
of the most effective ways of developing the process under consideration (Szargut et al., 1998). A true
concept of exergy and the insights it can provide into the efficiency, environmental impact and sustainability
of energy systems are required for the engineer or scientist working in the area of energy systems and the
environment (Dincer, 2000 and Dincer, 2002).

Theoretical analysis
For a general steady state and steady-flow process, the following balance equations are implemented to
detect the work and heat interactions, the rate of exergy decline, the rate of irreversibility, the energy and
exergy efficiencies (Kotas,1985 and Bejan, 1988).
The mass balance equation can be expressed in the rate form as
(1)
where is the mass flow rate, and the subscript ‘in’ explains inlet and ‘out’ for outlet.
The general energy balance can be expressed as
(2)
(3)
where is the rate of net energy transfer in, is the rate of net energy transfer out by heat, work and
mass, is the rate of net heat input, is the rate of net work
output.

81
Presuming no changes in kinetic and potential energies with any heat or work transfers, the energy balance
given in Eq. (3) can be simplified to flow enthalpies only:
(4)
where h is the specific enthalpy. The general exergy balance can be expressed in the rate form as

or (5)

with
(6)
where is the heat transfer rate through the boundary at temperature Tk at location k, is the work
rate, is the flow exergy, s is the specific entropy and the subscript zero shows properties at the dead state
of P0 and T0.
The exergy destroyed ( ) or the irreversibility ( ) may be expressed as follows:
(7)
where is the rate of entropy, while the subscript ‘0’ means conditions of the reference
environment (Dincer et al., 2003 and Dincer et al., 2004).
Different ways of formulating exergetic efficiency offered in the literature have been given in detail
elsewhere (Cornelissen, 1997). The exergy efficiency explains all exergy input as used exergy, and all
exergy output as utilized exergy.
Therefore, the exergy efficiency ε1 becomes

(8)

Often, there is a part of the output exergy that is out of use, i.e. an exergy wasted, to the
environment. In this case, exergy efficiency may be written as follows (Wall, 2003):

(9)

The rational efficiency is defined by Kotas and Cornelissen (Kotas, 1985 and Cornelissen, 1997) as
the ratio of the desired exergy output to the exergy used, namely

(10a)

where is all exergy transfer rate from the system, which must be regarded as forming the
desired output, plus any by-product that is produced by the system, whereas is the required
exergy input rate for the process to be performed. The exergy efficiency given in Eq. (10a) may also
express as follows (Tores et al., 1998):

(10b)

To define the exergetic efficiency both a product and a fuel for the system being analyzed are identified.
The product stands for the desired result of the system (power, steam, some combination of power and
steam, etc.). Accordingly, the definition of the product must be consistent with the purpose of purchasing
and using the system. The fuel stands for the resources expended to generate the product and is not
necessarily restricted to being an actual fuel such as a natural gas, oil, or coal. Both the product and the
fuel are expressed in terms of exergy (Moran,1999).

Description of cement process

The cement industry has an important role in the economy on the basis of its production. During the
production of cement, natural resources are used up in large amounts. The most important raw materials for
the production of cement are limestone (CaCO3) and clay or calcareous clay in which both components are
already naturally mixed. The components are milled and dried with flue gases from the clinker kiln. This
section of the process is given in Fig. 1.

82
Figure 1. A section of the process
Depending on the type of cement to be produced, the following products may be added to the dried
limestone immediately after: pyrite ash, fly ash from coal fired power plants, sandy clay and filter ash from
the electrostatic precipitator present. The mixture obtained is ground and immediately after fired in a rotary
furnace to cement clinkers. For heating, various fuels and other combustible materials, e.g. coal dust,
petroleum coke, etc., are used. Depending on the type of preheating of the material, it is separated from
grate and cyclone preheating, whereby the starting materials are preheated to 700-800°C. The raw
materials pass through the rotary furnace towards the flame. In the hottest zone, the material being fired
reaches temperatures of around 1450 °C. After fast cooling with ambient air, the clinkers are milled,
together with gypsum, to give ready cement. A part of the process is given in Fig. 2 (Çelen. 1998 and
Gürüz, 1997).

Figure 2. A part of cement production process

Results and discussion

Here, the energy and exergy modeling technique discussed in the previous section is applied to a raw mill
in the cement factory studied using actual data.
Energy analysis of the raw mill

83
In order to analyze the raw mill thermodynamically, the following assumptions are made:
(a) The system is assumed as a steady state, steady flow process.
(b) Kinetic and potential energy chances of input and output materials are ignored.
(c) No heat is transferred to the system from the outside.
(d) Electrical energy produces the shaft work in the raw mill.
(e) The change in the ambient temperature is neglected.
Under the above-mentioned conditions and using the actual operational data of the plant, an energy
balance is applied to the raw mill. The references for enthalpy, entropy and input energy are considered for
the calculations. The complete energy balance for the system is shown in Table 1. Energy flow of the raw
mill is also illustrated in Fig. 3.
In addition, the enthalpies of the materials going into and leaving the raw mill are given for the chemical
components in Table 2, while the energy balance is given in Table 1. Relatively good consistency between
the total heat input and total heat output is obtained.
Table 1 Mass and energy balances of the raw mill
3 3
INPUT MATERIAL T(K) m(kg/h) ʋ(Nm /h) Cp(kJ/kg.K) Cp(kJ/Nm .K) Q(kJ/h)

Limestone Tl ml Cpl 95 962 697.60

Clay Tc mc Cpc 14 681 654.40

Return from separator Tr mr Cpr 59 887 878.40

Moisture in the limestone Tml mml Cpml 42 638 780.80

Moisture in the clay Tmc mmc Cpmc 5 814 379.20

Gas Tg ʋg Cpg 521 342 976.00

Leaking air Tla ʋla Cpla 78 578 560.00

Dust in the gas Td ʋd Cpd 19 272 704.00

Heat from electrical

28 591 088.40
energy

Total 866 770 718.80

3 3
OUTPUT MATERIAL T(K) m(kg/h) ʋ(Nm /h) Cp(kJ/kg.K) Cp(kJ/Nm .K) Q(kJ/h)

Farine Tout mf Cpf 205 474 008.88

Gas Tout ʋg Cpg 247 582 464.00

Leaking air Tout ʋla Cpla 116 442 355.00

Moisture Tout mm Cpm 4 050 073.44

Steam Tout ms Cps 27 051 787.70

Total 600 600 689.02

84
Figure 3. Energy flow of the raw mill
Energy efficiency of the raw mill
Energy efficiency of the raw mill is calculated from the following relation

(11)

Using energy analysis values and Eq. (11), the energy efficiency of the raw mill is calculated as follows:

The overall thermal efficiency of the raw mill has been found to be 69% and is close to the best practice
with the current technological limitations. The waste heat has been estimated at about 31% of the energy
input. This has represented an improvement of about 31% in terms of primary energy efficiency of the raw
mill.

Exergy analysis of the raw mill

The irreversibility of each of the components is calculated from the exergy consideration, while it may also
be found using the entropy balance equations. Entropy balance of the raw mill is illustrated in Table 3, while
exergy balance of the raw mill is shown in Table 4. The following assumptions are made in the calculations.
(a) The system is assumed as a steady state, steady flow process.
(b) Chemical exergies of the substances are neglected.
(c) Kinetic and potential exergies of materials are ignored.
Total exergy values of the input and output materials are calculated to be respectively 98 894 113.15 and
10 712 208.10 kJ/h. Exergy flow of the raw mill is also illustrated in Fig. 4. The lost heat for the raw mill in
the energy flow is obtained to be 266 170 032.78 kJ/h in this study, whereas the lost heat for the raw mill in
the exergy flow is found to be 88 181 905.05 kJ/h.

85
Table 2 Enthalpy balance of the raw mill
3
INPUT MATERIAL T(K) T0(K) 3 Cp(kJ/kg.K) Cp(kJ/Nm .K) Δh(kJ/kg)
m(kg/h) ʋ(Nm /h)

Limestone Tl T0 ml Cpl 0

Clay Tc T0 mc Cpc 0

Return from separator Tr T0 mr Cpr 10 949 030.40

Moisture in the limestone Tml T0 mml Cpml 0

Moisture in the clay Tmc T0 mmc Cpmc 0

Gas Tg T0 ʋg Cpg 280 283 136.00

Leaking air Tla T0 ʋla Cpla 0

Dust in the gas Td T0 ʋd Cpd 10 361 344.00

3
OUTPUT MATERIAL T(K) T0(K) 3 Cp(kJ/kg.K) Cp(kJ/Nm .K) Δh (kJ/kg)
m(kg/h) ʋ(Nm /h)

Farine Tout T0 mf Cpf 37 565 885.28

Gas Tout T0 ʋg Cpg 45 264 384.00

Leaking air Tout T0 ʋla Cpla 21 288 630.00

Moisture Tout T0 mm Cpm 740 456.64

Steam Tout T0 ms Cps 4 945 756.20

86
Table 3 Entropy balance of the raw mill
3 3
INPUT MATERIAL T(K) T0(K) m(kg/h) ʋ(Nm /h) Cp(kJ/kg.K) Cp(kJ/Nm .K) ΔS(kJ/kg.K)

Limestone Tl T0 ml Cpl 0

Clay Tc T0 mc Cpc 0

Return from separator Tr T0 mr Cpr 33 494.51

Moisture in the limestone Tml T0 mml Cpml 0

Moisture in the clay Tmc T0 mmc Cpmc 0

Gas Tg T0 ʋg Cpg 630 320.76

Leaking air Tla T0 ʋla Cpla 0

Dust in the gas Td T0 ʋd Cpd 23 301.33

3 3
OUTPUT MATERIAL T(K) T0(K) m(kg/h) ʋ(Nm /h) Cp(kJ/kg.K) Cp(kJ/Nm .K) ΔS(kJ/kg.K)

Farine Tout T0 mf Cpf 114 918.94

Gas Tout T0 ʋg Cpg 138 469.65

Leaking air Tout T0 ʋla Cpla 65 124.69

Moisture Tout T0 mm Cpm 2 265.15

Steam Tout T0 ms Cps 15 129.71

87
Table 4 Exergy balance in the raw mill
Δh(kJ/kg) T0(K) ΔS(kJ/kg.K) Ex(kJ/h)
INPUT MATERIAL

0 0 0
Limestone T0

0 0 0
Clay T0

10 949 030.40 33 494.51 1 068 149.65

Return from separator T0

0 0 0
Moisture in the limestone T0

0 0 0
Moisture in the clay T0

280 283 136.00 630 320.76 94 338 512.23

Gas T0

0 0 0
Leaking air T0

10 361 344.00 23 301.33 3 487 451.27

Dust in the gas T0

Total 98 894 113.15

Δh(kJ/kg) T0(K) ΔS(kJ/kg.K) Ex(kJ/h)

OUTPUT MATERIAL

37 565 885.28 114 918.94 3 664 798.23

Farine T0

45 264 384.00 138 469.65 4 415 837.22

Gas T0

21 288 630.00 65 124.69 2 076 845.33

Leaking air T0

740 456.64 2 265.15 72 236.40

Moisture T0

4 945 756.20 15 129.71 482 490.92

Steam T0

Total 10 712 208.10

Figure 4. Exergy flow of the raw mill

88
Exergy efficiency of the raw mill
The exergy efficiency of the raw mill is calculated from

(12)

Using exergy analysis values and Eq. (12), the exergy efficiency of the raw mill is computed as follows:

The overall exergy efficiency of the raw mill was determined as 11%.

Conclusions
The aim of this study has been to determine energy and exergy efficiencies for a raw mill in a cement
factory. Energy balance, energy and exergy efficiencies of the raw mill have been analyzed using the actual
plant operational data.
The main conclusions from this study may be summarized as follows:
(a) Exergy analysis is a advanced tool, which has been successfully and effectively used in the design and
performance evaluation of energy-related systems.
(b) The energy efficiency value for the raw mill are obtained to be 69% in this study, whereas the exergy
efficiency value for that are found to be 11%.
(c) Heat losses that show up especially at the beginning stage of the process shows problem with the
efficiency of the system. Heat losses will decrease if necessary precautions are taken in the raw mill. It will
also cause saving of fuel at the rotary kiln.
(d) This study has indicated that exergy utilization in the raw mill has been worse than energy utilization. In
other words, this process represents a big potential for increasing the exergy efficiency. Some suggestions
are given below to increase the efficiency of the raw mill.
• The mixing temperatures of the materials, which constitute farine and gas coming from the preheater
cyclones and being the energy source of the raw mill should be reexamined. Cement plants are designed at
a full production capacity. The system productivity would be increased when the raw material input capacity
would be held at the highest capacity and when the turning speed of the farine mill and the material flow
rate could be decreased because the system output temperature would be raised.
• The waste energy in the raw mill and the leaving temperature of the farine could be recovered by means
of a heat recovery system. Thus, the efficiency of the system would be increased.

Nomenclature
E energy, kJ
energy rate, kW
Ex exergy, kJ
exergy rate, kW
h specific enthalpy, kJ/kg)
I irreversibility, exergy consumption, kJ
irreversibility rate, exergy consumption rate, kW
m mass, kg
mass flow rate, kg/s
P pressure, Pa
Q heat transfer, kJ
heat transfer rate, kW
s specific entropy, kJ/kg.K
entropy rate, kW
T temperature, K
W work, kJ
work rate or power, kW

Greek Letters
η energy (first law) efficiency, %
ε exergy (second law) efficiency, %
ψ flow exergy, kJ/kg

Indices
c clay
89
dest destroyed
d gas dust
f farine
gen generation
g gas
in input
m moisture
l limestone
la leaking air
mc moisture clay
ml moisture limestone
r return from separator
s steam
out outlet, existing
0 dead state or reference environment

References
[1] M. J. Moran.1982. Availability analysis. New Jersey: Prentice-Hall, Inc.
[2] Szargut, J., D. R. Morris and F.R. Steward.1998. Exergy analysis of thermal and metallurgical
processes. Hemisphere Publishing Corporation.
[3] I. Dincer. 2002. The role of exergy in energy policy making. Energy Policy 30 137–149.
[4] I. Dincer. 2000. Thermodynamics, exergy and environmental impact. Energy Sources 22 723–732.
[5] T.J. Kotas.1985. The exergy method of thermal plant analysis. Tiptree, Essex: Anchor Brendon Ltd.
[6] F. Çelen. 1998. Exergy analysis in a rotary burner with pre-calcination in dry system cement production.
Kırıkkale University Institute of Science & Technology, M.S. Thesis.
[7] H. K. Gürüz. 1977. Mass and energy balances in cement factories. J Chamber Chem Eng (Turkish).
[8] A. Bejan. 1988. Advanced engineering thermodynamics. Wiley, New York.
[9] Dincer, I., M. M. Hussain and I. Al-Zaharnah. 2003. Energy and exergy use in the industrial sector of
Saudi Arabia. Proceedings of the Institute of Mechanical Engineers 217 481–492.
[10] Dincer, I., M. M. Hussain and I. Al-Zaharnah. 2004. Energy and exergy utilization in transportation
sector of Saudi Arabia. Applied Thermal Engineering 24 525–538.
[11] R. L. Cornelissen. 1997. Thermodynamics and sustainable development: The use of exergy analysis
and the reduction of irreversibility. PhD thesis, University of Twente, The Netherlands.
[12] G. Wall. 2003. Exergy tools, in: Proceedings of the Institution of Mechanical Engineers. Wilson Applied
Science and Technology Abstracts Plus Text, pp. 125–136.
[13] Torres, E. A. and W. L. R. Gallo. 1998. Exergetic evaluation of a cogeneration system in a
petrochemical complex. Energy Conversion and Management 16–18 1845–1852.
[14] M. J. Moran. 1999. Engineering thermodynamics, in: F. Kreith (Ed.). Mechanical Engineering
Handbook, Boca Raton: CRC Press LLC.

90
 INVESTIGATION OF EXERGETIC AND ENVIRONMENTAL PERFORMANCE OF HFCs
WITH 20 YEAR AND 100 YEAR GWP METRIC

M. Z. Sog ut
Technical and Computer Applications Center, Military Academy, Ankara, Turkey
mzsogut@yahoo.com

Abstract
In recent years, Hydroflourocarbons (HFCs) used instead of CFCs, which have become the dominant in the
refrigeration sector, have lower ozone depletion potential (ODP) than CFCs. However, they have worse
performance from CFCs considering COP, exergy efficiency and global warming potential (GWP). Besides their
emission potentials are also higher performance taken into GWP potentials consideration.

In this study, first, COP and exergy efficiencies of HFCs are examined considering unit cooling capacity and
evaporator temperature -20 °C and 10 °C. Then, the total equivalent warming impact (TEWI) of each refrigerant
are found as based on metric values of GWP 20 and 100 years for each refrigerant. Besides, the changes of
COP, exergy efficiency and TEWI for each refrigerant according to R-22 are examined separately. Furthermore,
for each refrigerant, the changes according to R22, which is taken reference refrigerant, are examined separately.
In the end of the study, the assessments and recommendations about environmental performances of HFCs,
which are higher GWP than R-22 gas, have been made.

Key Words: Refrigerants, HFCs, Exergy, Global Warming Potential, TEWI

INTRODUCTION

Today, when climate change is effective, refrigerants, solvents, foams and fire extinguishers are important
substances strengthening global warming and causing ozone consumption (EPA,2004). It is noticed that
PFC’s (perfluorocarbons) and SF6s (sulphur hexafluorides) are used heavily in the production process of
these substances with intensive use of HFC. The GWPs (Global Warming Potentials) of these gases are far
more effective than their CO2 emissions and their atmosphere life reaches to thousand years. Refrigerants
are the ones using the HFCs most.

The air condition and cooling sector takes around 9% of the world energy consumption and it is a sector
where its emission effects caused by refrigerants are discussed in recent years with respect to sustainable
life and environment (ECCJ, 2003). The consumption of refrigerants is increased in order to meet the
comfort requirements needed in the varying environmental conditions as well as the demand increase in the
air conditioning and cooling systems due to economic requirements. This increase indirectly causes an
increase in the energy demand and emission potentials. Likewise, the refrigerant effects reached to around
80% in 2010 with 94,5 MMTCE (million metric tons of carbon equivalent). Estimations indicate that this
effect will continue with similar rates in 2020. These gases are followed by foams by around 12,27% with
14,44 MMTCE (EPA,2004).

It has been an international concern to ensure the control of production and consumption of the refrigerants
which are important catalysts in the weather change based on global warming and ozone layer depletion
with the emissions they caused. For decreasing this effect, Montreal Protocol and European Council
Directive (3093/94) restricted the use and sales of hydroclorofluorocarbon (HCFC) and chlorofluorocarbon
(CFC) gases that were used as safe fluids instead of refrigerants like ammonium which was commonly used
in cooling systems during 1930s (Horst, 2000). As an alternative to them, hydro fluorocarbons (HFCs) and
their mixtures like R-134a, R-404A, R-507, R-407C ve R-410A started to be used widely in market
applications. However, despite their low potential of ozone layer depletion (ODP), these gases have high
global warming potentials (GWP) indicating that there is still a problematic process with respect to
environmental impact (Halozen,2007).

The air conditioning and cooling systems, thermodynamically inappropriate system choice, wrong assembly,
missing and wrong capacity determination, repair and maintenance activities, uncontrolled leakages and
similar irreversibilities based on many effects cause more emission release. In determining the emission
effects of the system, the method of Total Equivalent Warming Impact (TEWI) developed for the objective is
a methodology for evaluation of greenhouse gases especially CO2 emissions for cooling applications
(IPCC/TEAP, 2005; Arthur, 2002).

This study examines separately the changes for the 20 and 100 years of GWP values of the emission
potentials caused by gases with their thermodynamical performances including the irreversibilities under
unit cooling load of the HFC fluids used in the air conditioning and cooling applications.

91
AIR CONDITIONING AND COOLING SYSTEMS AND HFCS

Air conditioning and cooling systems find wide application area like house type, commercial, industrial,
vehicle air conditioning and transportation (land, sea, air), cooling applications. Among these application
areas, refrigerant amounts varying from 50 gr to several tons depending on the cooling capacities of the
refrigerant systems (IPCC/TEAP, 2005).

It is observed that steam compressed cooling cycle is taken as basis throughout these cooling systems.
This model operating on a cycle in thermodynamical process is taken as basis in the study. The cooling
cycle, the refrigerant at low pressure is increased to high pressure and sent to the condenser by the
compressor. Condensation takes place in the condenser and the refrigerant turns to liquid under low
pressure through the expansion valve. Cooling takes place here with the effect of the heat from the internal
environment and the fluid comes to the compressor with the steam phase (Sogut et L.,2012). Cooling
performance is affected by the capacitive qualities of the mechanical system preferred in the system.
However, the effect of the cooling fluid is important. Choice of refrigerant for a system should include a safe
quality where the toxic and combustible effects are taken into consideration. On the other hand,
performance parameters should also be taken into consideration where atmospheric life, GWP and OD
qualities, exergetic efficiency, mechanical qualities and investment costs are taken into consideration.

Together with the development of the evaporation compressed cooling cycles starting from 1990s,
chlorofluorocarbons (CFCs) being used to replace refrigerants like ammonium in cooling systems were
replaced through the course of time by hydroclorofluorocarbon (HCFC) gases with higher performance than
CFCs like R22. The emissions caused by these gases became important catalysts in the ozone layer
depletion and global warming based climate change. For decreasing this effect, Montreal Protocol and
European Council Directive (3093/94) restricted the use and sales of hydroclorofluorocarbon (HCFC) ve
chlorofluorocarbon (CFC) gases that were used as safe fluids instead of refrigerants like ammonium which
was commonly used in cooling systems during 1930s (Halozen,2007). Nevertheless there is a tendency of
decrease in recent years in the use of these gases since 1994. Fig.1 includes the breakdown of refrigerants
between 1990 and 2006.

Fig.1 Refrigerant breakdown between 1990-2006(Clodic et al., 2010)

The refrigerant capacity consumed between 1990-2006 had an increase of around 120%. In this process, it
is seen that the CFC gases with rather bad environmental parameters are in a declining tendency but the
use of HFC gases is becoming more and more widespread while the sector mainly prefers HCFC gases
(Clodic et al., 2010; Jubb et al.,2002, Unep, 2006). In this flow trend, a consumption of around 566 kilo
tonnes is expected in the developed countries as of the end 2013 with respect to the consumption of HCFC
gases that are being used. This value is around 3.4 times more than the CFC gases. This effect
unfortunately indicates that the use of HCFC gases will continue in the sector. However, it can be said that
this effect is low especially in developed countries and that HFC gases are preferred instead.

As to the refrigerants, GWP is a valid measurement for the greenhouse gas emissions equal to the CO2
emissions of a refrigerant for a specific time period. This measurement is usually important for the decision

92
makes while UNEP/IPCC (United Nations Environment Programme/Intergovernmental Panel on Climate
Change) required 3 different time zones for the GWP values namely 20 years, 100 years and 500 years.
Among these values, GWP20 has been a reference measurement for the short evaluations and decisions.
For potential analyses usually GWP100 is used. GWP500 is a valid measurement in the assessment of long
term scenario. Table 1 includes the GWP20 and GWP 100 values and distribution of HFC gases per usage
areas which are becoming dominant in the sector with respect to the air conditioning and cooling systems.

Table.1 The most commonly used HFCs

HFC- Applications 20 Year 100 Year

Refrigerant GWP GWP
R-23 Low temperature refrigerant 12000 14800

R-32 Blend component of refrigerants 2330 675

R-125 Blend component of refrigerants 6350 3500

R-134a Refrigerant in domestic refrigerators, mobile 3830 1430

airconditioning,stationary air-conditioning, blend
component of refrigerants, foam blowing agent, aerosol propellant
R-143a Blend component of refrigerants 5890 4470

R-152a Blend component of refrigerants, foam blowing agent, possible 437 124
future refrigerant

R-227ea Refrigerant 5310 3220

R-245fa Foam blowing agent Possible future refrigerant 3380 1030

R-365mfc Foam blowing agent Possible future refrigerant 2520 794

R-404A Refrigerant blend: a leading alternative to HCFC-22 in air- 6010 3922

conditioning and commercial cooling

R-410A Refrigerant blend: a leading alternative to HCFC-22 in air- 4340 2088

conditioning, transport refrigeration

R-407C Refrigerant blend: a leading retrofit alternative to HCFC-22 in air- 4115 1774
conditioning, transport refrigeration
R-507 Refrigerant blend: a leading alternative to 5700 3300
HCFC-22 in air-conditioning and commercial cooling

Today, refrigerant solvents, foams and fire extinguishers are defined as the most important items
strengthening the global warming effect and causing ozone consumption (EPA,2004). The global warming
effects and distribution of these items are given in Fig.2.

Refrigerants
160 Solvents
140 Aerosols

120 Foams
Fire extinguishers
100
MMTCE

80
60
40
20
0
2005 2010 2015* 2020*
Years

Fig.2. Items with high ozone consumption and global warming effect (EPA,2004)

93
In the production processes of these items with heavy HFC use, PFCs (perfluorocarbons) and SF6s
(sulphur hexafluorides) are heavily used. Mainly the GWP effects of these gases are more effective than the
CO2 effects and the atmosphere lives of these gases reach to thousand years. Among them, the effects of
the refrigerant reached to 80% with 94,5 MMTCE (million metric tons of carbon equivalent) in 2010. The
estimates indicate that this effect will continue with the same rates also in 2020. These gases are followed
by foams by around 12,27% with 14,44 MMTCE.

In the sector based emission calculation, different models are preferred in many countries. For instance, the
Vintanging model developed by the energy environmental agency where data are collected from sectors
based on the simulation of the ozone depletion values of the refrigerants in the USA are used is found in the
European Union by dividing the gross local product cost of each country to the individual consumption of
European Union member countries. In other countries, ozone consumption estimates of the countries from
the United Nations Environment Program (UNEP) are used.

In scenarios created for the projection estimates, economic growth ratio, ozone consumption ratio and
sector date of the country are used together with the ozone consumption values stated in the Montreal
protocol or UNEP. For the calculation of the emission effects of the cooling systems in the sector, there are
two categories namely the emission during the equipment life time and emissions released in scrap status.
In the first approach, the leakage ratios and service processes of the fluids are taken into consideration. In
the second approach, the emissions released as a result of exposure of system equipment are calculated.

EMISSION ANALYSES FOR COOLING SYSTEMS

For the emission analysis of systems, usually the TEWI system where a cooling system is evaluated from
an integral point of view and the Potential Emissions of System Components approach where each
system component is evaluated from production to scrap process are taken into consideration. This process
defined by IPPC is an evaluation method which is taken separately for each system component. Emission
effect in the production process of this type of systems is an effect depending on the fluid transfer and has a
very low direct emission effect of 5-10%. During the life of the systems, emission effect, charge amounts,
repair and maintenance processes are effective. Emission effect for each element in systems during repair,
maintenance and operation process can be calculated as follows ((EPA,2004).

i k
EsJ  ( I a  I s ) QCJ i 1
i 1 (1)

Here,
Es : Emissions of the equipment serviced
la : Annual leakage ratio (normal operation process)
ls : Service leakage ratio (service process)
Qc: Quality of the chemical used in the new equipment (total charge amount of the specific item)
j: emission year
i: life process index
k: lifetime process

Scrap process of an element is also a cause of emission in the cooling system. Calculation of the emission
defined as scrap emission is based on the assumption that the scrap element releases a certain amount of
existing fluid inside to the atmosphere. Scrap emission;

Ed J  QCJ k 1 [1  (rm * rc )]
(2)
Here
Ed = Emission of the scrap element.
Qc = Quality of the chemical used in the new equipment (total charge amount of the specific item)
rm = The remaining fluid amount (depending on the total fluid)
rc = The recycle ratio of the remaining fluid.

Total emission potential for a cooling cycle is defined as the total of the scrap emission potential and
emission potential during lifetime. Total emission potential;

94
E J  E sJ  Ed J
(3)

The energy consumption of the system is also effective according to the preferred fluid. The emission effect
in this process is defined as the indirect effect based on energy. It is noticed that emission effects of the
refrigerant groups are evaluated in many organizations including the montreal protocol and that the
projections of these items are conducted (EPA, 2004). However, there have been different approaches for
the evaluation of the emission potentials of the cooling systems by taking lifetime processes into
consideration. Among them the most popular one is the TEWI method.

TEWI CONCEPT

At the meeting in Copenhagen, a concept for evaluating the total sum of global warming caused by
refrigeration systems, namely through refrigerant emissions and also through CO2 emissions from the fossil
fuel power plant which produces the energy for the refrigeration system was presented. It is called the Total
Equivalent Warming Impact, TEWI, and was developed by the US Department of Energy and Oak Ridge
National Laboratories (Horst, 2000).

An appropriate method of evaluating any refrigerant is the TEWI value, as defined in EN 378:2000. This is
an internationally accepted formula that may be used to directly compare the relative effect of different
systems. It represents the environmental effects, based on the global warming potential of the refrigerant,
and energy consumption for the plant over its lifetime. The higher the TEWI value(kgCO2/year), the poorer
the system in terms of its overall impact on the environment:

TEWI  [ GWP . L . n ]  [GWP . m . (1   re cov ery )]  [ n . E annual.  ]

Leakagelosses Recov ery losses Energy consumption (4)
Direct GWP Indirect GWP
Where TEWI is the total equivalent warming impact, GWP is the global warming potential, L is the amount
of leakage, n is the system operating time, m is the refrigerant charge, αrecovery is recovery or recycling
factor, Eannual is the annual energy consumption, β is CO2 emission produced by power
generation[(Moore,2005).

The global warming effects of different substances are compared with carbon dioxide using an index known
as the global warming potential or GWP. The GWP of a substance of the global warming impact due to the
emission of 1 kg of that substance is divided by the global warming impact of 1 kg of carbon dioxide, both
effects are measured over a 100 year time span (Dupont,2005).

The entire climatically relevant impact by the refrigeration process is composed of direct and indirect
emission of greenhouse gases. The direct impact includes the emission of working substances by the
greenhouse effect due to leakages, maintenance and losses during disposal. The indirect impact is related
to the CO2 emissions, particularly due to the generation of electrical energy in power stations. But also
those indirect CO2 emissions which are created by the energy consumption needed by supplying material,
material processing, maintenance, and waste disposal have been taken into consideration (Eurammon,
1996). In TEWI analysis, the CO2 emission per kWh of generated energy is taken into consideration for the
calculation of indirect TEWI contribution. The CO2 emission for the production of electricity is dependent
upon the mixture of energy carriers used in power plants. The burden on the Earth’s climate resulting from
the greenhouse effect is a global problem. The study is based on published data on CO 2 emissions from
power plants and utilizes the average value for the regions of North America, Japan and Europe. In this
study, the average value is 0.47 kg CO2/kWh (Hellmann et al., 1997).

RESULTS AND DISCUSSION

TEWI values for each widely used HFC fluids have been taken individually in the study based on the vapour
compressed cooling cycle taken as model. R-22 gas has been taken as reference fluid in the performance
evaluation of the HFC fluids. Basic operation parameters for a cooling system with vapour compression
considered under adiabatic conditions are accepted as given in Table 2.

95
Table. 2 Cycle parameters

Parameters Units Values

Evaporator Temp. (ºC) -10/0
Evaparator Superheating (ºC) 0
Condenser Temp. (ºC) 45
Condenser Superheating (ºC) 0
Refrigerating capacity (kW) 1
Leakage rate (%) Min (%3) , Max. (%30)

In the study, the R-227 and R-365 refrigerants have different operation parameters than the other HFC
fluids. For these gases, the evaporator and condenser temperatures have been examined for the
temperatures between -10 ºC and 35 ºC. However, it is not given here considering the shortness of the
work. However, the mass flow for the refrigerants is 6.678 g/s for R-22 gas for unit kW while it varies
between 7 g/s and 12 g/s. While refrigerant charge of 0.22 kg per kW cooling for R-22 gas is sufficient,
these values for HFCs are not below 0.45 g/s. However, the fluid charge amount is an important parameter
for the TEWI values. Determination of these values is a variant for each system. The charge amount in the
system has been evaluated in proportion with the flow speed.

COP and exergy values of each refrigerant have been examined in the study for each cooling load. Then
the improvement potentials of each fluid have taken into consideration and the emission analyses have
been examined in detail. Figure 3 includes the COP distributions based on evaporator temperatures
between -10/0 0C of the R-22 gas together with the HFCs examined.

4
3,8
3,6 R-22
R-32
3,4
R-125
3,2
R-134a
COP

3
R143a
2,8
R-152a
2,6 R-404A
2,4 R-410A
2,2 R-407C
2 R507
-10 -8 -6 -4 -2 0
Evaporator tempeature °C

Fig.3. COP distributions of HFCs and R-22 refrigerant

COP values of the R152a are the highest among the HFCs in the study. The performances of other HFCs
are below the R-22 gas. The performances of the widely used R134a, R404a and r-410a refrigerants are
worse than the R-22 gas with values respectively 1.8%, 17.56% and 10.87%. In addition, the R-123 and R-
507 gases have the worst COP values with respectively 2.5 and 2.91. Exergetic performances of HFCs
have been examined separately and the results are given in Fig.4.

96
 0,65
R-22
0,6 R-32
Exergy efficiency

R-125
0,55
R-134a
R-143a
0,5
R-152a

0,45 R-404A
R-410A
0,4 R-407C
-10 -8 -6 -4 -2 0
R-507
Evaporator temperature °C

Fig.4. Exergy efficiencies of HFCs and R-22 refrigerant

R-152a gas is prominent among the exergy efficiencies of HFCs. The average exergy efficiency of 59.55%
of this gas has a 5.06% better performance than the R-22 gas. In addition, it is seen that the R-152a gas
has 5.06%, 21.44% and 14.46% better performance than the R-134a, R-404A and R-410A gases which
dominate the market. R-407C gas has a high performance after the R-152a and R-134a gases among
HFCs. Among them, R-125 gas has a performance of 46.53% and R-507 gas has a performance of
49.72%.

Like in the thermal processes, high entropy is the biggest source of irreversibility in processes. In this type
of processes, energy recycle is an important factor not only in efficient operation of the processes, but also
the decrease of energy costs and sera gas emissions. Therefore, improvement potentials of the HFCs were
examined together with the R-22 gas. Fig.5 includes the compressor consumption together with the
recycling potentials.

0,45
0,4
0,35
Comp. Power
0,3 (kW)
Parameters

0,25 IP (kW)
0,2
0,15 IP Rate (%)
0,1
0,05
0
R-22 R-32 R-125 R-134a R-143a R-152a R-404A R-410A R407C R-507
Refrigerants
Figure 5 Compressor power, Improvement potential and rates of HFCs and R-22 gas

General average (0.3494 kW) of power consumption for each kW cooling capacity of HFCs has 11.53%
worse performance when compared to the R-22 gas. When we consider the dominance of HFCs in the
cooling sector, general energy consumption has also increased accordingly. However, R-152a gas has the
best performance with 0.3033 kW for the -10 ºC /0 ºC evaporator temperature with respect to power
consumption among HFCs. This gas is followed by R-134a gas with 0.32 kW and R-407C gas with 0.33 kW.

Despite the ozone depletion in HFC fluids is zero, most of them have high emission potentials due to high
GWP value. The TEWI effects of these gases have been examined separately according to the minimum
and maximum values. The leakage ratios for the TEWI effects have been taken as 3% and 30%
respectively according to minimum and maximum values and the αrecovery values have been taken as 3%
and 30%. Depending on these assumptions, TEWI values of the refrigerants for GWP20 have been found
and the results have been given in Fig.6.

97
 60000
Leakage rate %3
50000 Leakage rate %30
TEWI (kgCO2/year)

40000

30000

20000

10000

0
R-22 R-32 R-125 R-134a R-143a R-152a R-404A R-410A R-407C R-507
Refrigerants

Fig.6. TEWI values of HFCs refrigerants and R-22 for GWP20

The average TEWI values of HFCs for the GWP20 potentials have been found as 34134.6 kgCO 2/year.
When this value is considered with R-22, it is a higher value with 47.92%. This increase is not valid for all
HFCs. Among them, only R-152a gas is below R-22 with its value of 19326.4 kgCO2/year. The R-125 and
R-507 gases have 1.83 and 2.31 times higher TEWI values compared to R-22. Similar analyses have been
done for GWP 100 and the results are given in Figure 7.

45000
Leakage rate 3%
40000
Leakage rate 33%
35000
TEWI (kgCO2/year)

30000

25000

20000

15000

10000

5000

0
R22 R-32 R-125 R-134a R-143a R-152a R-404A R-410A R407C R-507
Refrigerants

Fig.7. TEWI values of HFCs refrigerants and R-22 for GWP100

The TEWI values of HFCs according to GWP100 have been calculated as 23063.78 kgCO2/year for 3%
leakage ratio and as 28682.76 kgCO2/year for 30% leakage ratio. However, when the R-22 TEWI values
are taken into consideration, the average HFCs have 14.30% higher value for 3% leakage ratio and 23.12%
higher value for 30% leakage ratio. For GWP100, only the TEWI values of R-152a gas were found to be
3.99% and 18.94% lower for the leakage ratios of 3% and 30%. For the TEWI values of the R-404A gas
used heavily for commercial applications have been found to be 26.73% higher for 3% and 52.24% higher
for 30%.

CONCLUSIONS

COP and exergy efficiencies under unit cooling unit of the HFC refrigerants have been examined separately
together with the power consumptions, irreversibility based improvement potentials and ratios. Then the
CO2 emission equivalents based on TEWI values have been calculated. The results of the study are given
below:

98
a. Among HFCs, R-152a gas has the best performance with a COP value of 3.32 and exergy efficiency of
61.47. COP and exergy efficiencies of other HFCs have been below R-22.
b. R-152a gas is more prominent with respect to compressor power consumptions, irreversibility based
improvement potentials and ratios. Among all these parameters, the worst gas was found to be R-125.
c. For both leakage ratios, all gases except R-152a have higher TEWI values than the R-22 gas.
d. The emission equivalent of the leakage ratio in TEWI values has been found 41.04% higher for GWP20
and 24.36% higher for GWP100.

Based on these results, the thermodynamic properties of HFCs except R-152a gas have higher values with
respect to CO2 emission equivalent of TEWI. In fact, these refrigerants with only low ozone depletion value
should either be developed or alternative refrigerants like CO2 should be preferred. In addition, choice of
exergy analyses is important for determining borders of irreversibility in these systems and for defining the
irreversibilities based on the system performances.

References

Arthur D. Little, Inc. (ADL). 2002. Global comparative analysis of HFC and alternative technologies for
refrigeration, air conditioning, foam, solvent, aerosol propellant, and fire protection applications, Final
Report to the Alliance for Responsible Atmospheric Policy Available at
http://www.arap.org/adlittle/HFCstudy3-22JD.pdf.

Clodic D., Barrault S. Saba S. 2010, Global inventories of the worldwide fleets of refrigerating and
airconditioning equipment in order to determine refrigerant emissions. The 1990 to 2006 updating.
ADEME/ARMINES Agreement 0874C0147– Final report–Centre Energetıque Et Procedes, Paris France.

DUPONT. 2005. Dupont refrigerants the science of cool, Du Pont de Nemours (Deutschland) GmbH,
Germany, www.refrigerants.dupont.com

ECCJ, 2002, Energy Conservation Center, Japon, http://www.eccj-or-jp/summary/local0303/eng/03-01.html,

http://www.eccj-or-jp/databook/2002-2003e/03-04.html, http://www.eccj-or-jp/databook/2002-2003e/03-
05.html.

EPA, 2004 Determination of comparative HCFC and HFC emission profiles for the Foam and Refrigeration
sectors until 2015 Part 1 – Refrigerant emission profiles, Centre D’energetıque, Paris France,
www.epa.gov/ozone/snap/emissions/dowloands/FoamEmissionProfiles_Part1.pdf

EPA, 2004. Analysis of Costs to Abate International Ozone-Depleting Substance Substitute Emissions, EPA
430-R-04-006, Environmental Protection Agency, Washington, D.C.,USA.

Eurammon, 1996. Evaluation of the environmentally friendly refrigerant ammonia according to the TEWI
Concept, NH3 for ecologically friendly future, Frankfurt, Germany http://www.eurammon.com

Hellmann J., and Barthélemy P. 1997. AFEAS-TEWI III study: results and evaluation of alternative
refrigerants, Solvay Fluor und Derivate GmbH Technical Service-product Refrigerants, Bulletin no.
C/11.97/06/E, page 5.

Halozen H. 2007, HFCs or the Old Refrigerants - what is the best Choice?, XII European Conference
Technological Innovations in Air Conditioning and Refrigeration Industry June 8-9, Politecnico of Milano,
Milano, Italy
http://www.centrogalileo.net/nuovaPA/Articolitecnici/INGLESECONVEGNO/XIIConvegnoEnglish/ISESSION
E/Halozan_eng.pdf

Horst K., September 2000, Refrigerant use in Europe, ASHRAE journal, www.ashraejournal.org

IPCC/TEAP.2005, Refrigeration, Draft of IPCC/TEAP Special Report on Ozone and Climate, Chapter 4,
Intergovernmental Panel on Climate Change/Technology and Economic Assessment Panel.Cambbridge
Univercity, Newyork,USA.

Jubb C., Bennett M. Fraser P. 2002.Inventories and Projections of Ozone Depleting and Synthetic
Greenhouse Gases used in Montreal Protocol Industries, Burnbank Consulting Pty.Ltd. Australia.
www.environmental.gov.au/atmosphere/ozone/publications/pubs/burnbank-report.pdf

99
Moore D. 2005. A comparative method for evaluating industrial refrigerant systems, Sabroe Ltd. (revA).
Nov. 2005, www. Sabroe org.

Sogut M.Z., Bulgurcu H., Yalçın E. 2012, Emission inventory based on refrigerants in cooling sector
National Air Conditioning and cooling education Symposium 5-7 December - Balıkesir/Turkey

UNEP, 2006 Report of the Refrigeration, Air Conditioning And Heat Pumps Technical Options Committee
2006 Assessment, ISBN: 978-92-807-2822-4, Job No: OZO/0948/NA, UNEP Nairobi, Ozone Secretariat,
Kenya,2007.

100
 A COMPACT DESIGN OF WATER HEATING-HUMIDIFICATION PROCESSES FOR SOLAR
HDD SYSTEMS

M. Ghazal, U. Atikol, F. Egelioglu

Mechanical Engineering Department, Eastern Mediterranean University, Gazimagusa, North Cyprus
maher.ghazal@emu.edu.tr, ugur.atikol@emu.edu.tr, fuat.egelioglu@emu.edu.tr

ABSTRACT
This work investigates the efficiency of a pilot solar humidification unit that is suitable to use in a
Humidification Dehumidification Desalination system (HDD). Water heating and air humidification processes
have been merged to take place in a direct contact water solar thermal collection unit. Experiments are
conducted under the weather conditions of N. Cyprus. Air passes through the humidifier reaches saturation
at the outlet. The difference between temperatures and relative humidity at the inlet and outlet of the
humidifier were recorded for different air flow rates (8.2, 10.4, and 12.6 kg/hr), in the period between the 1st
2
and the 14th of December. The experiment shows that for an average solar radiation of 600W/m and a
2
mass flow rate of 12.6 kg/hr of air; the increase in average absolute humidity was 0.62kg/hr.m . As the flow
rate increases the vapor carried per hour increases but the efficiency of the humidification process slightly
drops.

INTRODUCTION
Different desalination methods have been suggested since last three decades due to growing demands for
fresh water. With the exponential growth in industry, techniques used in desalination have advanced too.
Among all the techniques known in present, Reverse Osmosis and Multi Stage Flash systems are mostly
the dominant ones in terms of large scale desalination plants. On the other hand, medium and small scale
desalination techniques are competing for the lead in terms of independency, reliability, and desalination
cost. The importance of the small scale plants manifests itself in the rural areas, countryside and remote
places.

This work focuses on one of the latest small scale desalination techniques utilizing HDD. HDD systems have
had limited use in industry, due to the poor process efficiency and energy losses accompanied with them,
until the main energy source used in this technique has shifted to a renewable one, namely Solar Energy.

Several works have been done on this kind of desalination using different setups and systems. Narayan et al
2010 provided a comprehensive review of solar-driven humidification–dehumidification desalination. These
systems varied between open and closed systems with heated air or heated water setups. The technique is
versatile but the main idea is to bring dry air in contact with water to be humidified in an evaporator before
passing it through a condenser in which air is dehumidified to give fresh water.

This study investigates the moisture carrying capacity of air bubbles passing through water. In general, this
setup allows air to pass through a pipe with several holes, which is immersed in water, then spurge into the
water through the holes. The air comes in the form of bubbles through the pipe holes then ascend vertically
under the buoyancy effect. Similar experimental investigation has been done by El-Agouz, in 2010, but on a
system that uses electrical energy in order to heat sea water. In this study the main energy source (i.e.
energy used for water heating) utilized was solar energy. The evaporation chamber is merged with the solar
collector in one unit i.e. water heating and humidification process take place in the same inclined chamber.

In fact this technique is widely used by chemists to diffuse gasses in liquids in a system called column
reactor. But these systems have a challenging problem called the coalescence of bubbles. Bubble
coalescence takes place even at the generation point or while rising to the surface. Upon coalescence of

101
bubbles the superficial gas–liquid contact area decreases and thus mass and heat transfer to the bubble
decrease (i.e., humidification decreases).

In their study of the lower region of the bubble column, Perry and Green, 2001, found that bubble
coalescence depends on the dispersion device. They claimed that in the bubble generator design, the typical
separation between orifice centers should range from 2.5 to 4 times the orifice diameter. This result has been
considered in the design of the bubble generator used in this study. The coalescence process has been
studied in the central region of the bubble column in different work by Camarasa et al., 1999 and Pohorecki,
2001, where bubble break-up and coalescence processes reach equilibrium. They have determined the
bubble mean diameter was responsible for the superficial area available for transport phenomena between
the gas phase and the liquid phase. It is evident from these studies that the smaller the bubble diameter, the
greater becomes the heat and mass transfer to it.

Mariano et al., 2007, studied the Effect of coalescence of bubbles at sieve plate on mass transfer revealing
that the results show that although coalescence decreases mass transfer rate from bubbles; the deformable
bubble generated can, in certain cases, balance the decrease in mass transfer rate due to the reduction in
superficial area.

Chen et al., 2011, investigated the coalescence of bubble pairs rising in a stagnant liquid numerically and
compared the results with the experimental ones. The numerical results indicated that the rising velocity of
the trailing bubble was larger than that of the leading bubble. They indicated that both of the leading bubble
and the trailing bubble rose faster than the single bubble. This is in fact another factor that affects the rate of
heat and mass transfer to the bubbles since faster rising bubbles decrease the interaction time and thus
decreases the amount of heat and mass transferred.

The main objective of the present study is to investigate experimentally the result of merging the
humidification and water heating processes of the HDD in an inclined bed. The effects of different
parameters (i.e., temperatures and different air mass flow rates) on the moisture carrying capacity of air
bubbles are also studied.

EXPERIMENTAL SETUP
The experimental setup presented in this study comes with a new conceptual solar humidifier by which two
stages of the HDD system are accomplished, namely, heating and humidification. The major components
which were used in the setup are; the humidifier bed, the compressor, and the data retrieving devices, see
Fig. 1. These components are briefly explained in the following sections.

The Humidifier Bed

The humidifier bed is made using Plexiglas for framing, galvanized metal sheet is used as an absorber, and
glazing from glass sheets for the aperture. The whole bed then was insulated and enclosed in a wooden box,
except the aperture side was glazed allowing sun light to penetrate through to the 2 mm absorber plate by
which the water contained in the cavity was heated by solar energy. The aperture is double glazed with 8
mm thick inner glass, in order to withstand the pressure of water underneath the humidifier, and 4 mm thick
outer glass having width of 50 cm and length of 100 cm. A half inch diameter and 60 cm long copper pipe
was inserted all the way through a hole, which was drilled from one side of the bed, to the opposite side of
the bed and was fixed internally. The pipe had 2 mm diameter holes drilled apart with 10 mm center-to-
center distances along 500 mm of its length, which is immersed in the bed. The purpose of this pipe is to
generate air bubbles into the water. The distance between the glazing and the absorber plate was 5mm. To
allow air to exit the system, one inch plexiglas pipe was attached to a hole in the upper part of the frame. The
whole bed was mounted on a frame and was tilted due south to face the sun.

102
Fig. 1. The humidification system.

Compressor and controllers

An electrically driven two-piston compressor was used to supply air to the system. The compressor was
equipped with a 200L of storage tank, air flow controller, and pressure controllers to control the air in the tank
as well as the inlet air to the system.

Data Acquisition and measuring devices

The data collected in this study was air flow rate, global solar radiation, water temperature, inlet and outlet air
temperatures and relative humidity. Solar intensity on the surface of the system was measured using a
Pyranometer coupled with a digital multimeter. The Eppley radiometer Pyranometer has the accuracy of ±0.5
2
for the range from 0 to 2800W/m . Water temperature was snapped using two thermocouples which were
connected to Xplorer GLX data logger. Inlet and outlet relative humidity and temperatures of the air were
acquired using an Omega Thermo Hygrometer RH411 with measurement range of 5 to 98% relative humidity
and 0 to 49ºC Temperature. The measurement accuracy of the hygrometer is ±0.8ºC for temperature and
±3% for relative humidity. The air flow rate was measured using a hot wire anemometer.

EXPERIMENTAL PROCEDURE AND DATA PROCESSING

Procedure
The main working principle is to bring unsaturated air (i.e. air bubbles) into contact with water to saturate the
air in a process called humidification. In this setup air bubbles travel freely through still water in the cavity to
ensure maximum contact surface area, thus greater mass transfer rate can be achieved.

Water was filled in the cavity to about two third in order to make room for the volume of the air bubbles. The
south facing bed allowed the solar energy to penetrate through the glazing to the black painted absorber
plate. The absorber in turn transferred the absorbed heat to the contained water and increased its
temperature. Compressed air was directed through the copper pipe and bubbles were generated in the
water. While ascending upwards, heat and mass transfer to the bubbles took place.

103
Data of the solar intensity, water temperature, and air relative humidity and temperatures at the inlet and
outlet of the evaporator were recorded hourly from 9:30 am to 3:30 pm in mid-winter for three different air
flow rates (8.2, 10.4, and12.6 kg/h.)

Fig. 2. Working principal of the humidifier.

Data processing
The absolute humidity calculation was made possible using the data snapped by the electronic Hygrometer.
The inlet and outlet relative humidity and temperatures of air were inserted to a computer-aided
thermodynamic table that gave all the thermodynamic properties of air including the most important of which,
the absolute humidity. Then the difference between the inlet and outlet absolute humidity was calculated to
give the increase in water vapor content in air through the process. Humidification Process Efficiency was
determined according to the following consideration.

Humidification Process Efficiency is considered as the ratio of actual increment in humidity to the maximum
possible increment in vapor content at the saturated temperature of air. And it is given by:

(1)

RESULTS AND DISCUSSION

Figure 3 shows the inlet and the outlet relative humidity of the air passing through the humidifier for different
air flow rates (i.e., 8.2, 10.4, and 12.6 kg/hr) on different days. Although the inlet relative humidity varied for
each day, the outlet relative humidity was almost 100% with slight decrease at flow rate of 12.6 kg/hr.

Fig.3. The inlet and outlet relative humidity of air for different flow rates.

104
Figure 4 shows the inlet and outlet temperatures of air as well as the water temperature. It can be seen that
there is a gap between the passing air temperature and water temperature for all of the flow rates. As the
flow rate increases the difference between the inlet and outlet temperatures decreases. This behavior comes
as a result of two reasons; the first one is the chilling effect of air on water due to higher flow rate and thus
higher heat transfer rate. The second reason is that the energy of vaporization, which comes mainly from
water, needed for phase change increases as it is directly proportional to the amount of vaporized water, and
thus to the air flow rate.

Fig.4. Effect of air flow rate on the temperature differences between the air inlet and outlet.

Figure 6 illustrates the average moisture content differences between the inlet and outlet air for each flow
rate (i.e. 8.2, 10.4, 12.6 kg/hr.) in addition to the average efficiency calculated using Eq. (1). It can be clearly
seen that mass transferred is greater for greater flow rates whereas efficiency of the humidification process
decreased slightly.

Fig.6. Effect of higher flow rates on the productivity and humidification efficiency.

CONCLUSIONS
This study has proposed a novel solar humidification unit for use in HDD. The effect of water temperature,
and flow rate on the productivity and efficiency of the system has been studied and the following points were
concluded:

1. Direct contact humidification is an effective technique to be used in solar HDD systems. Merging
water heating and air humidification processes in a compact design, results in efficient humidification
and efficient use of space for small scale HDD systems.
2. Although the system hasn’t reached high temperatures, due to the weather mid-winter, air comes out
from the humidifier as saturated.

105
 3. Although there was slight drop in the efficiency of the humidification process as the flow rate
increases, the average increment in moisture content per hour increased and thus the productivity
also increased.

NOMENCLATURE
T temperature, C°
RH relative humidity, %

Greek Letters
η efficiency of the humidification process, %

ω absolute humidity, kg water vapor/kg dry air

Subscripts
w water
a air
in inlet
out outlet
sat saturated, saturation

REFERENCES
Camarasa, E., Vial, C., Poncin, S., Wild, G., Midoux, N., Bouillard, J., 1999. Influence of coalescence
behaviour of the liquid and of gas sparging on hydrodynamics and bubble characteristics in a bubble column.
Chemical Engineering and Processing 38, 329–344

Chen R.H., W.X.Tian, G.H.Su, S.Z.Qiu, YukiIshiwatari, YoshiakiOka. Numerical investigation on coalescence
of bubble pairs rising in a stagnant liquid. Chemical Engineering Science 66 (2011) 5055–5063

El-Agouz S.A., 2010. A new process of desalination by air passing through seawater based on
humidification-dehumidification process. Energy 35: 5108-5114

Mariano Martin, Francisco J. Montes, Miguel A. Galan. Bubble coalescence at sieve plates: II. Effect of
coalescence on mass transfer. Superficial area versus bubble oscillations. Chemical Engineering Science 62
(2007) 1741 – 1752

Narayan G. Prakash, Mostafa H. Sharqawy, Edward K. Summers, John H. Lienhard, Syed M. Zubair b, M.A.
Antar. 2010. The potential of solar-driven humidification–dehumidification desalination for small-scale
decentralized water production. Renewable and Sustainable Energy Reviews 14:1187–1201

Perry, R.H., Green, D.W., 2001. Engineering manual of liquids 3 4a Edition. McGraw-Hill, Madrid.

Pohorecki, R., Moniuk,W., Zdrojkowski, A., Bielski, P., 2001. Hydrodynamics of a pilot plant bubble column
under elevated temperature and pressure. Chemical Engineering Science 56, 1167–1174.

106
PERFORMANCE EVALUATION OF THE BIO-CHAR HEAVY METAL REMOVAL PRODUCED
FROM TOMATO FACTORY WASTE
1 1 1 1 2
E. Önal , N. Özbay , A.Ş. Yargıç , R.Z. Yarbay Şahin and Ö. Gök
1
Chemical&Process Engineering Department, Faculty of Engineering, Bilecik Seyh Edebali University, Bilecik,
TURKEY
2
Chemical Engineering Department, Hakkari University, TURKEY.

eylem.onal@bilecik.edu.tr, nurgul.ozbay@bilecik.edu.tr, seyda.guler@bilecik.edu.tr,

zerrin.yarbay@bilecik.edu.tr, ozgulgok@hakkari.edu.tr

ABSTRACT
Bio-char is a carbon-enriched and porous material produced from a variety of biomass. When bio-char is
produced from biomass, approximately 50% of the carbon that the plants absorbed as CO 2 from the
atmosphere is, ‘’fixed’’ in the charcoal. Bio-char is similar in its appearance to charcoal and activated
carbon.

In this study tomato factory waste has been used for the production of the bio-char. Biomass with a mean
particle size was carbonized at 623 K in a furnace. Biomass and bio-char were characterized by using
elemental analyses Fourier Transform Infrared Spectroscopy (FTIR), and scanning electron microscope
(SEM) analysis.

The adsorption capacity of the bio-char produced with carbonization of tomato factory waste has been
evaluated with the Co (II) ion removal to investigate the effects of pH, amount of adsorbent, initial
concentration of the aqueous solution ,adsorption time and solution temperature. To describe the
equilibrium isotherms Langmuir and Freundlich models were applied. Pseudo-first order and pseudo-
second order kinetic models were used to find out the kinetic parameters and mechanism of adsorption
process with increasing adsorbent dosage from 1 to 10 g/l in the batch mode, the final heavy metal
concentrations have been reduced from 59 to 8 ppm and removal efficiencies have been increased from
60% to 82% respectively.

Experimental results showed that, tomato factory waste char seems to be an effective and alternative
adsorbent precursor for the removal of heavy metal ions from aqueous solutions due to its high adsorption
capacity, low cost and availability.

INTRODUCTION
The growth of human population, urbanization and industries contribute to environmental contamination by
hazardous metals, which is of very important and serious environmental concern for all living organisms,
especially for humans. Heavy metal ions, such as cobalt, accumulated through the food chain; even at low
concentrations, have damaging influences on human health. Cobalt has toxic effects like imparting
neurotoxicological disorder, genotoxicity, carcinogenicity, cardiomyopathy and bronchial asthma
(Lazarevica et al., 2012; Liu et al. 2013).

Up to now, numerous technologies have been developed for the removal of heavy metals from wastewater
such as chemical precipitation, ion exchange, liquid–liquid extraction, membrane filtration, biosorption, and
electro-coagulation. However, the application of these conventional methods have been impeded by some
inherent limitations, involving high capital and maintenance cost, expensive equipment, high sensitivity to
operational conditions, significant energy consumption, incomplete metal removal or sludge generation, etc.
Thus, adsorption has been received increasingly attention in recent years owing to its high efficiency, cost-
effectiveness and easy handling (Liu et al. 2013).

The application of various types of bio-char which is a carbon-enriched and porous material produced
from a variety of biomass, as low-cost adsorbents, for the control of environmental pollution in terms of
heavy metals removal from aqueous media is currently of great interest. Various adsorbents have been
2+
examined for removal of Co from water, but agricultural wastes like tomato wastes, which have relatively
high carbon and present porous structure, have not been explored fully. The present research is to explore
an inexpensive and efficient adsorbent from agricultural waste, specifically tomato waste, to replace the
existing commercial adsorbents such as activated carbon and nanosorbents. In the present study, bio-
char from tomato waste was characterized using FT-IR and SEM techniques and used as an adsorbent
2+
for Co ion removal. Batch mode studies were used to study the effect of pH, adsorbent dose, contact
time, initial concentration of metal, and solution temperature. In addition, the adsorption mechanism was
determined (Lakshmipathy and Sarada, 2013).

107
MATERIAL AND METHODS

Materials and preparation of Bio-char

Tomato waste was collected from Marmara Region. Tomato waste first washed with distilled water to
remove impurities like dust, air-dried at room temperature, ground in a ball-mill, sieved and stored in a
dark room Tomato waste with a mean particle size was carbonized at 623 K in a furnace. The bio-char
was classified as BC.
Stock solution of cobalt was prepared by dissolving a weighed amount of cobalt in distilled water. The
experimental test solutions were prepared by diluting the respective stock solution of cobalt with distilled
water and mixing them in the desired proportion.

Analytical techniques
Flame atomic absorption spectrophotometer (GBC 933 AA) was used to determine the concentration of
the metals. Fourier Transform Infrared (FTIR) spectroscopy (Perkin Elmer Spectrum 100) was used to
identify the functional groups and functional molecules that facilitate metal ion uptake in the bio-char.
Scanning electron microscopy (SEM) were performed with a Zeiss Supra 40VP microscope.

Adsorption studies
Batch adsorption studies were performed in 50-ml conical flasks by varying 2 to 8 for pH, 0.05 to 0.5 g for
adsorbent dose, contact time, 25 to 125 mg/L with an increase of 25 for initial concentration of metal, and
solution temperature. The suspension was separated using Whatmann filters, and supernatant was
subject to Atomic Absorption Spectrometer (AAS) to determine the residual metal concentration. For the
quantitative estimation of metals adsorbed on the surface of bio-char, the Eq. (1) was used:

(1)
−1
Where qe is the metal uptake (mg g ) by bio-char, Co and C1 are initial and final metal concentrations
−1
(mg L ), V is the solution volume (L), and M is the mass of the adsorbent (g) (Lakshmipathy and Sarada,
2013).

RESULTS AND DISCUSSION

Characterization of materials
The composition of the tomato waste and bio-char are given in Table 1. Bio-char has higher carbon
content than tomato waste which makes them more carbonaceous material. Nitrogen is increased at
levels ranging from 3.776 to 5.032 wt. %. Carbonization increased the nitrogen content.

Table 1. Characteristics of tomato waste and bio-char

Elemental analysis (%)
Sample Code C H N O*
Tomato Waste 49.689 7.430 3.776 39.105
BC 65.776 2.152 5.032 27.039
* By difference

Fourier transform infrared (FT-IR) technique was applied to investigate the differences on the
morphology after carbonization and the FT-IR spectra of tomato waste and bio-char are shown in Figure
−1
1. The FTIR spectrum of the tomato waste shows a broad peak at 3280 cm and the peak is attributed
to the O–H in hydroxyl functional group in alcohols, phenolic and carboxylic acids. Two peaks between
-1
2921 and 2851 cm were assigned to the asymmetric and symmetric tension modes of the aliphatic C-H.
-1
The peaks seen at nearly 1750 – 1640 cm indicates aromatic C=C and C=O tensions. The peak
intensities of the asymmetric and symmetric tension modes of the aliphatic C-H at around between 3000
-1
- 2900 cm were decreased according to increasing temperature from 350 to 550. The peak seen at
-1
nearly 1600 cm indicates aromatic ring. When FT-IR spectrum of the tomato waste and bio-char were
investigated, they found similar to each other but some changes are observed in chemical structure and
also a decrease in OH-, CH2- groups, and olefinic structure.

Surface morphology of the tomato waste and bio-char were investigated by scanning electron
microscope (SEM) and presented in Figure 2. Magnification used in SEM analysis for tomato waste and
BC was 10000x and 3000x, respectively. As seen in Figure 2, the absence of porosity is evident for
tomato waste and displays a platelet shape. However, the surface of the BC was including some pores of
varying sizes. The external surface of the BC has a little sponge-like morphology and consists of some
cavities.

108
 Fig.1. Infrared spectra of tomato waste (a), BC (b).

(a) (b)
Fig. 2. SEM images of; tomato waste (a), bio-char (b)

Effect of pH
Hydrogen ion concentration in the adsorption is considered as one of the most important parameters that
influence the adsorption behavior of metal ions in aqueous solutions. It affects the solubility of the metal
ions in the solution, replaces some of the positive ions found in the active sites and affects the degree of
ionization of the adsorbate during the reaction. In this study, the pH of the solution was adjusted by
adding 0.1 M HCl or 0.1 M NaOH and measured with a Thermo Scientific Orion 3 Star pH-meter. The
effect of initial pH on the biosorption of cobalt (II) ions onto BC was evaluated within the pH range of 2–8
with 0.25 g/50 ml adsorbent dosage. Studies beyond pH 8 were not attempted because precipitation of
the ions as hydroxides would be likely (Vimala and Das, 2009).

From Figure 3, it could be seen that cobalt adsorption increased along with the increase of pH of the
adsorbate solution. This pH dependency of adsorption efficiency could be explained by the functional
groups involved in metal uptake and metal chemistry. At low pH values, protons occupy most of the
adsorption sites on the adsorbent surface and less cobalt ion could be sorbed because of electric
repulsion with protons on adsorbent. When the pH values increased, adsorbent surfaces were more
negatively charged and the adsorption of metal ions (positive charge) increased. Decrease in adsorption
at higher pH (pH >4.0) is due to the formation of soluble hydroxylated complexes of the metal ions and
their competition with the active sites, as a consequence, the retention would decrease again.

109
 Fig. 3. Effect of pH on cobalt (II) adsorption capacity (initial cobalt concentration: 50 mg/L, contact time:
60 min, temperature: 293 K, adsorbent dosage: 0.25 g)

Effect of adsorbent dosage

The number of available sites and exchanging ions for adsorption depends upon the amount of
adsorbent in the adsorption process. The effect of adsorbent concentration on the metal removal
efficiency is presented in Figure 4. The cobalt uptake was found to increase with increasing
concentration of adsorbent up to an adsorbent dosage of 0.2 g/l. However, beyond this dosage the
increase in removal efficiency was slightly increased and this may be related to reduction in
concentration gradient (Vimala and Das, 2009).

Fig. 4. Effect of adsorbent dosage on cobalt (II) adsorption capacity (initial cobalt concentration: 100
mg/L, contact time: 60 min, temperature: 293 K, pH: 7)

Effect of initial Co concentration

The effect of initial cobalt concentration on the metal removal efficiency is given in Figure 5. The initial
concentration of the cobalt in the solution remarkably influenced the equilibrium uptake of cobalt. It was
noted that initial concentration increased the sorption of cobalt as is generally expected due to
equilibrium process. This increase in uptake capacity of the adsorbent with the increase in initial cobalt
concentrations is due to higher availability of cobalt ions for the adsorption. Moreover, higher initial
concentration provides increased driving force to overcome all mass transfer resistance of metal ions
between the aqueous and solid phase resulting in higher probability of collision between metal ions and
adsorbents. This also results in higher metal uptake (Vimala and Das, 2009).

110
Fig. 5. Effect of initial cobalt concentration on cobalt (II) adsorption capacity (pH: 7, contact time: 60 min,
temperature: 293 K, adsorbent dosage: 0.25 g)

Effect of contact time and solution temperature

In this study, the effect of temperature as a function of time was studied in an isothermal bath shaker at
293, 303, and 313 K. Adsorption of cobalt (II) by BC as a function of time is shown in Figure 6. In the
case of 293 K, 303 K and 313 K solution temperature, the removal efficiency almost increased. The
highest removal efficiency was reached at 313 K solution temperature. The improvement in adsorption
with temperature may be attributed to increase in the number of active surface sites available for
adsorption on activated carbon. In addition, the decrease in the thickness of the boundary layer
surroundings with temperature would be also a reason.

Fig. 6. Effect of contact time and solution temperature on cobalt (II) adsorption capacities (pH: 7, initial
cobalt concentration: 50 mg/L, adsorbent dosage: 0.25 g)

Adsorption isotherms
The analysis of equilibrium data to construct adsorption isotherms is usually important for design of
adsorption systems. Adsorption isotherms express the mathematical relationship between the quantity of
adsorbate and equilibrium concentration of adsorbate remaining in the solution at a constant
temperature. The most common types of isotherms are Langmuir and Freundlich models. The Langmuir
isotherm model is applicable to monolayer adsorptions and can be expressed as Eq. (2):

111
 (2)

Where Ce is concentration of metal ions in solution at equilibrium (mg/L), q e is the amount of metal ions
at equilibrium in unit mass of adsorbent (mg/g), qmax and b are the Langmuir coefficient related to
adsorption capacity (mg/g) and adsorption energy (L/mg), respectively. The Langmuir isotherm
parameters qmax and b were determined from slope (1/qmax) and intercept (1/bqmax) of the plot of Ce/qe vs.
Ce (Fig. 11), respectively. The empirical Freundlich isotherm is based on adsorption on a heterogeneous
surface and express by the Eq. (3):

(3)
−1
The sorption parameters are as follows: qm is the maximum adsorption capacity (mmol g ), KL the
3 −1
Langmuir constant related to energy of sorption (dm mmol ), KF and n are the Freundlich coefficient
related to adsorption capacity and intensity, respectively. The values of Freundlich constants n and K F
were obtained from the slope (1/n) and intercept (ln KF) of the plot of ln qe vs. ln Ce (Fig. 12), respectively
(Gupta et al., 2012).

The plots give straight lines, indicating that the adsorption of Co(II) onto BC followed the Langmuir
isotherm model. The Langmuir and Freundlich isotherms constants and regression coefficients are listed
2
in Table 2. For Freundlich isotherm, the R values were all higher than 0.97, and it indicated monolayer
coverage by Co(II) ions of the BC and a homogenous distribution of the active sites on the surface of the
BC.

Table 2. Isotherm constants for the adsorption of cobalt onto BC

Isotherms Parameters
Langmuir
qm (mg/g) 166.67
KL (L/mg) 0.02927
2
R 0.952

Freundlich
1/n
KF ((mg/g) (L/mg) ) 1.3647
n 1.40
2
R 0.972

Adsorption kinetics
In order to investigate the mechanism of sorption, two kinetic models were used to test the experimental
data: the pseudo-first-order equation proposed by Lagergren, and the pseudo-second-order kinetic
model proposed by Ho and McKay. The respective linear forms of the equations are given in Eq. (4) and
Eq. (5):

(4)

(5)
2+
where t (min) is the contact time, qe and qt (mmol/g) are the amounts of sorbed Co at equilibrium and
−1
time t, respectively, and k 1 (min ), and k2 (g/mmol. min) are the rate constants of the pseudo-first-order,
and pseudo-second-order model, respectively (Lazarevica et al., 2012).
2+
As seen from Table 3, the sorption data of Co can be satisfactorily described by the pseudo-first-order
equation, since values of the correlation coefficient is close to 1. Cobalt adsorption onto BC consisted of
chemical adsorption due to the fact that pseudo-first order kinetic model suggested that adsorption
process involved chemisorption mechanism.

112
 Table 3. Kinetic parameters for the adsorption of cobalt onto BC

Pseudo-first order kinetic model Pseudo-second order kinetic model

T
k1 qe,cal k2 qe,cal
(K) R2 R2
(min-1) (mg/g) (g/(mg∙min)) (mg/g)
-3
293 4.606x10 4.8083 0.896 0.0126 8.547 0.998
303 0.0115 2.1478 0.445 0.01 8.9285 0.988
313 0.0184 1.555 0.924 0.0249 8.6206 0.998

CONCLUSIONS
In the present work, the cobalt adsorption characteristics of bio-char obtained from tomato waste was
examined. Batch sorption experiments were conducted to examine the efficiency of the bio-char on cobalt
binding. The equilibrium data were analyzed using Langmuir, and Freundlich isotherm models. The
Freundlich model correlated the experimental data the best. Kinetic evaluations showed that the cobalt
adsorption process follows the pseudo-first order kinetic model. Overall, this study showed that bio-char
with high cobalt removal capacity could be prepared from tomato waste through simple treatment methods
at low temperature such as 623 K.

NOMENCLATURE

FTIR Fourier Transform Infrared Spectroscopy

SEM Scanning Electron Microscope Analysis
Co initial metal concentration, mg/L
C1 final metal concentration, mg/L
Ce concentration of metal ions in solution at equilibrium, mg/L
V solution volume, L
M mass of the adsorbent, g
t contact time, min
qe amount of metal ions at equilibrium in unit mass of adsorbent, mg/g
qt amounts of sorbed Co2+ at time t, mmol/g
qmax Langmuir coefficient related to adsorption capacity
b Langmuir coefficient related to adsorption energy, L/mg
RL dimensionless constant separation factor
k1 rate constant of the pseudo-first-order, 1/min
k2 rate constant of the pseudo-second-order model, g/mmol.min

REFERENCES

Lazarevica, S., I. Jankovic-Castvana, B. Potkonjak, D.Janackovic, R.Petrovica, 2012. Removal of Co2+ ions
from aqueous solutions using iron-functionalized sepiolite, Chemical Engineering and Processing 55: 40– 47.

Liu, Y., M. Chen, Y. Hao. 2013. Study on the adsorption of Cu(II) by EDTA functionalized Fe3O4 magnetic
nano-particles, Chemical Engineering Journal 218: 46–54.

Lakshmipathy, R., N.C. Sarada. 2013. Application of watermelon rind as sorbent for removal of nickel and
cobalt from aqueous solution, International Journal of Mineral Processing, article in press.

Vimala, R. and N. Das. 2009. Biosorption of cadmium (II) and lead (II) from aqueous solutions using
mushrooms: A comparative study, Journal of Hazardous Materials 168: 376–382.

Gupta, N., A.K. Kushwaha, M.C. Chattopadhyaya. 2012. Adsorptive removal of Pb2+, Co2+ and Ni2+ by
hydroxyapatite/chitosan composite from aqueous solution, Journal of the Taiwan Institute of Chemical
Engineers 43: 125–131.

113
 EFFICIENT ANIONIC DYES ADSORPTION ON ACTIVATED ALGERIAN CLAYS
a b c
S. Zen , F.Z. El Berrichi , A. Maoui
aLaboratoire de chimie appliquée, Université de 8mai45, BP 401, Guelma, Algeria
b
Laboratoire de chimie physique, Université de 8mai45, BP 401, Guelma, Algeria
cLaboratoire de génie civil et hydraulique, Université de 8 mai 45, BP 401, Guelma, Algeria
email address: *zora172001@yahoo.fr

ABSTRACT

The discharge of the tanning industry constitutes enormous nuisances for the environment and exactly for the
human health, in particular the various dyes which are used excessively to improve the dye industry.
Therefore, the use of local clays in the treatment of the discharge has an important economic and
environmental interest and with the aim for value the local’s clays from Algerian East, we have selected the
Djebel Debagh «DD3» and Tamazert “AT” clays low-cost materials abundant in highly weathered soils and .
Adsorption experiments were accomplished using tannery dyes namely: coriacide bordeau 3B, derma blue
R67 and the coriacide brown 3J adsorbed on activated kaolin by acid treatment. The adsorption capacity of
all three dyes exceeds 90 % observed after 40 to 80 min. Natural and treated clay samples were
characterized by chemical analyses, powder X-ray diffraction, FTIR. Finally, the obtained results confirmed
the applicability of local clays as an efficient adsorbent for anionic dyes and from dyes wastewater.

Keywords - « DD3 » and “AT” clays, Activation, Adsorption Kinetics, Anionic dyes

1. INTRODUCTION

Textile and tannery effluents are two of the most polluting industrial wastes, and the problems of treatment
and disposal of such wastes require much attention. The presence of dyes in aqueous effluent such as in
river stream can be noticed easily because dyes are colored and highly visible. The discharge of dyes directly
into aqueous effluent can endanger living organism and its aquatic ecology because most dyes are toxic.
They have negative environmental effects, causing oxygen impoverishment and light transmission
attenuation in aquatic ecosystems affecting both fauna and flora [Maes, 1994]. It is difficult to remove the
dyes from the effluents, since most dyes are stable to light and heat and are not biodegradable [Robinson et
al.,2001]. Several conventional methods for the treatment of effluents containing dyes are available [Akar et
al.2006], and the most efficient one is the adsorption process, because it is simple in terms of operation and
can remove the contaminant even at very low concentration. However, the choice of the adsorbent is based
on economical and practical reasons and clay minerals are natural materials, due to their low cost and better
adsorption capability, appears to be a more convenient route for removing organic pollutants and of heavy
metal ions [Debey et al.,2009; Gupta et al.,2009; Liu et al.,2007 and Adebowale et al.,2005]. Acid treatments
of clay consist to modify its structure by changing the properties of surface, porosity and acidity. The results
obtained through acid activation vary according to the type of clay used’ its nature’ the acid concentration’ the
temperature and the time of treatment [Srasra,1987]. Acid activated clay is widely used in industry as a
bleaching and decontaminating agent “adsorbent” catalyst and catalyst support [Belver et al.,2002; Teng et
al.,2006]. Therefore, the aim of this study is to investigate the adsorption of specific anionic dyes, widely used
in the tannery, such as: derma blue R67, coriacide brown 3J and coriacide bordeau 3B on Algerian clays
from Djebel Debagh”DD3” and EL Milia “AT” in which Algeria possesses estimated at millions of tons, in the
East, that need to be valorized. To explore the feasibility of this clay, Bentonite was chosen to test the
adsorption capacity to release the anionic dyes after treatment.

2. EXPERIMENTAL

2.1 Materials
Two commercial clays, DD3 and AT, were chosen to represent abundant raw materials without special degree
of purity. The kaolin (DD3), gray in color, was obtained from the region of Guelma (Djebel Debagh) in Algeria
and supplied by ETER (ceramic company) (Guelma, Algeria). The type of kaolin is much rarer other clay
minerals [Papke, 1971]. The second clay used is an Algerian kaolin “AT”, type kaolinite came from EL Milia
deposit “TAMAZERT” from Jijel region (Algeria), and supplied by ETER (Algeria). In order to obtain the acid-
activated clays (DD3) and AT, the acid treatment of was carried out with 0.1N H2SO4 acid [Benguella et
al.,2009]. The cation-exchange capacity (CEC) was measured in order to evaluate the potential use of these
clays for adsorption. The exchange was followed by the ammonium acetate method with a concentration of
-3 -1
2.0 mol dm at pH 8.0, giving 15meq.g for both clays: DD3 and AT.

114
The chemical composition in mass % and loss on ignition (at 1000°C) of natural AT and DD3 is shown in
Table 1.

Table 1. Chemical analysis of natural Tamazert “AT” and DD3 clays

Component DD3 AT
Content mass%
SiO2 41.97 71.10
Al2O3 38.00 18.6
Fe2O3 0.12 0.58 à 1.22
MgO 0.07 0.39
CaO 0.20 0.02
Na2O - 0.22
K2O - 0.82 à 206
NO2 - 0.37
TiO2 - -
BaO - 0.04
SO3 0.75 -
MnO 1.34 -
LOI 16.80 5.68
LOI: Loss on ignition at 1000°C

2.2 Characterisation
Mineralogical compositions of representative clays samples were determined by XRD using air-dried. XRD
patterns of DD3 and AT clays (before and after activation) were collected on a X-Pert Propanalytical
diffractometer using Ni filtered Cu-Kα radiation (λ=1.5406 Å, 30 kV, 30 mA and automatic monochromator).
The diffractograms were recorded in the range 2 = 6–65°. The scanning speed was 1°/min. The FTIR
-1
spectra of natural and treated AT and DD3 with H2SO4 acid were obtained in the region 4000–500 cm using
FT-IR spectrometer, type Perkin Elmer Spectrum one model, at room temperature dispersed in KBr discs.

2.3 Dyes solutions

The reactive dyes used as adsorbates are bifunctional dyes, they were provided by the Stahl Iberica of Spain
and they were simulated by aqueous solutions of organic anionic dyes industrially used in tanning industry,
namely: coriacide bordeau 3B, derma blue R67 and coriacide brown 3J. Synthetic test dye solution was
prepared by dissolving an accurately weighed amount of dye (1 g/L) in distilled water and subsequently
diluted to required concentrations. Their chemical formulas and complete structure diagrams are not
available.
On a Photolab Spektral WTW UV–vis spectrophotometer, the coriacide bordeau 3B, derma blue R67 and
coriacide brown 3J absorbed at 510, 610 and 430 nm, respectively. Intrinsic pH values of 6.1, 5.6 and 6.3
were measured at 25 °C with a Consort C 831 potentiometer. These pH values did not vary immediately after
addition of clay but only after few minutes once dye adsorption started.

2.4 Adsorption kinetics

In addition to Djebel Debagh “DD3” and Tamzert “AT” clays, the bentonite was chosen to test the adsorption
capacity to release the anionic dyes after treatment.
The adsorption process was conducted by adding a known amount of activated clays Tamazert “AT” or
Djebel Debagh”DD3” into 500 mL of 100 mg/L dye solution at a constant stirring speed of 450 rpm. Dye
adsorption kinetics was investigated at 20 °C and natural pH=4 for 200min.
Two milliliters of samples were drawn at suitable time intervals. The samples were then centrifuged for 15min
at 5000rpm and the left out concentration in the supernatant solution were analysed through UV visible at
maximum wavelength.
It was investigated the effect of clay amount (2g ,6g) to the removal rate of coriacide bordeau 3B, derma blue
R67 and coriacide brown 3J dyes on AT, DD3 and bentonite clays in the experiments.

115
The amount (mg) of dye adsorbed per unit weight of calys: Djebel Debagh”DD3”, Tamazert “AT” and
−1
bentonite at time t, Qt (mg. g ) and the percentage removal (P %) were determined by the following Eq. (1)
and Eq. (2):

Qt = (C0-Ct) V/m (1)

P% = 100(Co − Ct)/Ci (2)
Where Qt +3is the dye concentration on adsorbent at any time t (mg/g activated clay), C 0 and Ct (mg/L) are
the initial and liquid phase concentrations of dye adsorbed at time t respectively.
V is the volume of dye solution (L) and m is the mass of activated used (g).

3. RESULTS AND DISCUSSION

3.1 Characterisation of adsorbents

The interlayer spacing or d001 spacing is measured from the top of the corresponding Si tetrahedral silica
sheet (T) to the top of the Si tetrahedral sheet of the following layer [Maidan et al.,2006;Gungor et
al.,2001].The X-ray diffractograms of the AT and DD3 clays, before and after acid treatment, are given in Fig.
1. The AT and DD3 clays yielded, 7 peaks in the range of 6–40° (2). The basal spacing is 7.26 Å and 7.13 Å,
respectively for kaolinite (K) in DD3 and AT. The XRD pattern (Fig. 1.a) showed sharp reflections at d=4.42Å;
3.62 Å (kaolinite) and the major impurities are quartz (reflection at 3.96 Å), calcite (reflection at 2.45 Å and
2.02 Å). Natural AT shows (Fig. 1.c) reflections at 4.24 Å, 3.33 Å attributed to kaolinite and quartz, and calcite,
as impurities, with reflections at 2.56 Å and 1.98 Å, respectively. Identical XRD patterns of clays (Fig. 1.b and
d) indicate the acid treatment does not affect intensely the structure of the components and shows that the d
(001) band of AT and DD3 shifts to 7.23 Å and 7.33 Å, respectively, showing an increase of the basal
distance. Furthermore, acid treatment with H2SO4 0.1N, dissolves major impurities such as calcite, quartz
and dolomite for both materials.

Fig. 1. XRD patterns of the natural and activated clays; (a and c): natural DD3 and AT; (b and d): activated DD3 and AT: K= kaolinite and impurity
phases (C= Calcite, D= Dolomite and Q= Quartz) are shown

The infrared Spectroscopy constitutes for the mineralogists a tool for characterization of the crystallinity of
clays by the observation of the relative intensities of the bands of vibrations of hydroxyls of structure. The
FTIR spectrums of natural and activated clays are shown in Fig. 2.
As we can see, all clays natural and activated exhibit two moderately intense bands between 3620.64 and
-1
3750cm , which might be ascribed to the stretching frequencies of the OH functional groups of co-ordination
-1
water and of the hydration OH stretching. The absorption peaks between 1620 and 1630.53cm can be taken
as both due to OH stretching vibration and (H2O) deformation. The rest of the bands between 450 and 1095
-1 3+ 3+ 2+
cm in natural clays are due to stretching vibration of Si–O, Si–O–Si, OH attached to (Al , Fe , and Mg )
groups, and the silica quartz impurities. After acid treatment, a significant difference is not observed between

116
original and activated AT and DD3. Under these mild conditions, the FTIR curves showed that a weak
destruction of the layers and interlayer space of activated clay was carried [Salem et al.,2009].

Fig. 2. Infrared Spectra of naturel and activated Tamazert “AT” and Djebel Debagh”DD3” clays

3.2. Kinetics of adsorption

3.2.1 Effect of contact time
The influence of contact time on removal of blue derma R67, coriacide bordeau 3B and coriacide brown 3J
by 6g of activated Tamazert clay ”AT”, kaolin Djebel Debagh (DD3) or bentonite at pH 4 and 20°C with an
initial dye concentration 20mg/l are shown in Fig. 3. It is evident that all clays are efficient to adsorb dyes with
different efficiencies and the removal of dyes was rapid and strong in the initial stages of contact time (2 min)
followed by a slow increase until reaching equilibrium, due to the abundant availability of active sites on the
clay surface, and with the gradual occupancy of these sites, the sorption becomes less efficient. The shape
of the curves for DD3, AT and bentonite clays are similar. This indicates a monolayer formation of the dye on
the external surface [Al Ghouti et al.,2003]. The adsorption of the activated DD3 and AT is faster than that of
activated bentonite for three dyes except the coriacide bordeau 3B in activated AT. Moreover, the maximum
removal percentage of three dyes adsorbed (P%) is higher for the activated DD3 (98%) than for the activated
AT (94%) and more higher than for activated bentonite (82%). To reach equilibrium for anionic dye as blue
derma R67, it takes 40, 60 and 80 min for activated DD3, AT and bentonite, respectively. At the equilibrium,
the activated DD3, AT and bentonite fixes more blue derma R67 than the coriacide brown 3J and coriacide
bordeau 3B. The results of dye removal at equilibrium obtained are given in Table 2. Similar results have
been obtained on a textile dyes adsorbed by natural and activated bentonite [Benguella et al.,2009].

Fig. 3. Effect of contact time on the removal of anionic dyes on activated DD3 (a), AT (b) and bentonite (c).

Table 2. Results of % removal of dyes obtained at equilibrium

Anionic Dyes Activated The percentage
clays removal (P %)of
anionic dye at
equilibrium
Derma blue R67 98.03
Coriacide brown 3J DD3 93.83
Coriacide bordeau 3B 96.83
Derma blue R67 94,42
Coriacide brown 3J AT 89,55
Coriacide bordeau 3B 65,73
Derma blue R67 Bentonite 82.70
Coriacide brown 3J 76.0
Coriacide bordeau 3B 79.76

117
 3.2.2 Effect of mass of DD3 kaolin

The adsorption of three anionic dyes on the activated DD3 amount was studied on mass of kaolin of 2g in
constant dye solution of 20mg/l at pH = 4 and 20°C at constant stirring speed of 450 rpm. The effect of clay
amount on dye removal is shown in Fig. 4. Decreasing mass of activated DD3 induced an increasing in
removal percentage of adsorbed dyes with the maximum of 97% and it takes 40 min for coriacide bordeau 3B
and 75min for blue derma R67 and coriacide brown 3J, to reach equilibrium. From Fig. 4 and Table 3, it was
clear that the decrease of mass (2g) of DD3 influences slightly the percentage removal of dyes adsorbed and
its removal efficiency. This may be explained by the fact that a large adsorbent amount (6g) reduces the
saturation of the adsorption sites and correspondingly, the number of such sites per unit mass comes up
resulting in comparatively more adsorption at higher adsorbent amount. Furthermore, the results obtained in
dye adsorption experiments onto kaolinite showed that the adsorption capacity was higher with increasing
amount clay [Errais,2011; Dogan et al., 2009].

Fig. 4. Effect of clay amount on the removal of anionic dyes on activated kaolin “DD3”

Table 3. Results of % removal dyes obtained at equilibrium

Anionic Dyes Activated clays The percentage removal
(P %)of anionic dye at
equilibrium
Derma blue R67 96.81
Coriacide bordeau 3B DD3 96.87
Coriacide brown 3J 92.87

3.2.3 Effect of clay nature

Fig. 5 shows the effect of clay nature on the anionic adsorption dye at initial dye concentrations of 20mg/L at
pH=4 for derma blue as function of contact time on 6g of activated bentonite or DD3 or AT. In order to
characterise the adsorption capacity on kaolin of the anionic dye “blue derma R67” which was fixed more
than coriacide bordeau 3B and brown 3J, we have compared with activated bentonite. The removal of derma
blue R67 was rapid in the initial stages of contact time (Fig. 5) and gradually decreased with lapse of time
until equilibrium. The rapid adsorption observed during the first 5 min is probably due to the abundant
availability of active sites on the kaolinite surface, and with the gradual occupancy of these sites, the sorption
becomes less efficient. The time required to attain this state of equilibrium was termed as the equilibrium time
and the amount of dye adsorbed at the equilibrium time reflected the maximum dye adsorption capacity of the
adsorbent under these particular conditions [Hameed,2008]. The time necessary to reach this equilibrium is
about 40 min for DD3 with higher removal percentage (98%). However, it appears from Fig. 5 the rapid
adsorption of derma blue R67 on AT and bentonite clays was observed during the first 3 min and this is
followed by a strong increase of adsorption after 60 min for activated AT with higher removal percentage
(94%) than that of bentonite (82%) (Table 4). The shapes of the curves of activated DD3 and bentonite are
similar (Fig. 5). This indicates a monolayer formation of the dye on the external surface [Al Ghouti et
al.,2003]. Thus the main process involved in anionic derma blue dye adsorption on activated DD3 kaolin may
be the attraction on the broken edges of clay particles where charges become positive at acidic pH. That is

118
why although acidic pH conditions were found to enhance the dye removal. On the other hand, the anionic
dyes of the tannery industry used in this study are acids. The effect of clay nature on dye adsorption observed
in this study was explained by electrostatic interaction between kaolinite or bentonite and dye molecules. As
the pH of the system (dye +kaolinite) is acid (pH=4), knowing that the pH has a relatively high value for
kaolinite (2–4.6) [Schroth et al.,1997] , the number of negatively charged sites decreases and the number of
positively charged sites increases in activated DD3. Therefore, the extent of dyes adsorbed on kaolinite
tended with pH system, which can be attributed to the electrostatic attraction between the positively charged
surface and the negatively charged dye molecule [Dogan et al.,2009]. Also, lower adsorption of anionic dyes
+
on activated bentonite is because of the low presence of H ions competing with dye anions for the
adsorption sites.

Fig. 5. Effect of clay nature on the removal of derma blue R67 on activated ”DD3” , AT and bentonite

Table 4. Results of % removal dyes obtained at equilibrium

Anionic Activated The percentage

Dyes clays removal (P %)of anionic
dye at equilibrium
DD3 98.03
Derma AT 93.50
blue R67 Bentonite 82.70

4. CONCLUSION

Algerian clay, DD3 from Djebel Debagh and AT from El Milia region, activated by acid treatment can be
employed as an effective and promising low cost alternative adsorbents for the removal of anionic dyes from
tannery wastewater. A comparative study of adsorption of derma blue R67, coriacide brown 3J and coriacide
bordeau 3B evidences the high adsorption capacity of dye on the activated DD3 and AT than activated
bentonite. An increase in the initial dye concentration enhances the interaction between dyes and clays “DD3
and AT”, resulting in greater and fast adsorption capacity with the equilibrium was attained after 40 to 60min
for anionic dyes by comparison with activated bentonite, the adsorption capacity is low for all the three anionic
dyes and the time to reach equilibrium was 80 to 140min. The results showed that the adsorption anionic
dyes on activated DD3 and AT could be explained by the electrostatic attraction between the positively
charged surface and the negatively charged dye molecule in the acidic medium (natural pH=4). Furthermore,
the experiments conditions (acidic medium, activated adsorbent, low temperature) of kinetics adsorption
study were found to enhance the dye removal great adsorption efficiency.
Compared to standard bentonite clay, activated DD3 and AT have a good and high potential for adsorption
removal of anionic dye from aqueous solution

5. REFERENCES

[1] M.Maes. 1994. Paris. « Les résidus industriels, traitement valorisation et législation », Entreprise Moderne d'Edition
Technique et Documentation.

[2] T. Robinson, G. McMullan, R. Marchant, and P. Nigam. 2001. “Remediation of dyes in textile effluent: a critical review
on current treatment technologies with a proposed alternative”, Bioresource Technology 77: 247–255.

119
[3] T. Akar, T.A. Demir, I. Kiran, A. Ozcan, A.S. Ozcan, and S. Tunali. 2006.” Biosorption potential of Neurospora crassa
cells for decolorization of Acid Red 57 (AR57)dye”. Journal of Chemical Technology and Biotechnology 81: 1100–1106.

[4] S.P. Dubey, K. Gopal,and J.L. Bersillon. 2009. ” Utility of absorbents in the purification of drinking water: A review of
characterization, efficiency and safety evaluation of various absorbents”. Journal of Environmental Biology 30: 327–332.

[5] V.K. Gupta, N. Suha.2009. “Application of low-cost adsorbents for dye removal - A review », Journal of Environmental
Management 90: 2313–2342.

[6] P. Liu, L. Zhang. 2007. “Adsorption of dyes from aqueous solutions or suspensions with clay nano-adsorbents”.
Separation and Purification Technology 58: 32–39.

[7] K.O.Adebowale., I.E.Unuabonah, and B.I.Olu-Owolabi. 2005. “Adsorption of some heavy metal ions on Sulfate- and
Phosphate-modified Kaolin”. Applied Clay Science 29:145-148.

[8] E.Srasra. 1987. « Caractérisation minéralogique, propriété physico-chimique et application des argiles du gisement
Haidoudi ». Thèse de 3eme cycle, Fac. Sc. Tunis : 171.

[9] C.Belver, M.Angel Bañnares Munoz, and M.Vicente.2002. “Chemical Activation of a Kaolinite under Acid and
Alkaline Conditions”. Chemistry of Materials. 14: 2033–204.

[10] Teng, Min-Yu and Lin, Su-Hsia.2006. “Removal of Basic Dye from Water onto Pristine and HCl-Activated
Montmorillonite in Fixed Beds”, Desalination 194: 156–165.

[11] K.G.Papke. 1971. “Halloysite deposits in the Terraced Hills”, Washoe County, Nevada: Clays and Clay Minerals 19:
71-74.

[12] B. Benguella, A. Yacouta-Nour, 2009. “Adsorption of Bezanyl Red and Nylomine Green from aqueous solutions by
natural and acid-activated bentonite”, Desalination 235: 276–292.

[13] M. Majdan, S. Pikus, Z. Rzaczynska, M. Iwan, O.Maryuk, R. Kwiatkowski and H. Skrzypek. 2006. “Characteristics of
chabazite modified by hexadecyltrimethylammonium bromide and of its affinity toward chromates”. Journal of Molecular
Structure 791: 53–60.

[14] N. Güngör and S. Karaoglan. 2001” Interactions of polyacrylamide polymer with bentonite in aqueous systems”,
Materials Letters 48: 168–175.

[15] A.Salem and L.Karimi. 2009. “Physico-chemical variation in bentonite by sulfuric acid activation”. Korean Journal of
Chemical Engineering 26,no.4: 980-984.

[16] M.A. Al-Ghouti, M. Khraisheh, S.J. Allen, and M.N. Ahmad. 2003. “The removal of dyes from textile wastewater: a
study of the physical characteristics and adsorption mechanisms of diatomaceous earth”, Journal of Environmental
Management 69: 229–235.

[17] E. Errais. 2011. ”Réactivité de surface d’argiles naturelles, étude de l’adsorption des colorants anioniques ». Thèse
de doctorat. Université de Strasbourg :114.

[18] M. Dogan, M. Hamdi Karaoglu, M. Alkan.2009. “Adsorption kinetics of maxilon yellow 4 GL and maxilon red GRL
dyes on kaolinite”. Journal of Hazardous Materials 165: 1142–1151.

[19] B.H. Hameed. 2008. “Equilibrium and kinetic studies of methyl violet sorption by agricultural waste”, Journal of
Hazardous Materials 154: 204–212.

[20] B.K. Schroth, G. Sposito. 1997. “Surface charge properties of kaolinite”, Clays Clay Minerals 45 no.1: 85–91.

120
 EXERGOECONOMIC ANALYSIS OF A CASCADE ACTIVE MAGNETIC REGENERATIVE
REFRIGERATION SYSTEM
a b a
H. Ganjehsarabi , I. Dincer , A. Gungor
a
Department of Mechanical Engineering, Faculty of Engineering, Ege University,
TR-35100 Bornova, Izmir, Turkey
b
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology, 2000 Simcoe Street North,
“Oshawa, Ontario L1H 7K4, Canada
Email : hadi.ganjehsarabi@gmail.com, ibrahim.dincer@uoit.ca, ali.gungor@ege.edu.tr,

ABSTRACT
In this paper, an exergoeconomic analysis of cascade active magnetic regenerative refrigeration system
operating on a regenerative Brayton cycle is conducted with respect to various system design parameters. In
exergy analysis, a thermodynamic model is developed in order to determine exergy destruction rates and
calculate the exergy efficiency of the system. In the economic analysis, investment cost rates are calculated
with respect to equipment costs, which are determined by cost correlations for each system component, and
capital recovery factors. Thus, by combining the two analyses, an exergoeconomic model is created whereby
the exergy streams are identified and cost equations are allocated for each component. The results of both
exergetic and exergoeconomic analyses show that increasing the fluid mass flow rate decreases the exergy
efficiency, and increasing the specific exergetic cooling decreases the cost per unit of cooling.

Keywords: AMR cycle, MCE, Energy, Exergy, Exergoeconomic.

INTRODUCTION
Active magnetic regenerative (AMR) refrigeration systems are considered an environmentally benign solution that
could be used as an attractive alternative to vapor compression refrigeration systems (Engelbrecht et al., 2006,
2010). AMR system utilizes a varying magnetic field, magnetocaloric material (MCM) and heat transfer fluid to
facilitate heat transfer between reservoirs.
Barclay et al. (1982) presented and patented the Active Magnetic Regenerator Refrigeration (AMRR)
system in 1982. Barclay, (1983) showed that the possibility of achieving significantly larger temperature spans than
the adiabatic temperature rise of the magnetic material by using the magnetic material simultaneously as a heat
regenerating medium and as a refrigerating medium. Engelbrecht et al. (2006) developed a 1-D transient model of
a single regenerator sphere undergoing an idealized passive and active regenerative process and evaluated the
impact of the internal temperature gradients and their effect on the performance of the sphere in the context of a
regenerative cycle. Li et al. (2011) have numerically analyzed 2-dimentional porous media model of room
temperature AMRR and compared the 2-D model and 1-D model. Their work shows that the AMRR system can
lose 22% of cooling power caused by air convection and conduction in y-axis. Tura et al. (2012) evaluated the
performance of a permanent magnet magnetic refrigerator (PMMR) with gadolinium parallel plates. Their results
shows that, in order to obtain both temperature spans and cooling capacities comparable to those obtained with
packed spheres; significantly thinner plates and channels are needed. Li et al. (2006) studied the impact of the
physical properties of heat transfer fluids and operating frequency on the AMRR performance. They determined
that a liquid is more favorable than a gas for being used as heat transfer fluid in AMRR systems. Nellis et al.
(2006) developed a method in order to examine the effect of the entrained fluid heat capacity in the passive
regenerator. They found out that the entrained fluid heat capacity has a significant impact on the performance of a
passive regenerator. Rowe (2011a, 2011b) derived expressions which describe the thermodynamic quantities of an
idealized AMR cycle in steady state. Aprea et al. (2012) carried out a comparison between three different models
in order to simulate the thermal behavior of an AMR cycle. They found that for all the three models the COP
increases with the regenerator’s volume. Kitanovski et al. (2009) suggested that for having high temperature
differences cascade or regeneration stages have to be considered. Bjørk et al. (2011) proposed an experimental
data-based correlation for the total mass and the cost of magnet and magnetocaloric materials needed to
construct a magnetic refrigerator.
The primary objective of this study is to develop a practical model in order to perform both energy and
exergy analyses of a cascade AMR cycle. The specific objectives are given as follows:
 To simulate a cascade AMR refrigeration system in order to predict the performance of the overall system.
 To perform energy and exergy analyses of the system, including calculations of the exergy efficiency and
exergy destruction of the system.
 To conduct a parametric study of the effects of various selected design parameters on the COP and overall
exergy efficiency of the system.
 To develop an exergoeconomic model which determines the cost per unit of cooling with respect to specific
exergetic cooling power.

121
SYSTEM DESCRIPTION
The schematic of a cascade AMR system is shown in Fig.1, which consists of two Brayton cycles: (I) the higher-
temperature cycle and (II) the lower-temperature cycle. During magnetization, the two stage is moved into a
magnetic field, where the temperature of the magnetocaloric materials increases as a result of adiabatic
magnetization. In the isofield cooling process, the heat transfer fluid is blown into the second stage from the cold
end to the hot end. Subsequently, the fluid flows from the cold end of the first stage to the hot end, while
maintaining the magnetic field, absorbs the heat from the both stages and rejects the heat to the surrounding in
the hot heat exchanger. In the adiabatic demagnetization process, both stages are cooled by the demagnetization
effect. Finally, the fluid moves from the hot end of first stage to the cold end and travels from the hot end of the
second stage to the cold end, where it exits at a temperature lower than the cold heat exchanger and produces the
cooling power of the system.

Magnet
HHEX CHEX
TH Tblow

QH AMR QC

Trej TC

Magnet

Fig. 1: Schematic of an AMRR system (HHEX: hot heat exchanger, AMR: active magnetic regenerator, CHEX: cold heat exchanger).

The T-s diagram of idealized cascade AMR cycle is shown in Fig. 2. The first and second stage materials
are Gdx and Tb1-x alloys. The thermo-magnetic properties of the regenerator's compositions are calculated by using
the molecular field theory and the Debye approximation. The heat transfer fluid in the simulation is water–glycol
mixture (50% by weight) (for details see Aprea et al., 2011).

H high
T c
d
c
I
d H low
II b
a
b
a
s
Fig. 2: T-s diagram of theoretical (ideal) cascade AMR refrigeration system.

ANALYSES
In this study, the main aim is to address the thermodynamic performance of the cascade AMR refrigeration system
and to compare these results in terms of magnetic field and heat transfer fluid mass flow rate. In order to
accomplish this, the governing equations and numerical procedures used by Matlab must be outlined in order to
obtain sufficient data for the thermodynamic assessment. Finally, these procedures are used to gain insights on
the effectiveness of the cascade AMR refrigeration system through energy and exergy analyses. In the
development of heat transfer and thermodynamic models, several simplifying assumptions are made as follows:
 The effect of viscous dissipation on the energy balance on the fluid flow throughout the bed is neglected.
 The heat transfer fluid is incompressible (ρf=constant).
 The temperature and velocity profile of heat transfer fluid is uniform during the period of flow blowing.
 The axial heat conduction in the regenerator is neglected.
 Heat loss to the environment in regenerator is negligible.
 The properties of magnetic material (except for the specific heat) are constant in the regenerator.

122
After some simplification based on the above assumptions, the energy balance expression of the heat transfer
fluid (f) and the magnetic material (s) can be written as

T f T f
m f cp  mc p L  hA(Ts  T f ) (1)
t x
Ts
ms cs  hA(T f  Ts ) (2)
t
where the subscripts f and s refer to fluid and solid, respectively. Here, m is the mass flow rate, c is the heat
capacity per unit mass, L length of the regenerator and fluid and A is the heat transfer area.
The heat transfer coefficient is calculated by using a correlation given by Roshenow et al.(1985):
1
Nu f  2  1.1Re0.6 P Pr 3
P (3)

Energy analysis
The cooling capacity, heat rejection and the rate of magnetic work performed on the regenerator during cyclic
steady-state operation are the basic outputs of this model. The cooling capacity and heat rejection to the
environment are calculated by numerically through the integrals as follows:
2

Qrefrigeration   m()cP (TC  TfII (,0))d (4)


1

1  2  3  4

Qrejection   m()cP (Tf I ( ,L)  TH )d (5)

1  2  3

The pressure drop in the heat transfer fluid flow can be obtained by using the Ergun correlation (Kaviany, 1985):
P  1    f  1    f
2

 180   w inf  1.8   w 2inf (6)

x    dp    dp

The power required for moving the heat transfer fluid through the regenerator is written as
m(t)(PC  PH )
Wp  ( 2  4 ) (7)
pf
The coefficient of performance (COP) of the system can be found as
Q refrigeration
COP  (8)
WP  WM

Exergy analysis
The exergy analysis of the process determines the exergetic variables such as exergy destruction rate and exergy
efficiency for a thermodynamic evaluation. The irreversibility in the cascade AMR cycle can be calculated through
the exergy destruction rate as follows:
Ex des  Exin  Ex out  WM  WP  Ex q (9)

where Ex q is the exergetic cooling power as

TH
Ex q  Qc (  1) (10)
Tc
The specific exergy cooling of the system can be defined as the ratio of exergetic cooling power to the
multiplication of the applied magnetic field, H, and the total volume of refrigerant used, VMCM (Rowe, 2011).
Ex q
 (11)
HVMCM
The exergy efficiency for the cascade AMR refrigeration system can be calculated as follows:
Ex(Qrefrigeration )
ex  (12)
WP  WM

Exergoeconomic analysis
Economic model
Exergoeconomics is an exergy-based method that identifies and calculates the location, magnitude, causes and
costs of thermodynamic inefficiencies in an energy conversion system. The real inefficiencies in such a system are
the exergy destruction (associated with the entropy generation within a system or a component) and the exergy
loss (associated with the transfer to the environment of exergy that is not further being used in this or another
system) (Bejan, 1996, Dobrovicescu et al. 2011).The cost balances are generally formulated for each component
separately as

123
C
e
e,k  Cw,k  Cq,k   Cin,k  Zk
in
(13)

where
Cin  cin Ex in (14)
Cw  c w W (15)
Cq  cq Ex q (16)
Here, ci , ce , cw and cq donate average costs per unit of exergy in dollars per kW ($/kW).
The cost rate expression of equation (13) can be written as
 e e )k  cw Wk  cqEx q,k  (cin Exin )k  Zk
(c
e
Ex
in
(17)

where Z k is obtained by first calculating the capital investment and O&M costs associated with kth component and
then calculating of these costs per unit of time of system operation.
In the cost balance formulation Eq. (13), there is no cost term directly associated with the exergy
destruction of each component. Accordingly, the cost associated with the exergy destruction in a component or
process is a hidden cost. If one combines the exergy and exergoeconomic balances, one can obtain the following:
Ex F,k  Ex P,k  Ex D,k (18)
The cost rate associated with the exergy destruction within the component is defined as
CD, k  cF , k ExD, k (19)
where ExD , k is the exergy destruction of the component which is evaluated by using exergy balances.
A cost balance applied to the entire system shows that the sum of cost rates associated with all existing
exergy stream equals the sum of cost rates of all entering exergy streams plus the appropriate charges due to
capital investment and operating and maintenance expenses.
cq Ex q  celect WM  celect WP  Z (20)
Here Cq, cost per unit of cooling, Celect, is the unit cost of electricity, which is taken as 0.075 $/kWh.
On the economic side, the capital investment can be calculated with respect to the purchase cost of
equipment and capital recovery as well as maintenance factor over the number of operation hours per year as
shown below;
Z  CRF  
Zk  k (21)
N  3600
where Zk is the purchase equipment cost of the magnetic refrigerator components, N is the annual number of
operation hours for the unit, and  is the maintenance factor, which generally taken as 1.06. CRF is the capital
recovery factor which depends on the interest rate (i) and equipment life time (n), and is determined here as
follows:
i  (1  i) n
CRF  (22)
(1  i) n  1

RESULTS AND DISCUSSION

In this study, numerical modeling and a thermodynamic analysis of cascade active magnetic regenerative
refrigeration system through energy and exergy is conducted. In this regard, COP and exergy efficiency are
examined for performance assessment. The parametric studies considering variation of various design parameter
and some performance results are presented accordingly.
The parameters for model inputs that were used in order to carry out the analysis are summarized in Table
1. In order to investigate the influence of magnetic field on the cascade AMR cycle three various magnetic field
(1.5, 1.75 and 2 T) were chosen. The impact of the magnetic field on AMR cycle is investigated in more detail in
Bjørk et al. (2011). Fig. 3 shows the refrigeration capacity as a function of the fluid mass flow rate for a AMR
cascade system. As the fluid mass flow rate increases, the system is capable of producing refrigeration in direct
proportion to the mass flow rate so the refrigeration capacity increases.
It can be seen that an increase of the mass flow rate increases the refrigeration capacity of the system
until a certain value of mass flow rate, after where the cooling capacity of the system decreases. It is because of
the mass flow rate becoming too large that it overwhelms the magnetocaloric effect exhibited by the bed and,
thereafter, the refrigeration capacity is decreased. For cases of H= 1.5, 1.75, 2 T when the increment of magnetic
field is ∆H= 0.25 T, the increasing degree of refrigeration capacity is 30.1 % and 69.2 %, respectively. The higher
the magnetic field strength is, the greater refrigeration capacity is.

124
Table 1.Parameters of the simulation used for numerical analysis.

Parameters Values Dimensions

dp 600 m
L 0.1 M
 25 % -
D 0.01 M
x1 0.72 -

m bd 41.8 G
I

x11 0.92 -

m bd II 41.5 G

m wg 0.028 kg s -1

H low 0 T

H high 1.5 T

1  3 0.2 S

2  4 0.5 S

TH 293 K
TC 255 K

Fig. 3. Cooling capacity as functions of fluid mass flow rate at different magnetic field values.

Fig. 4 shows the general behavior of the COP as a function of the fluid mass flow rate;. Figure 4 shows
that an increase of the fluid mass flow rate decreases the COP due to fluid mass flow rate becoming too large for
the regenerator.
The variations in exergy efficiency with changes in the mass flow rate of heat transfer fluid are presented
in Fig. 5. According to figure , exergy efficiency drops as the mass flow rate of heat transfer fluid increases for all
applying magnetic field. In addition, as magnetic field increases the exergy efficiency increase at any chosen value
of mass flow rate.
Fig.6 depicts the effect of rising in fluid mass flow rate on the exergy destruction rate inside the
regenerator. It is found that when fluid mass flow rate increases, all curves reach their maximum values and then
decrease slightly with increasing fluid mass flow rate. On the other hand, excessive mass flow cause to decrease
in cooling power and due to increase in work pump resulting in decrease in the cascade AMR performance. Fig. 7
illustrates the total exergy destruction of the system as a function of the fluid mass flow rate. It can be observed
the exergy destruction increases with the increase of the fluid mass flow rate as a result of required pumping
power. The analytical results show that in order to reach optimal performance, mass flow rate should be designed
carefully regarding to various operating conditions.
In this study, an exergoeconomic analysis is also conducted where the cost formation can be determined
for the cascade AMR refrigeration system. The cost for the magnetocaloric material and magnet material is taken

125
to be 20 $Kg -1 and 40 $Kg -1, respectively. The variation of cost per unit of cooling with respect to the specific
exergetic cooling power at various magnetic fields is shown in Fig. 8. It can be observed that an increase in the
specific exergetic cooling power causes a decrease in the cost per unit of cooling.

Fig. 4. COP as functions of fluid mass flow rate at different magnetic field values.

Fig. 5. Exergy efficiency against fluid mass flow rate at different magnetic field values.

Fig. 6. Exergy destruction of regenerator as functions of fluid mass flow rate at different magnetic field values.

126
Fig. 7. Total exergy destruction of the system versus the fluid mass flow rate at different magnetic field values.

Fig. 8. Cost per unit of cooling as functions of the specific exergetic cooling power at different magnetic field values.

CONCLUSIONS
In this study, a mathematical model is developed to investigate the performance characteristics of a cascade
active magnetic regenerative refrigeration system. Also, the model was applied in order to predict the cascade
AMR performance characteristics at various selected design parameters. The main findings of this study are
summarized as follows:
 The refrigeration capacity initially raise at lower mass flow rate up to the point that it attains its maximum value
and eventually decrease with the increase of fluid mass flow rate.
 An increase in the magnetic field from 1.5 T to 2 T, the COP and exergy efficiency of the cascade AMR
refrigeration increases.
 Both COP and exergy efficiency decreases as the mass flow rate increases.
 An increase in fluid mass flow rate increases the overall exergy destruction rate of the system as a result of
pumping power requirements.
 An Increase in the specific exergetic cooling decreases the cost per unit of cooling.

NOMENCLATURE
2
A Heat transfer area, m
B constant magnetic field,  H
c specific heat, J kg K -1
COP coefficient of performance
D diameter of the regenerator section, m
dP diameter of the particles,  m

127
Ex exergy flow rate (W)
h convection coefficient (W m -2 K -1)
H magnetic field, A m -1
Hmax maximum magnetic field, A m -1
k thermal conductivity, W m -1 K -1
L length of the regenerator, m
m mass, kg
m mass flow rate, kg s -1
M magnetization per unit mass, A m 2 kg -1
MCM Magnetocaloric material
Nu Nusselt number, dimensionless
Pr Prandtl number, dimensionless
Q heat transfer rate,W
Re Reynolds number, dimensionless
s specific entropy (J kg -1 K -1)
t time coordinate, s
T temperature, K
V volume, L
x axial position, m
x mass fraction
W work, kJ s -1
P pressure drop, Pa
Greek Letters
 porosity
0 permeability of free space (m kg s -2 A -2)
 density kg m-3
 efficiency (-)
 The specific exergy cooling of the system (W T -1 L -1)
Subscripts
ad adiabatic
bd bed
bdI first stage bed
bdII second stage bed
c cold or low field
D demagnetization
des destruction
ex exergy
f fluid
H hot or high field
inf undisturbed flow
I first stage of the cascade system
II second stage of the cascade system
M magnetic
P pump
s solid
wg water–glycol mixture

REFERENCES
Aprea, C., Greco, A., Maiorino, A. 2011. A numerical analysis of an active magnetic regenerative cascade system,
International Journal Energy Research 35:177–188.
Aprea, C., Greco, A., Maiorino, A. 2012. Modelling an active magnetic refrigeration system: A comparison with different
models of incompressible flow through a packed bed, Applied Thermal Engineering 36: 296-306.
Barclay, JA., Steyert, WA.1982. Active magnetic regenerator. US Patent No. 4,332,135.
Barclay, JA.1983. Wheel-type magnetic refrigerator. US Patent No. 4,408,463.
Bejan, A., Tsatsaronis, G., Moran, M. 1996. Thermal Design and Optimization. Wiley: New York.
Bjørk, R., Engelbrecht, K. 2011. The influence of the magnetic field on the performance of an active magnetic
regenerator (AMR), International journal of refrigeration 34: 192-203.
Bjørk, R., Smith, A., Bahl, C.R.H., Pryds, N. 2011. Determining the minimum mass and cost of a magnetic refrigerator.
International journal of refrigeration 34: 1805-1816.

128
Dobrovicescu, A., Tsatsaronis, G., Stanciu, D., Apostol, V. 2011. Consideration upon Exergy Destruction and
Exergoeconomic Analysis of a Refrigerating System, Revista de Chimie 62(12): 1168-1174.
Engelbrecht, K., Bahl, C.R.H. 2010. Evaluating the effect of magnetocaloric properties on magnetic refrigeration
performance, Journal of applied physics 108: 123918.
Engelbrecht, K., Nellis, G.F., Klein, S.A. 2006. Predicting the performance of an active magnetic regenerator refrigerator
used for space cooling and refrigeration, HVAC&R RESEARCH 12: 1077-1095.
Engelbrecht, K.L., Nellis, G. F., Klein S.A. 2006. The effect of internal temperature gradients on regenerator matrix
performance, Journal of Heat Transfer 128 (10): 1060-1069.
Kaviany, M. 1995. Principles of Heat Transfer in Porous Media. Springer: New York, NY, 33, 46–47, 130, 228–229.
Kitanovski, A., Egolf, P.W. 2009. Application of magnetic refrigeration and its assessment, Journal of Magnetism and
Magnetic Materials 321: 777–781.
Li, J., Numazawa, T., Nakagome, H., Matsumoto, K. 2011. Numerical modeling on a reciprocating active magnetic
regenerator refrigeration in room temperature, Cryogenics 51: 347–352.
Li, P., Gong, M., Yao, G., Wu, J. 2006. A practical model for analysis of active magnetic regenerative refrigerators for
room temperature applications, International Journal of Refrigeration 29: 1259-1266.
Nellis, G.F., and Klein, S.A. 2006. Regenerative heat exchangers with significant entrained fluid heat capacity.
International Journal of Heat and Mass Transfer 49:329–40.
Rohsenow, W.M, Hartnett, J.P., Ganic, ENI. 1985. Handbook of Heat Transfer, vol. 6. McGraw-Hill: New York, NY, 10–11.
Rowe, A. 2011a. Thermodynamics of active magnetic regenerators: Part I. Cryogenics, 52:111–118.
Rowe, A. 2011b. Thermodynamics of active magnetic regenerators: Part II. Cryogenics, 52:119–128.
Rowe, A. 2011. Configuration and performance analysis of magnetic refrigerators, International Journal of
Refrigeration, 34, 168-177.
Tura, A., Nielsen, K.K., Rowe, A. 2012. Experimental and modeling results of a parallel plate-based active magnetic
regenerator; International Journal of Refrigeration 35 (6): 1518-1527.

129
 THERMODYNAMIC AND COST ANALYSES OF A RESIDENTIAL HYBRID PV-FUEL CELL-
BATTERY SYSTEM FOR A CANADIAN HOUSE

M. Hosseini, I. Dince and M. A. Rosen

Faculty of Engineering and Applied Science, University of Ontario Institute of Technology
2000 Simcoe St North, Oshawa, ON, L1H 7K4, Canada

ABSTRACT
A residential photovoltaic (PV)-based hydrogen fuel cell system is analyzed using energy and exergy methods,
and its monthly performance is investigated. ThePV system is accompanied by a water electrolyser for
hydrogen production, a lead acid battery pack, and a solid oxide fuel cell (SOFC) for reconverting the hydrogen
produced to electricity during periods of solar unavailability. The solar irradiance is based on a monthly average
in Toronto in 2011. The energy and exergy analyses results reported include the PV power output and the
shares attributable to the battery and the SOFC in meeting the electrical demand. The exergy destruction of the
main components and the overall efficiencies are presented. A cost analysis is performed to determine the
electricity unit cost over the system lifetime.

Keywords: Photovoltaic, Hydrogen, Fuel Cell, Energy, Exergy

1. INTRODUCTION
Solar photovoltaic (PV) systems are capable of converting 10–20% of solar energy into electricity with zero
greenhouse gas emission during their operation. These systems can be implemented in residential applications
for electricity generation. However, solar radiation is intermittent, and supplying the electricity of a house
requires the use of some energy storage options. While the PV system supplies the electricity demand of the
house, the surplus electricity generated can be used in a water electrolyser for hydrogen production. Hydrogen
is stored in compressed hydrogen tanks, to be fed to a fuel cell later in the night or during periods of solar
unavailability. With this hybrid system, not only is the power demand is supplied, but a certain amount of the
thermal demand is also met.
This type of hybrid PV-fuel cell-battery system is the focus of several studies (Shabani et al. 2010, Lagorse et
al. 2008, Morrison et al. 2006, Hosseini et al. 2013). The use of solar energy in a renewable electricity
generation system for a residential area in Italy is studied by Santarelli et al. (2004), with solar energy being
one of three main renewable energy resources considered. The result revealed that wind energy is not capable
of meeting the demand, for that specific area. Since solar availability is more constant rather than micro-hydro
energy throughout the year, smaller hydrogen storage capacity is required. Uzunoglu et al. (2009) investigate a
renewable energy power generation system which utilizes hydrogen and ultra-capacitors as energy storage
options. The power generation system provides the electricity demand of a house based on the solar irradiance
data in Turkey. The main contribution is the modeling and analysis of a renewable based energy resource in a
residential fuel cell hybrid system with short and long term storage options. The parametric design and dynamic
behaviour of the hybrid system is examined.
The components of the PV-fuel cell-battery system have been modeled separately by researchers.
Sukamongkol et al. (2002) study the performance of a PV system with simulation. Chenni et al. (2007) and
Nordin and Omar (2011) validate the results of detailed mathematical models with I-V characteristic and
maximum power point of real PV modules. A direct coupling of photovoltaic systems to water electrolysers for
hydrogen production is the main focus of the study by Clarke et al. (2009), who found that the system cost and
therefore hydrogen production cost decreases with their proposed configuration. This is achieved while
retaining minimum energy loss and maximum safety for the system. Solid oxide fuel cells are increasingly being
applied in stationary power generation, especially in combined heat and power (CHP) systems and in remote
areas. These systems have been studied comprehensively,in terms of experimentation, modeling,
thermodynamic analysis and implementation (La Oet al. 2007, Dincer et al. 2007, Motahar and Alemrajabi
2009, Knightet al. 2007).
The performance of the PV-fuel cell system depends on weather conditions; here, the system is analyzed for a
Canadian house with relevant solar data. The investigation of the hybrid PV-fuel cell system based on energy
and exergy (quality of energy) determines the rates of energy and exergy flows in each component. The results
of the analyses are used for calculating system outputs, efficiencies, and exergy destruction rates. The main
contribution of the present research is the analysis of the hybrid PV-fuel cell system on a monthly basis. The

130
daily average solar irradiance during each month of operation is taken as the input to the PV modules. The total
daily average outputs of the system components are reported, along with system efficiencies. The annual total
exergy destruction of the system main components is presented, and the electricity unit cost is calculated for
the system lifetime.

2. SYSTEM DESCRIPTION
The configuration of the hybrid PV-fuel cell-battery system,shown in Figure 1,is selected to meet the electricity
requirement of a Canadian house. Solar energy is converted to electricity by the PV modules. The electricity
generated is directed to a load controller, which makes the decision of energy distribution within the system
components and the house. The load controller directs a part of the electricity generated to meet the power
demand. The electrolyser converts the surplus electricity to hydrogen, which is stored in a pressurized
hydrogen tank. During periods of solar unavailability the fuel cell and battery cover the load. The operation of
the battery is limited to the conditions where the electric load exceeds the solid oxide fuel cell (SOFC) nominal
power.

DC Current
Battery Pack

Electrolyzer DC Current

Load Controller
H2
DC/AC
Converter
Hydrogen Water
Tank
AC Current
Air

DC Current
Heat SOFC Power Load

Steam
Stack Gas

Absorption Thermal
HRSG Cooling Air
Chiller Load
Water

Figure 1: Schematic of the residential PV-hydrogen-fuel cell CHP system.

The SOFC is fed with the hydrogen (the fuel, which is stored in the storage tank) and ambient air as the
oxidant. Hydrogen is preheated via heat transfer from an external source, while air preheating is performed
utilizing the heat from the SOFC stack gases, which have adequate energy for recovery for heating or hot water
production. This is accomplished using a heat recovery steam generator, which generates low pressure
saturated steam. As shown in Figure 1, the steam generated in the Heat Recovery Steam Generator (HRSG)
can be used in an absorption chiller for cooling purposes, or it can be used for hot water production.
The operational algorithm is simplified and illustrated in Figure 2. Once the PV output power and the load
demand are calculated, the control system decides on the operation of the system components. A lower
demand than the PV output results in the operation of the electrolyser or the charging of the battery pack. If the
demand is higher than the electric power of the photovoltaic system, the fuel cell and the battery will be in
operation.

131
3. ENERGY AND EXERGY ANALYSES
The modeling and energy and exergy analyses of the hybrid PV-fuel cell-battery system are briefly presented in
this section.

Electricity Demand
Solar Irradiance
Pdemand

Calculate PV Electric Power

Output ∆P=PPV - Pdemand
PPV

∆P ≥ 0 No

Yes Yes |ΔP| ≤ FCnominal No

PSOFC = FCnominal
Pelectrolyser = PVpower – Pcharge, Battery PSOFC = FCnominal
Pbattery = |ΔP|- PSOFC

Figure 2: Solving algorithm for providing the detached house with electricity.

3.1 Photovoltaic System

A previous paper (Hosseini et al. 2013)reported energy and exergy analyses of a residential PV-fuel cell-
battery for a Canadian detached house. That analysis is used in this work. The current-voltage characteristics
are calculated as a function of light current and reverse saturation current. The cell temperature and series
resistance affect the I-V characteristics, as well. The general equation to obtain the I-V characteristics of the PV
modules can be expressed as(Chenni et al. 2007)

(1)

where is solar insolation W/m2, is the solar insolation at the design condition, is calculated based on
the manufacturer data for short circuit and maximum point currents, and is the manufacturer supplied
temperature coefficient of short-circuit current (A/°C). The maximum power output of the PV modules is at a
point which voltage and current have their maximum values. The control system for the PV system is designed
so that the system operates at the maximum power point.The energy and exergy efficiencies of the PV cell at
maximum power point are given by

(2)

(3)

132
3.2 Water Electrolyser
For surplus electricity from the PV system, , the energy efficiency relation of the electrolyser is used to
calculate the produced hydrogen flow rate:

(4)

3.3 Solid Oxide Fuel Cell

The Butler-Volmer equation is used to relate the current and voltage output of the solid oxide fuel cell. Relevant
voltage drops due to ohmic, activation, and concentration losses are considered in the calculation of SOFC
electric power output according to Dincer et al. (2007) and Motahar and Alemrajabi (2009). The energy and
exergy efficiencies for the fuel cell system are

(5)

(6)

The purpose of the hybrid system is to provide a detached house with zero-emission electricity. The system
also provides a part of the heating/cooling demand. This is performed through heat recovery from the fuel cell
stack gases.

3.4 Heat Recovery Steam Generator

The HRSG is treated as a heat exchanger. Sub-cooled water enters the economizer section and receives heat
from the gases leaving the HRSG. After it heats up to the saturation temperature at the boiler pressure, water
vapor is formed inside the pipes. The saturated steam generated in the HRSG leaves the unit for hot water
production or cooling purposes.

3.5 Hybrid PV-Fuel Cell-Battery System

The overall efficiencies of the hybrid system consider all the useful outputs, and the inputs to the system. The
electric power demand of the house and the heat recovered from the fuel cell exhaust gases are considered as
outputs. The inlets to the overall hybrid PV-Fuel Cell-Battery system are solar irradiance and the heat required
to preheat the fuel feed to the SOFC.

(7)

(8)

4. ASSUMPTION AND DATA

The thermodynamic and cost analyses are based on the following assumptions and data:
 The average daily demand is used of a 140 m2 detached house (22.2 kWh/day).
 The solar irradiance is based on a monthly average in Toronto in 2011. The calculations are for each
month, separately.
 The Proton Exchange Membrane PEM electrolyser is selected to operate at 30 bar and 65%efficiency.

133
  The hydrogen generated by the electrolyser process is stored at 25 bar on a seasonal storage basis.
The size of the storage tank is determined based on the seasonal need for hydrogen.
 The fuel cell is an atmospheric SOFC.
 The fuel cell nominal power is 1 kW (after accounting for 4% internal consumption). The size of the fuel
cell determines the size of the PV array, the hydrogen storage tank and the battery.
 The annual amount of the stored hydrogen must be positive; therefore a PV panel with 86 modules is
selected. The nominal power output of each module is 210 W (at standard test conditions (STC)).

Figure 3: Hourly electric power demand of the detached house broken down by hour of the day (Knight et al. 2007).

Figure 3 shows the daily electricity demand of the house. These sets of data are used in the previous work of
the authors (Hosseini et al. 2013) for comparison of the performance of the system in summer and winter of
2011. Here, the data, which are adopted from the study for International Energy Agency (Knight et al. 2007),
are used to perform the thermodynamic analysis and cost evaluation of the system for each month of
operation.

Table 1: Equipment purchase and operation and maintenance costs

System and component Nominal size Unit cost Cost
PV system (SunPower 2012)
PV system 18.06 kW 5156 $/kW US$ 93113.66
O&M (Enbar 2010) 47$/kW.year US$/year848.82

Fuel Cell system (Braun 2011)

SOFC, Balance of Plant (BOP) 1 kW 2296.73 $/kW US$ 2296.73
O&M 5% of purchase US$/year 114.84

Stored hydrogen 117.29 kg (3909.28 kWh)

3
Hydrogen storage tank 58.12 m at 25 bar, 300 K 4 $/kWh US$ 15636.73
O&M 3% of purchase US$/year 469.10

Electrolyser
PEM Electrolyser, BOP, Installation 14.45 kW 586.5 $/kW US$ 8474.93
O&M 5% of purchase US$/year 423.75

Battery, BOP, Installation 5 kWh 185 $/kWh US$ 937.50

O&M 3% of purchase US$/year28.12

134
Total hybrid system capital cost US$ 120459.55
Operation and maintenance US$/year 1888.63
Total O&M cost over the system lifetime (25 years) US$ 47115.75
Total estimated system cost over 25 years US$ 167575.30
The University of Toronto publishes solar irradiance data in Toronto over the past 6 years. Here, the 2011 solar
irradiance presented in Figure 4 is used (University of Toronto 2012). Although the average monthly solar
radiation on 24 hour basis is used in the analyses, Figure 5 is presented to illustrate the significant difference
between the summer and winter solar availability. The performance of the PV modules, and as a result the
hybrid system components, are strongly dependent on the solar irradiance, as discussed further in section 5.
The cost model considers only the purchase, installation labour, and operation and maintenance of the system
components. The details are provided in Table 1.

5. RESULTS AND DISCUSSION

The load profile depends on the instantaneous demand, which requires precise measurement of electricity
consumption. However, the thermodynamic analyses can be performed based on average values to determine
the initial sizes of the system components, the pattern of energy flow in the electricity generation system, and
the sources of losses. The results can also help developing a control strategy and dynamic model for further
investigations.

Figure 4: Average daily solar irradiance over each month of 2011 (adapted from University of Toronto 2012).

In this study, the hourly solar irradiance and load demand are considered in the analyses. The data for solar
irradiance are derived from the average values over each month of the year 2011. However, the presented
results only show the total daily generation/consumption of each system component. The system is considered
to perform the same throughout the month.
The electric power supply by each component of the hybrid PV-fuel cell-battery system (in kWh per day) is
presented in Figure 5. The results are reported for each month, considering average values of solar irradiance
in each month. Figure 5 presents the daily share of power supply of the PV modules, fuel cell, and batteries in
each month. With the increase in solar irradiance in the summer months (Figure 4), the PV electric power
output increases significantly. Therefore, most of the electricity demand is directly met by the PV output, and
the fuel cell has its minimum share in supplying the demand. To quantify this significant difference, the power
penetrations of the PV system and the fuel cell-battery are shown in Figure 6. Due to less solar irradiance in
Fall-Winter months, the fuel cell-battery penetration in demand is as high as 80% in January and December.
The supply share of the fuel cell-battery of the demand decreases with the rise in solar availability in summer
months.

135
Figure 5: Electric power output of the PV-fuel cell system component on a daily basis as a monthly average.

Figure 6: PV and SOFC-battery power penetration.

The surplus electricity generated by the PV modules is directed to the electrolyser. The daily hydrogen
production rate is shown in Figure 7, with July having the maximum production rate as 2.33 kg/day. The fuel
cell hydrogen consumption is related to its power output rate; therefore, more hydrogen is consumed by the
SOFC during winter days. The difference between the electrolyser hydrogen production rate and the SOFC fuel
consumption rate is taken as the amount of the hydrogen which needs to be stored in or provided by the
hydrogen storage tank. The positive values for hydrogen storage in Figure 5 illustrate more production than
what is consumed by the fuel cell. The negative values mean that the electrolyser is not capable of producing
enough hydrogen for the SOFC, and hydrogen is supplied by the storage tank. The hydrogen storage tank
must be large enough to cover the total consumption during the months with more consumption than
production.

Table 2: Total yearly exergy destruction in the major hybrid system components

Component PV Electrolyser SOFC HRSG

Exergy destruction, kWh/year 81974.77 4734.40 5249.03 510.18

136
The annual exergy destruction in the hybrid PV-fuel cell-battery system is shown in Table 2, which shows that
the PV modules are the main source of exergy destruction in the system. Improving the efficiency of the PV
cells leads to a lower exergy destruction. Exergy destruction in the fuel cell is caused by several phenomena,
e.g. electrochemical reactions, and high temperature preheating of the air and hydrogen flows.

Figure 7: Monthly hydrogen storage/consumption.

The efficiencies of the system components change when their operating conditions deviate from the nominal
conditions. The variations of the efficiencies of the PV and fuel cell systems are presented in Figure 8. In
summer days, the average daily solar irradiance is closer to the standard test condition of the photovoltaic
cells. However, the fuel cell appears to work at its nominal power rate throughout the year. Higher values are
obtained for exergy efficiency because solar exergy, the input to the PV module, is less than solar energy. In
the case of the fuel cell, the exergy of the required heat for hydrogen preheating is less than its energy content,
which results in higher efficiencies based on exergy rather than energy.

Figure 8: Energy and exergy efficiencies of the PV and SOFC systems.

However, interesting results are obtained for the overall efficiencies of the hybrid PV-fuel cell-battery system.
According to Figure 9, the efficiencies differ significantly with month. Higher values are reported for Fall-Winter
months in which the fuel cell-battery penetration is a maximum, and lower values are reported for Spring-

137
Summer months in which the PV modules provide a greater portion of the demand. The difference in the
efficiencies of PV and fuel cell systems is the cause of such results.

Figure 9: Energy and exergy efficiencies of the hybrid PV-fuel cell-battery system.

Figure 10: Annual electricity unit cost during the hybrid system lifetime.

The unit cost of electricity is a function of the total estimated cost and the total electricity production over 25
years of operation. The total annual electricity production is the sum of the daily production over 365 days of
the year. Therefore, the electricity unit cost (EUC) is 0.83 US$/kW.However, another analysis is made to
estimate the cost of electricity in each year of operation of the system. If the capital costs are assumed to be
paidoff in 10 years, consideringthe inflation rate (3%), the electricity unit cost will follow the trend shown in
Figure 10. A significant drop is observed for the 11th year of operation, since the only costs of the system are
associated with operation and maintenance costs.

6. CONCLUSIONS
Several conclusions can be drawn from the thermodynamic and cost analyses reported here of a hybrid PV-
fuel cell-battery system as applied to a Canadian house. The nominal maximum power output of the PV
modules is 18.06 kW, which is 5 times higher than the demand maximum power. The photovoltaic system is
sized to meet the demand either directly or by storage of hydrogen generated in the electrolyser. Due to less
solar irradiance in Fall-Winter months, the fuel cell-battery penetration in supplying of the demand is as high as

138
80% in January and December. The supply share of the fuel cell-battery in meeting the demand decreases with
the rise in solar availability in Summer months. The maximum H2 production rate is in July (2.3 kg/day), and the
hydrogen consumption is maximum in January (0.99 kg/day).The efficiencies differ significantly with the time of
the year. Higher values are observed for fall-winter months in which the fuel cell-battery penetration is a
maximum, and lower values are observed for Spring-Summer months in which the PV module provides a
greater portion of the demand.The electricity unit cost is 0.83 $/kWhbased on a 25-year economic evaluation
period of the hybrid PV-fuel cell-battery system.

ACKNOWLEDGMENT
The authors acknowledge the support provided by the Natural Sciences and Engineering Research Council of
Canada.

NOMENCLATURE
Specific exergy, kJ/kg
Energy flow rate, kW
Exergy flow rate, kW
Solar irradiance, kW/m2
PV light current, A
Reverse saturation current, A
PV electric current, A
Boltzmann constant
Manufacturer supplied temperature coefficient of short-circuit current, A/°C
Lower heating value, kJ/kg
Mass flow rate, kg/s
Power, kJ
Heat transfer rate, kW
Series resistance of the PV cells, Ohm
Temperature, K
Electric voltage, V
Work rate, kW

Greek letters
PV cell shape factor
Energy efficiency, %
Exergy efficiency, %

Subscripts
Ambient condition
PV cells
Hydrogen
Input to the electrolyser
Maximum power
Solid oxide fuel cell

Acronyms
CHP Combined heat and power
PV Photovoltaic
SOFC Solid oxide fuel cell
STC Standard test condition

REFERENCES
Shabani,B., J. Andrews, and S. Watkins. 2010. Energy and cost analysis of a solar-hydrogen combined heat
and power system for remote power supply using a computer simulation. Solar Energy 84:144-155.

139
Lagorse,J., M. G. Simoes, A. Miraoui, and Ph. Costerg. 2008. Energy cost analysis of a solar-hydrogen hybrid
energy system for stand-alone applications.Int. J. Hydrogen Energy 33:2871–2879.
Morrison, I. B., M. Mottillo, A. Ferguson, H. Ribberink, L. Yang, and K. Haddad. 2006. The simulation ofa
renewable-energy-powered hydrogen-based residential electricity system.Second National IBPSA-USA
Conference, Cambridge, MA, 2-4 August2006.
Hosseini, M., I. Dincer, and M. A. Rosen. 2013. Hybrid solar-fuel cell CHP Systems for residential applications:
Energy and exergy analyses. Journal of Power Sources 221:372-380.
Santarelli,M., M. Cali, and S. Macagno. 2004. Design and analysis of stand-alone hydrogen energy systems
with different renewable sources. Int. J. Hydrogen Energy 29:1571–1586.
Uzunoglu, M., O.C. Onar, and M. S. Alam. 2009. Modeling, control and simulation of a PV/FC/UC based
hybridpower generation system for stand-alone applications. Renewable Energy 34:509–520.
Sukamongkol , Y., S. Chungpaibulpatana, and W. Ongsakul. 2002. A simulation model for predicting the
performance of a solar photovoltaic system with alternating current loads. Renewable Energy 27:237–258.
Chenni , R., M. Makhlouf, T. Kerbache, and A. Bouzid. 2007. A detailed modeling method for photovoltaic cells.
Energy 32:1724–1730.
Nordin, A. and A. Omar. 2011. Modeling and simulation of photovoltaic (PV) array and maximum power point
tracker (MPPT) for grid-connected PV system. 3rd International Symposium and Exhibition in Sustainable
Energy & Environment, Melaka, Malaysia, 1-3 June 2011.
Clarke, R.E., S. Giddey, F.T. Ciacchi, S.P.S. Badwal, B. Paul, and J. Andrews. 2009. Direct coupling of an
electrolyser to a solar PV system for generating hydrogen. Int. J. Hydrogen Energy 34:2531–2542.
La O, G. J., H. J. In, E. Crumlin, G. Barbastathis, and Y. Sh. Horn. 2007. Recent advances in microdevices for
electrochemical energy conversion and storage.Int. J. Energy Res. 31:548–575.
Colpan, C.O., I. Dincer, and F. Hamdullahpur. 2007. Thermodynamic modeling of direct internal reforming solid
oxide fuel cells operating with syngas. Int. J. Hydrogen Energy 32:787–795.
Motahar,S. and A. A. Alemrajabi. 2009. Exergy based performance analysis of a solid oxide fuel cell and steam
injected gas turbine hybrid power system. Int. J. Hydrogen Energy 34:2396–2407.
Zhang, X., S. H. Chan, G. Li, H. K. Ho, J. Li, and Zh. Feng. 2010. A review of integration strategies for solid
oxide fuel cells.J. Power Sources 195:685–702.
Knight, I., N. Kreutzer, M. Manning, M. Swinton, and H. Ribberink. 2007. European and Canadian non-HVAC
electric demand and DHW load profiles for use in simulating the performance of residential cogeneration
systems. Annex 42, International Energy Agency, Energy Conversions in Buildings and Community System
Programme, May 2007.
Weather Data, Department of Geography, University of Toronto Mississauga. Web link:
http://www.utm.utoronto.ca/geography/resources/meteorological-station/weather-data, Date Accessed: June 26
2012.
SunPower Corporation. 2012. Document #001-42023, 210 Solar Panel. Web link:
www.prevailingwindpower.com/sunpower.pdf, Date Accessed: June 26 2012.
Enbar, N. 2010. PV O&M Best Practices. Utility/Lab Workshop on PV Technology and Systems. National
Renewable Energy Laboratory, 8-9 November 2010.
Braun, R. J., S. A. Klein, and D. T. Reindl. 2011. Assessment of Solid Oxide Fuel Cells in Building Applications,
Phase 1: Modeling and Preliminary Analyses. Prepared for Energy Center of Wisconsin, Report 207-R,
November 2011.

140
 RETROFITTING OF A HEAT EXCHANGER NETWORK
BY USING EXERGY ANALYSIS

N. Akgün*, Z. Özçelik, B. Kuryel

Chemical Engineering Department, Ege University, İzmir, Turkey
nursen.akgun.1986@gmail.com, zehra.ozcelik@gmail.com, beno.kuryel@gmail.com

ABSTRACT
Energy is a driving force behind the social and economic improvement of a country. Thus, most of the industrial
studies focus on design methods dealing with energy recovery and waste minimization in chemical processes. In
industrial processes, most of the energy is used for heating and cooling purposes. So, an effective design is
required for heating and cooling systems. Pinch Analysis (PA) is an effective method used in heat recovery.
However, it deals with the amount of energy rather than its quality. In such a case in addition to its amount, the
quality (in other words the availability) of energy is examined by using exergy analysis. Real processes in a plant
or energy system are irreversible and some energy is degraded. The main purpose of exergy is to discover the
causes of imperfections in a thermal or power system and recover the energy which is degraded.
In this study, an exergy analysis is carried out on a system that is retrofitted by using pinch analysis. As a result of
pinch analysis, 38.98 % of hot utility and 18.26 % of cold utility saving is obtained in addition to an energy recovery
up to 12.51 % is achieved. By applying exergy analysis to the heat exchanger network (HEN) retrofitted by PA, the
exergy losses depending on the imperfections, irreversibilities, frictions and interactions between substances are
calculated. Thus, by identifying the major causes of energy destruction within the system, some part of energy
destroyed is recovered. Thereby, significant energy savings are achieved with lower cost and higher efficiency. As
a result of exergy analysis, it is seen that 3.57 % of exergy is destroyed in Retrofit 1 whereas the exergy
destruction is 1.41 % in Retrofit 2.
Keywords: Energy, Exergy, Pinch Analysis, Retrofitting.

INTRODUCTION
The past decade has seen significant industrial and academic efforts devoted to the development of process
design methodologies that target energy conservation for a large variety of chemical process industries (Raskovic
et al., 2009). Especially energy saving is the most important issue in the petrochemical industry associated with
cost. The petrochemical industry is a capital intensive industry consuming much energy and the energy cost
contributes significantly to the total cost (Yoon et al., 2007).
The majority of energy consumed in industry is typically used mainly for heating and cooling purposes. Efficient
design of heating and cooling systems in industry is therefore vital and can be accomplished through design
optimal heat recovery network using tools such as Pinch Analysis (PA). PA is a systematic technique for the
design of thermally efficient systems. It allows the designer to identify the minimum heating and cooling
requirements and maximum energy recovery potential by identifying a thermodynamic bottleneck or the pinch
point for heat recovery (Wan et al., 2010).
However PA is concerned only with energy conservation, and therefore it cannot show how or where
irreversibilities occur in a system or process. To determine the irreversibilities, the exergy analysis method is
applicable. Exergy is defined as maximum amount of work, which can be produced by a system when it comes to
equilibrium with a reference environment. The standard conditions of the earth atmosphere are considered as the
thermodynamic state of the environment. Exergy analysis can clearly indicate the locations of energy degradation
in a process that may lead to improved operation or technology. It can also quantify the quality of heat in a reject
stream. So, the main aim of exergy analysis is to identify the causes and to calculate the true magnitudes of
exergy losses (Dincer et al., 2001). Pinch analysis deals with the quantity of energy and its aim is to reduce the
consumption of energy. However, exergy analysis deals with the quality of energy and targets to reduce the loss of
energy. Therefore, combination of pinch and exergy analysis could be an appropriate an appropriate tool for
simultaneous study of heat and power.

EXERGY CONCEPT
The increase of the energy efficiency in industry has always been one of the technical-scientific goals of
managers, engineers and scientists, but the problem has now become a priority. The experience of the last three
decades in the implementation of energy saving programs has shown that it is possible to obtain up to 15 %
savings by first level measures, i.e. “energy conservation”, avoiding wastes, like effective maintenance, increased
consciousness, etc. However, in order to go further in these efforts it is necessary to apply second and third level
measures, normally requiring additional investment, and doing so, it is necessary to answer the following
questions:
 Where and how energy is degraded in industry
 How to reduce energy degradation

141
As it is well known energy cannot be created nor destroyed but only transformed (1st law). However in all real
processes its quality decreases (2nd law) and the increasing relevance of the second law is the result of two
significant changes produces at the international level of energy policies. As shown in figure below, the exergy
concept is able to deal with three E’ s model because of its implications not only in terms of energy but also in
terms of ecology (Rivero, R., 2002).

METHODOLOGY
The 1st law of thermodynamic is based on the conservation of energy. However, losses occur when the efficiency
of a device or process deviates from the efficiency that would occur if the device or process were ideal. The value
of a loss is a measure of this deviation from ideality. In addition, exergy losses allow the location, type and cause
of a loss or inefficiency to be clearly identified. This information is critical for efforts to increase the exergy
efficiency (Dinçer, I. and Rosen, M. A., 2007)
The physical, kinetic, chemical, nuclear magnetic, electrical and surface tension forms the total exergy of the
system but for the sake of simplicity and since the effects of the nuclear, magnetic, electrical, and surface tension
are negligible. The total exergy is assumed to be formed from physical , chemical , kinetic and
potential exergies as shown in Eqn.(1).

(1)

The sum of kinetic, potential and physical exergies is also mechanical exergy. Although exergy is an extensive
property, it is often convenient to work with unit-of-mass or molar basis. The total specific exergy on mass basis is
given by Eqn. (2).

(2)

When evaluated relative to the environment, the kinetic and potential energies of a system are in principle fully
convertible to work as the system is brought to rest relative to the environment, and also they correspond to the
kinetic and potential exergies, respectively.

(3)

=gz (4)

Considering a system at rest relative to the environment, the kinetic and potential exergy values are equal to zero:

(at rest) (5)

If there is no chemical reaction, the chemical exergy is also zero ( . Thus, the exergy is equal to physical
exergy given as Eqn. (6).

(6)

Since the irreversibilities are inevitable in a real process, some exergy is destroyed within the process. This results
from the entropy change. Hence, for processes that undergo just physical change, the exergy destruction takes
form as Eqn. (7) and Eqn. (8).

142
 (7)

(8)

RESULTS AND DISCUSSION

In this study, an existing VCM plant is taken into consideration and exergy analysis is applied on the process that
has been retrofitted by pinch analysis. In the process 28 hot and 17 cold steams exist. By making heat and
material balances according to the stream data taken from the process, the heat capacities and heat loads of each
stream are calculated by using ASPEN Plus and ASPEN HYSYS (Table 1). The minimum approach temperature
is assumed to be 10 ˚C and the pinch temperature (found as 102.5˚C) and minimum requirement of hot and cold
utilities are obtained.

Table 1. Data of process streams

HOT STREAMS COLD STREAMS
Stream Tin Tout CP Stream Tin Tout CP
Number (°C) (°C) (kW/°C) (kW) Number (°C) (°C) (kW/°C) (kW)
H1 102 40 87.82 5138 C1 40 140 0.70 69.8
H2 40 10 6.51 195.4 C2 -28 145 2.87 497.1
H3 95 65 5.31 159.4 C3 60 150 4.05 364.2
H4 73 44 11.38 330 C4 11 150 2.33 323.8
H5 45 5 0.98 39.3 C5 42 71 5.27 152.9
H6 92.5 81.5 803.47 8838 C6 97.5 98.5 857.5 857.5
H7 81 56 62.39 1559.7 C7 62 78 6.98 111.7
H8 88.8 40 1.96 95.7 C8 100 101 526.7 526.7
H9 56 10 6.99 321.7 C9 60 103 20.48 880.7
H10 100 80.5 5.67 110.6 C10 103 160 22.45 1279.4
H11 148 80 147.98 10062.6 C11 160 203.7 115.77 5059.1
H12 80 40 49.82 1993 C12 102 103 1699 1699
H13 85.4 -10.3 0.48 45.9 C13 159 160 3925.4 3925.4
H14 -28 -32 284.48 1137.9 C14 75.3 76.3 691.5 691.5
H15 102 65 17.47 646.6 C15 110 111 5061.4 5061.4
H16 156.4 60 9.62 927.6 C16 130 131 3479.9 3479.9
H17 40 39 2937.7 2937.7 C17 79.9 80.9 507 507
H18 40 35 7.44 37.2
H19 75.4 35 8.57 346.2
H20 81 60 187.24 3932.1
H21 90 65 297.49 7437.2
H22 44.5 37 98.73 740.5
H23 37 10 2.86 77.3
H24 1250 300 3.58 3403.4
H25 68 43 45.08 1127.1
H26 68 45 21.03 483.7
H27 45 40 5.2 26
H28 116.9 50 34.8 2328.3

Then the grid diagram is constructed for two alternative cases named as Retrofit 1 and Retrofit 2. In the case of
Retrofit 1, it is avoided using heaters and coolers with small heat loads. Instead, using more exchangers with
smaller heat loads is preferred. The temperatures around exchangers and heat loads of heaters, coolers and
exchanger are calculated. In the case of Retrofit 2, it is tried to use heat exchangers with higher heat loads. The
temperatures around exchangers and their heat loads are calculated both for Retrofit 1 and Retrofit 2 and the
results are given in Table 2. The exchangers eliminated after network relaxation are shown in red color. So their
heat loads are not taken into consideration in the calculation of heat recovery. Also the heat loads of utilities
(coolers and heaters) for each case are calculated and given in Table 3 and 4. The heat loads before and after
network relaxation are given since after network relaxation the heat loads of some of the utilities are increased with
the addition of heat loads of eliminated exchangers.
Then the exergy analysis is applied on Retrofit 1 and 2 in order to see the types, locations and magnitudes of
thermal losses within the process. In exergy analysis, first of all the type of utilities for heaters and coolers are
decided and the required amounts are calculated. To quantify the exergy of a system, first of all both the system
and surroundings are specified. The reference environment is assumed to have uniform temperature and
pressure. In this study the temperature and pressure of environment are taken as 1 bar and 25 °C, respectively.
Also the system is assumed to be at steady-state. For simplicity the kinetic and potential exergies are ignored.
Since there isn’t any chemical reaction, the compositions of the streams are constant leading the chemical exergy

143
to be zero. So, only the physical exergy values of each stream need to be calculated. The enthalpy depends on
the temperature and the entropy is affected by both temperature and pressure. So, the program for each stream is
constructed on Aspen HYSYS. The enthalpy and entropy values of the streams are found and tabulated as Table
5, and 6 and 7. The physical exergy of streams are calculated by using Eq.(9).

(9)

During calculations, the exergetic efficiency and exergy losses are calculated. Then the contribution of each
equipment to the whole process is calculated and shown on Table 5, 6 and 7. The exergy calculations for Retrofit
1 are shown in detailed on these tables. The final results of Retrofit 2 are given on Table 9. It is seen that in
Retrofit 1 most of the exergy is destroyed in heat exchangers and coolers (43.13 and 43.51 %, respectively)
whereas in Retrofit 2 the exergy destruction mainly occurs within the coolers (63.16 %).
The comparison of Retrofit 1 and Retrofit 2 is given in Table 8 and Table 9. It seems that Retrofit 1 is better for
higher energy recovery. However, the exergy analysis states that the exergetic efficiency of Retrofit 2 is higher.
So, an optimization can be done by making cost analysis.

CONCLUSIONS
This study deals with the energy saving strategies in a real VCM plant by applying exergy analysis to a heat
exchanger network (HEN) retrofitted by Pinch Analysis (PA). It examines the existing process and 2 alternative
retrofit cases introduced by PA.
Results obtained by PA indicate that there is a remarkable improvement in energy usage:
 According to the pinch analysis made before, the yield reduction is 0.5 % for cold and 1.06 % for hot utilities.
Also the energy recovery is only 0.34 %.
 But according to Retrofit 1, the reductions in cold and hot utilities are 18.26 % and 38.98 %, respectively. Also
the energy recovery increases up to 12.51 %.
 On the other hand, Retrofit 2 states that there is 15.25 % saving of cold and 32.58 % saving of hot utilities.
Also, in Retrofit 2, the recovered energy is 10.45 %.
 If we compare the requirement for hot and cold utilities and recovered energy between the process streams
obtained by pinch analysis, it seems that Retrofit 1 gives the best results.
Results obtained by exergy analysis shows that the exergetic efficiency of Retrofit 2 is higher:
 In Retrofit 1, most of the exergy is destroyed in exchangers and coolers (43.13 and 43.51 %, respectively).
 In Retrofit 2, the exergy destruction mainly occurs within the coolers (63.16 %).
 The total exergy destruction in Retrofit 1 is 3.57 % whereas in Retrofit 2 it is just 1.41 %.
As a further investigation cost analysis can be added on this study.

144
Fig.1. The grid diagram of Retrofit 1 (Akgun et al., 2012)

145
Fig.2. The grid diagram of Retrofit 2 (Akgun et al., 2012)

146
Table 2. Calculations of exchangers after pinch analysis (Akgun et al., 2012)
RETROFIT 1 RETROFIT 2
HOT COLD HOT COLD
Heat Heat
Exchangers T in T out T in Tout Exchangers Tin Tout Tin Tout
(°C) (°C) (°C) (°C) (°C) (°C) (°C) (°C)
1 [H28-C6] 116.9 107.5 97.5 97.88 327.3 1 [H11-C6] 113.3 107.5 97.5 98.5 857.5
2 [H11-C6] 111.08 107.5 97.88 98.5 530.2 2 [H11-C12] 124.78 113.3 102 103 1699
3 [H16-C9] 119.22 107.5 97.5 103 112.7 3 [H11-C15] 148 124.78 110 110.679 3436.7
4 [H11-C8] 114.64 111.08 100 101 526.7 4 [H28-C8] 116.9 107.5 100 100.62 327.3
5 [H16-C10] 156.4 119.22 103 118.94 357.9 5 [H16-C9] 119.215 107.5 97.5 103 112.7
6 [H11-C12] 126.12 114.64 102 103 1699 6 [H16-C10] 156.4 119.215 103 118.942 357.9
7 [H11-C15] 148 126.12 110 110.64 3237.3 7 [H24-C16] 1250 300 130 130.978 3403.4
8 [H24-C11] 850.67 300 160 177.05 1973.68 8 [H3-C5] 97 65 42 71 152.9
9 [H24-C10] 1109.9 850.67 118.94 160 921.5 9 [H10-C7] 100 80.5 62 78 110.6
10 [H24-C4] 1144.07 1109.9 97.5 150 122.3 10 [H6-C17] 92.5 91.87 79.9 80.9 507
11 [H24-C3] 1203.46 1144.07 97.5 150 212.6 11 [H6-C14] 91.87 91 75.3 76.3 691.5
12 [H24-C2] 1241.7 1203.46 97.5 145 136.9 12 [H11-C1] 107.5 107.23 40 97.5 40.1
13 [H24-C1] 1250 1241.7 97.5 140 29.7 13 [H11-C3] 107.23 106.2 60 97.5 151.6
14 [H3-C5] 95 65 42 71 152.9 14 [H11-C9] 106.2 101 60 97.5 768
15 [H10-C7] 100 80.5 62 78 110.6 15 [H16-C4] 107.5 91.74 11 97.5 151.6
16 [H6-C14] 92.5 91.64 75.3 76.3 691.5 16 [H28-C2] 107.5 97.15 -28 97.5 360.2
17 [H6-C17] 91.64 91 79.9 80.9 507
18 [H16-C3] 107.5 93.2 38.12 97.5 138.3
19 [H27-C3] 45 40 26.96 38.12 26
20 [H18-C3] 40 35 11 26.96 37.2
21 [H28-C1] 107.5 106.35 40 97.5 40.1
22 [H28-C9] 106.35 82.28 60 97.5 768
23 [H11-C2] 107.5 106.16 28.62 97.5 197.7
24 [H23-C2] 37 10 1.69 28.62 77.3
25 [H5-C2] 45 5 -12 1.69 39.3
26 [H13-C2] 85.4 -10.3 -28 -12 45.9
27 [H11-C3] 106.16 105.13 60 97.5 151.6
9934.1 8302.6
kW kW

147
Table 3. Calculations of coolers and heaters for Retrofit 1 after pinch analysis (Akgun et al., 2012)
HOT COLD
Equipment
Stream Type Stream
(kg/h) (kg/h)
Cooler 1 H1 Mixture 37064 102 40 CW 350000 30 42 5138
Cooler 2 H2 Mixture 17321 40 10 Freon 3365 2 3 195.4
Cooler 3 H4 Mixture 3000 73 44 CW 22300 30 42 330
Cooler 4 H6 EDC 108875 91 81.5 CW 443000 30 44 7639.5
Cooler 5 H7 Mixture 19269 81 56 CW 90450 30 44 1559.7
Cooler 6 H8 EDC 870 88.8 40 CW 25900 30 33 95.7
Cooler 7 H9 Mixture 4505 56 10 Freon 5540 2 3 321.7
Cooler 8 H 11 Mixture 114600 127 80 CW 403500 30 44 6957.9
Cooler 10 H 14 Mixture 10365 -28 -32 Freon 17360 -38 -37 1137.9
Cooler 11 H 15 Mixture 41636 102 65 CW 52500 30 40 646.6
Cooler 12 H 16 Mixture 23907 93.12 60 CW 25870 30 40 318.7
Cooler 13 H 17 VCM 35200 40 39 CW 476800 30 35 2937.7
Cooler 14 H 19 VCM 19000 75.4 35 Freon 5965 2 3 346.2
Cooler 15 H 20 Mixture 49000 81 60 CW 212785 30 45 3932.1
Cooler 16 H 21 Mixture 73000 90 65 CW 402465 30 45 7437.2
Cooler 17 H 22 EDC 7989 44.5 37 Freon 12755 2 3 740.5
Cooler 18 H 25 Mixture 50000 68 43 CW 114365 30 38 1127.1
Cooler 19 H 26 Mixture 15576 68 45 CW 58045 30 38 483.7
Cooler 20 H 28 Mixture 15081 84.28 50 CW 69165 30 44 1192.9

Heater 1 C 11 EDC 58082 177.05 203.7 Steam 11000 221 220 3085.22
Heater 2 C 13 EDC 937570 159 160 Steam 8000 171 170 3925.4
Heater 3 C 15 Mixture 544000 110 111 Steam 9000 121 120 5061.4
Heater 4 C 16 Mixture 440000 130 131 Steam 59000 141 140 3479.7

Table 4. Calculations of coolers and heaters for Retrofit 2 after pinch analysis (Akgun et al., 2012)
HOT COLD
Equipment
Stream Type Stream
(kg/h) (kg/h)
Cooler 1 H1 Mixture 37064 102 40 CW 350000 30 42 5138
Cooler 2 H2 Mixture 17321 40 10 Freon 3365 2 3 195.4
Cooler 3 H4 Mixture 3000 73 44 CW 22300 30 42 330
Cooler 4 H5 Mixture 511 45 5 Freon 630 -20 -19 39.3
Cooler 5 H6 EDC 108875 91 81.5 CW 443000 30 44 7639.5
Cooler 6 H7 Mixture 19269 81 56 CW 90450 30 44 1559.7
Cooler 7 H8 EDC 870 88.8 40 CW 25900 30 33 95.7
Cooler 8 H9 Mixture 4505 56 10 Freon 5540 1 2 321.7
Cooler 10 H 11 Mixture 114600 125.53 80 CW 390700 30 44 6738.6
Cooler 11 H 13 Mixture 360 85.4 Freon 725 -26 -25 45.9
Cooler 12 H 14 Mixture 10365 -28 -32 Freon 17215 -41 -40 1137.9
Cooler 13 H 15 Mixture 41636 102 65 CW 52500 30 40 646.6
Cooler 14 H 16 Mixture 23907 144.68 60 CW 55100 30 42 814.4
Cooler 15 H 17 VCM 35200 40 39 CW 476800 30 35 2937.7
Cooler 16 H 18 VCM 17728 40 35 Freon 640 1 2 37.2
Cooler 17 H 19 VCM 19000 75.4 35 Freon 5965 2 3 346.2
Cooler 18 H 20 Mixture 49000 81 60 CW 212785 30 45 3932.1
Cooler 19 H 21 Mixture 73000 90 65 CW 402465 30 45 7437.2
Cooler 20 H 22 EDC 7989 44.5 37 Freon 12755 2 3 740.5
Cooler 21 H 23 Mixture 760 37 10 Freon 1325 -1 0 77.3
Cooler 22 H 25 Mixture 50000 68 43 CW 114365 30 38 1127.1
Cooler 23 H 26 Mixture 15576 68 45 CW 58045 30 38 483.7
Cooler 24 H 27 Mixture 6000 45 40 CW 5275 30 34 26
Cooler 25 H 28 Mixture 15081 116.9 50 CW 135000 30 44 2328.3

Heater 1 C1 Mixture 2680 40 140 Steam 120 160 159 69.8

Heater 2 C2 Mixture 11625 -28 145 Steam 855 160 159 497.1
Heater 3 C3 Mixture 16235 60 150 Steam 625 160 159 364.2
Heater 4 C4 C2H2 4467 11 150 Steam 555 160 159 323.8
Heater 5 C8 EDC 55230 100 101 Steam 855 121 120 526.7
Heater 6 C 10 EDC 54000 103 160 Steam 2225 170 169 1279.4
Heater 7 C 11 EDC 58082 160 203.7 Steam 9690 221 220 5059.1
Heater 8 C 13 EDC 937570 159 160 Steam 6830 170 169 3925.4
Heater 9 C 15 Mixture 544000 110 111 Steam 8280 130 129 5061.4
Heater 10 C 16 Mixture 440000 130.978 131 Steam 6867 180 150 76.3

148
Table 5. Exergy calculations on heat exchangers for Retrofit 1
Stream hin hout h0 Sin Sout S0 Ein Eout ∆E ∑Ein ∑Eout ED % ED
EXCHANGERS ε ε (%) % ED OVERALL
(kg/h) (kW) (kW) (kW) (kW) (kW) (kW)
HOT H 24 10818 -736.5 -747.2 -2255 7.144 7.137 5.244 4420.4 4388.7 31.6
HE 1 0.887 88.7 4474.8 4471.2 3.6 0.091 0.039
COLD C1 2680 65.9 106.1 -0,260 4.257 4.360 4.537 54.4 82.5 28
HOT H 24 10818 -747.2 -796.3 -2255 7.137 7.104 5.244 4388.7 4243.7 145.1
HE 2 0.856 85.6 4582.3 4561.4 20.9 0.530 0.228
COLD C2 11625 -2484 -2443 -2534 4.296 4.398 4.694 193.6 317.8 124.2
HOT H 24 10818 -796.3 -872.2 -2255 7.104 7.052 5.244 4243.7 4019.5 224.2
HE 3 0.905 90.5 4547.8 4526.5 21.3 0.541 0.233
COLD C3 16235 -3646 -3598 -3707 4.121 4.242 4.379 304.2 507 202.8
HOT H 24 10818 -872.2 -915.6 -2255 7.052 7.021 5.244 4019.5 3891.4 128.1
HE 4 0.926 92.6 4180.9 4171.4 9.5 0.241 0.104
COLD C4 4467 8838 8940 8712 6.014 6.272 6.177 161.4 280 118.6
HOT H 24 10818 -915.6 -1239 -2255 7.021 6.762 5.244 3891.4 2939 952.4
HE 5 0.918 91.8 5659.1 5580.9 78.2 1.982 0.855
COLD C 10 54000 -1553 -1491 -1680 1.765 1.914 1.399 1767.8 2641.9 874.1
HOT H 24 10818 -1239 -1881 -2255 6.762 5.986 5.244 2939 1068.1 1870.9
HE 6 0.220 22.0 5780.6 4320.7 1459.9 36.987 15.951
COLD C 11 58082 -1491 -1464 -1680 1.914 1.975 1.399 2841.6 3252.6 411
HOT H 11 114600 -841.8 -941.6 -1158 2.634 2.390 1.941 9514.2 6531.4 2982.8
HE 7 0.533 53.3 7719 6326.3 1392.7 35.284 15.216
COLD C 12 317580 -677.9 -658.6 -653.6 1.402 1.453 1.560 -1795.2 -205.1 1590.1
HOT H 16 23907 -1497 -1553 -1680 1.903 1.768 1.399 1131.6 782.1 349.5
HE 8 0.881 88.1 2591.5 2549.9 41.6 1.054 0.455
COLD C 10 54000 -1575 -1553 -1680 1.706 1.765 1.399 1459.9 1767.8 307.9
HOT H 11 114600 -941.6 -962.5 -1158 2.390 2.338 1.941 6531.4 5907.5 623.9
HE 9 0.779 77.9 7951.7 7814.1 137.6 3.486 1.503
COLD C8 55230 -1580 -1546 -1680 1.696 1.788 1.399 1420.3 1906.6 486.3
HOT H 16 23907 -1553 -1579 -1680 1.768 1.724 1.399 782.1 616.8 165.4
HE 10 0.678 67.8 2129.9 2076.6 53.2 1.349 0.582
COLD C9 54000 -1583 -1575 -1680 1.685 1.706 1.399 1347.8 1459.9 112.1
HOT H 11 114600 -962.5 -991.6 -1158 2.338 2.291 1.941 5907.5 5018.5 888.9
HE 11 0,863 86.3 8125.6 8004 121.7 3.083 1.329
COLD C6 89900 -1573 -1540 -1669 1.692 1.783 1.405 2218.2 2985.4 767.3
HOT H 28 15081 -15480 -15530 -15900 4.157 4.046 2.959 1634 1436.2 197.8
HE 12 0,123 12.3 3827.8 3654.3 173.5 4.395 1.895
COLD C6 89900 -1574 -1573 -1669 1.691 1.692 1.405 2193.8 2218.2 24.3
HOT H3 14110 -1576 -1617 -1668 1.684 1.569 1.408 333.5 184.1 149.4
HE 13 0.928 92.8 413.5 402.7 10.8 0.273 0.118
COLD C5 14230 -1763 -1725 -1785 1.503 1.620 1.432 79.9 218.6 138.6
HOT H 10 14736 -1580 -1617 -1680 1.696 1.622 1.399 378.9 235.1 143.9
HE 14 0.715 71.5 601.4 563.9 37.6 0.952 0.411
COLD C7 18736 -1632 -1610 -1680 1.549 1.612 1.339 222.5 328.8 106.3
HOT H6 108875 -1238 -1258 -1680 2.579 2.578 1.399 12475 11871 604.1
HE 15 0.957 95.7 6693.7 6667.9 25.8 0.654 0.282
COLD C 14 98280 324.2 347.2 560.8 0.792 0.865 1.785 -5781.5 -5203.2 578.3
HOT H6 108875 -1258 -1279 -1680 2,578 2.578 1.339 11826 11190 635.1
HE 16 0.719 71.9 12664 12486 178.4 4.519 1.949
COLD C 17 124560 -896.9 -882.4 -919 1.644 1.696 1.729 838.2 1294.9 456.7
HOT H 16 23907 -1569 -1590 -1680 1.724 1.670 1.399 683.2 552.7 130.5
HE 17 0.978 97.8 717 714.1 2.9 0.073 0.031
COLD C4 4467 8727 8838 8712 5.688 6.014 6.177 33.8 161.4 127.6
HOT H 27 6000 -12210 -12260 -12280 3.573 3.088 3.464 112.1 49 63.1
HE 18 0.361 36.1 123.1 82.8 40.4 1.022 0.441
COLD C4 4467 8707 8727 8712 5.622 5.688 6.177 11 33.8 22.8

149
Exergy calculations on heat exchangers for Retrofit 1 are continued…
HOT H 18 17728 264.9 257.1 560.8 0.616 0.591 1.785 -1313.2 -1348.6 35.4
HE 19 0.932 93.2 -1335.2 -1337.6 2.4 0.061 0.026
COLD C4 4467 8678 8707 8712 5.524 5.622 6.177 -21.9 11 32.9
HOT H 28 15081 -15530 -15540 -15900 4.046 4.033 2.959 1436.2 1395.6 40.5
HE 20 0.913 91.3 1453.6 1450.1 3.5 0.089 0.038
COLD C1 2680 12.2 65.9 -0.2603 4.100 4.257 4.537 17.4 54.4 37
HOT H 28 15081 -15531 -15730 -15900 4.033 3.764 2.959 1433.3 627.9 805.5
HE 21 0.883 88.3 2070.1 1975.6 94.5 2.393 1.032
COLD C9 54000 -1634 -1583 -1680 1.541 1.685 1.399 636.8 1347.8 711
HOT H 11 114600 -981.6 -988.3 -1158 2.291 2.274 1.941 5336.9 5137.1 199.8
HE 22 0.881 88.1 5354.4 5330.7 23.7 0.600 0.259
COLD C2 11625 -2543 -2484 -2534 4.117 4.296 4.694 17.5 193.6 176.1
HOT H 23 760 -1249 -1635 -1634 2.716 1.529 1.509 74.9 -0.3 75.2
HE 23 0.983 98.3 18.5 17.2 1.3 0.032 0.014
COLD C2 11625 -2568 -2543 -2534 4.033 4.117 4.694 -56.4 17.5 73.9
HOT H5 511 -1725 -2002 -2137 4.513 3.748 3.026 53.2 16.6 36.6
HE 24 0.955 95.5 -38.2 -39.8 1.6 0.042 0.018
COLD C2 11625 -2580 -2568 -2534 3.986 4.033 4.694 -91.4 -56.4 35
HOT H 13 360 -1235 -1793 -1650 2.578 1.242 1.396 38.5 -13.9 52.5
HE 25 0.840 84.0 -96.9 -105.3 8.4 0.212 0.092
COLD C2 11625 -2595 -2580 -2534 3.932 3.986 4.694 -135.5 -91.4 44.1
HOT H 11 114600 -988.3 -993.2 -1158 2.274 2.262 1.941 5137.1 4990.7 146.4
HE 26 0.986 98.6 5297 5294.9 2.1 0.054 0.023
COLD C3 16235 -3679 -3646 -3707 4.081 4.121 4.379 159.9 304.2 144.3
total = 95507 91559 3947

Table 6. Exergy calculations on heaters for Retrofit 1

Stream hin hout h0 Sin Sout S0 Ein Eout ∆E ∑Ein ∑Eout ED % ED
HEATERS ε ε (%) % ED OVERALL
(kg/h) (kW) (kW) (kW) (kW) (kW) (kW)
Heater HOT Steam 11000 -13110 -14790 -15810 6.286 2.519 0.364 7797.6 2952.1 4845.6
0.902 90.2 11050 10575 474.5 38.780 5.184
1 COLD C 11 58082 -1464 -1178 -1680 1.975 2.578 1.399 3252.6 7623.7 4371.1
Heater HOT Steam 8000 -13150 -15200 -15810 6.657 2.041 0.364 5561.5 1262.4 4299.1
0.915 91.5 50923 50556 366.5 29.955 4.005
2 COLD C 13 937570 -1493 -1477 -1680 1.912 1.948 1.399 45362 49294 3932.6
Heater HOT Steam 9000 -13210 -15410 -15810 7.116 1.528 0.364 6078 927.3 5150.8
0.934 93.4 21896 21554 341.6 27.922 3.733
3 COLD C 15 544000 -1550 -1516 -1663 1.741 1.828 1.408 15818 20626 4809.1
Heater HOT Steam 5900 -13190 -15330 -15810 6.917 1.739 0.364 4025.4 730.3 3295.1
0.988 98.8 17552 17511 40.9 3.343 0.447
4 COLD C 16 440000 -995.8 -967.4 -1115 1.470 1.541 1.129 13527 16781 3254.2
total = 101421 100198 1224

150
Table 7. Exergy calculations on coolers for Retrofit 1
Stream hin hout h0 Sin Sout S0 Ein Eout ∆E ∑Ein ∑Eout ED % ED
COOLERS ε ε (%) % ED OVERALL
(kg/h) (kW) (kW) (kW) (kW) (kW) (kW)
Cooler HOT H1 37064 -3827 -4309 -4308 4.465 3.072 3.211 4629.4 25.5 4603.9
0.965 96.5 7368.6 7207.8 160.9 4.040 1.758
1 COLD CW 350000 -15870 -15820 -15900 3.032 3.204 2.959 2739.2 7182.3 4443.1
Cooler HOT H2 17321 -3552 -3595 -3560 4.060 4.095 4.492 90.5 -120.6 211.1
0.842 84.2 -100.6 -134 33.4 0.838 0.365
2 COLD Freon 3365 -5749 -5540 -5520 1.361 2.115 2.343 -191.1 -13.4 177.7
Cooler HOT H4 3000 -1379 -1793 -1823 2.700 1.521 1.425 343.4 23 320.4
0.885 88.5 518.2 481.2 37 0.928 0.404
3 COLD CW 22330 -15870 -15820 -15900 3.032 3.204 2.959 174.8 458.2 283.5
Cooler HOT H6 108875 -1259 -1550 -1680 2.578 1.783 1.339 11796 3595.9 8199.6
0.825 82.5 15263 13831 1431.6 35.952 15.641
4 COLD CW 443000 -15870 -15810 -15900 3.032 3.232 2.959 3467.1 10235 6768.1
Cooler HOT H7 19269 -1168 -1459 -1544 2.829 1.997 1.724 1864.7 418.4 1446.2
0.794 79.4 2572.6 2274.5 298 7.485 3.256
5 COLD CW 90450 -15870 -15820 -15900 3.032 3.204 2.959 707.9 1856.1 1148.2
Cooler HOT H8 870 -1261 -1656 -1680 2.579 1.476 1.399 94.1 5.3 88.8
0.721 72.1 296.8 272.1 24.8 0.622 0.271
6 COLD CW 25900 -15870 -15860 -15900 3.032 3.076 2.959 202.7 266.7 64
Cooler HOT H9 4505 -1188 -1445 -1544 2.858 2.153 1.724 410 110.5 299.6
0.977 97.7 95.4 88.5 6.9 0.174 0.076
7 COLD Freon 5540 -5749 -5540 -5520 1.361 2.115 2.343 -314.6 -22 292.6
Cooler HOT H 11 114600 -983.2 -1158 -1158 2.262 1.823 1.941 5309 93.9 5215.1
0.982 98.2 8467 8374.1 92.9 2.333 1.015
8 COLD CW 403500 -15870 -15820 -15900 3.032 3.204 2.959 3157.9 8280.2 5122.2
Cooler HOT H 14 10365 -2578 -2973 -2519 3.936 2.318 4.684 -116 -1136.8 1020.8
0.997 99.7 -1302.3 -1305.5 3.2 0.080 0.035
10 COLD Freon 17360 -5795 -5559 -5520 1.183 2.182 2.343 -1186.3 -168.7 1017.6
Cooler HOT H 15 41636 -677.9 -734.4 -653.6 1.401 1.245 1.560 -235.1 -843.4 608.3
0.873 87.3 175.8 98.3 77.5 1.947 0.847
11 COLD CW 52500 -15870 -15830 -15900 3.032 3.176 2.959 410.9 941.7 530.8
Cooler HOT H 16 23907 -1590 -1634 -1680 1.670 1.542 1.399 552.7 281.7 270.9
0.965 96.5 755.2 745.8 9.4 0.235 0.102
12 COLD CW 25870 -15870 -15830 -15900 3.032 3.176 2.959 202.5 464 261.6
Cooler HOT H 17 35200 563.8 263.3 560.8 1.571 0.611 1.785 81.6 -2621.9 2703.6
0.890 89 3813.3 3516.9 296.4 7.443 3.238
13 COLD CW 476800 -15870 -15850 -15900 3.032 3.105 2.959 3731.6 6138.8 2407.2
Cooler HOT H 19 19000 324.3 257.1 560.8 0.792 0.591 1.785 -1117.2 -1445.3 328.1
0.960 96 -1455.9 -1469 13.1 0.328 0.143
14 COLD Freon 5965 -5749 -5540 -5520 1.361 2.115 2.343 -338.8 -23.7 315.1
Cooler HOT H 20 49000 -873.9 -1163 -1202 2.407 1.553 1.431 4133.7 489.3 3644.4
0.903 90.3 5799 5446.9 352.1 8.843 3.847
15 COLD CW 212785 -15870 -15810 -15900 3.032 3.204 2.959 1665.3 4957.6 3292.3
Cooler HOT H 21 73000 -1245 -1611 -1663 2.583 1.570 1.408 7880.4 972.3 6908.1
51 9376.8 32 0.901 90.1 11030 10349 681.1 17,105 7,442
16 COLD CW 402465 -15870 -15810 -15900 3.032 3.204 2,959 3149.8 6227
HOT H 22 7989 -1297 -1630 -1680 2.608 1.560 1,399 48
782.9 76
102 680.8
Cooler
0.990 99.0 58.5 51.4 7.1 0,179 0,078
17 COLD Freon 12755 -5749 -5540 -5520 1.361 2.115 2.343 724.4 -50.7 673.7
Cooler HOT H 25 50000 -12520 -12620 -12680 3.882 3.513 3.408 2057.6 796.9 1260.7
39 1755.9 0.683 68.3 2952.7 2552.9 399.9 10,041 4,369
18 COLD CW 114365 -15870 -15840 -15900 3.032 3.148 2.959 895.1 860.9
HOT H 26 15576 -2387 -2499 -2574 5.572 5.327 5.179 79
766.6 308.51 458.1
Cooler
0.954 95.4 1220.9 1199.7 21.14 0,531 0,231
19 COLD CW 58045 -15870 -15840 -15900 3.032 3.148 2.959 454.3 891.2 437
Cooler HOT H 28 15081 -15630 -15780 -15900 3.764 3.315 2.959 1046.7 465.4 581.4
68 1086.9 0.939 93.9 1588 1552.4 35.7 0,897 0,390
20 COLD CW 69165 -15870 -15838 -15900 3.032 3.176 2.959 541.3 545.63
47 5 total = 59116 55134 3981

151
Table 8. Comparison of existing process and alternative cases
PINCH RESULTS
Existing Process Retrofit 1 Retrofit 2
Heat Loads
54208 44531.3 46169.3
25217 15551.7 17183.3
270 9934.1 8302.6
Number of Equipment
# Heat Exch. 2 26 9
# Heaters 15 4 10
# Coolers 26 20 25
Umin 43 50 44
Saving of Utilities (%)
Cold 0.5 18.26 15.25
Hot 1.06 38.98 32.58
Total Heat Recovery (%) 0.34 12.51 10.45

Table 9. Comparison of Retrofit 1 and 2 after exergy analysis

EXERGY RESULTS
Retrofit 1 Retrofit 2
OVERALL OVERALL
Equipment ∑Ein (kW) ∑Eout (kW) ED (kW) % ED Equipment ∑Ein (kW) ΣEout (kW) ED (kW) % ED
Exchangers 95507 91560 3947 43.13 Exchangers 63254 61443 1810 27.74
Coolers 59116 55134 3982 43.51 Coolers 59745 55622 4122 63.16
Heaters 101421 100198 1224 13.37 Heaters 340031 339437 593 9.10
OVERALL 256044 246891 9153 100 OVERALL 463030 456504 6526 100
% ED 3.57 % ED 1.41

Acknowledgement
The technical support of PETKİM and KARSAN are sincerely acknowledged

NOMENCLATURE
mass flow rate, kg/h Greek Letters
heat load, kW ε exergetic efficiency
Cp specific heat, kJ/kg°C Δ difference
CP heat capacity flow rate, kW/°C Σ total
T temperature, ºC Subscripts
h enthalpy, kJ/kg rec recovered
S entropy, kJ/kg°C D destroyed
P pressure, bar c cold utility
HE heat exchanger h hot utility
HEN heat exchanger network min minimum
PA pinch analysis in inlet
H hot stream out outlet
C cold stream Superscripts
U utility 0 reference condition
E exergy, kW PH physical
V velocity, m/s CH chemical
2
g gravitational acceleration, m /s KN kinetic
Z height, m PT potential

152
REFERENCES

Raskovic, P. and Stoiljkovic, S. 2009. Pinch design method in the case of a limited number of process streams.
Energy 34:593-612

Yoon, S.G., Lee, J. and Park, S. 2007. Heat integration analysis for an industrial etylbenzene plant using pinch
analysis. Applied Thermal Engineering 27:886-893

Wan Alwi, S. R. and Manan, Z. A. 2010. STEP-a graphical tool for simultaneous targeting and design of heat
exchanger network. Chemical Engineering Journal 162:106-121

Dincer, I. and Cengel Y. A. 2001. Energy, entropy and exergy concepts and their roles in thermal engineering.
Entropy 3:116-149

Akgun N., Ozcelik Z. and Kuryel B. 2012. Debottlenecking and retrofitting by pinch analysis in a chemical plant. 6 th
International Ege Energy Symposıum and Exhibition Proceedings 6:214-228

Dincer, I and Rosen, M. A. 2007. Exergy: Energy, Environment and Sustainable Development. Elsevier

Rivero, R. 2002. Application of the exergy concept in the petroleum refining and petrochemical industry. Energy
Conversion and Management 43:1199-1220

AspenTech Software, Aspen Plus 11.1, Aspen HYSYS 2006.5

APPENDICES
Exergy calculations of the 1st heat exchanger in existing process are shown below:

153
 COMPARISON OF THE ENERGY PERFORMANCE AND STATIC LOADS ON A
BUILDING INTEGRATED WIND, SOLAR AND RAIN WATER HARVESTER
a b b
A. Gurbuz , U. Akbulut , M. Ozsaban
a
Civil Engineering Department, Faculty of Engineering, Recep Tayyip Erdogan University, Rize, Turkey
b
Mechanical Engineering Department, Faculty of Engineering, Recep Tayyip Erdogan University, Rize, Turkey
ali.gurbuz@erdogan.edu.tr, ugur.akbulut@erdogan.edu.tr, mert.ozsaban@erdogan.edu.tr

ABSTRACT
In recent years, sustainable energy and green buildings are developed as important issues. Many helpful systems
were developed to profit from advantages of different renewable energy sources. In the other point, the research
on dissadvantages of sustainable energy systems is very rare and papers focus on only economic issues. In this
paper, an innovative building integrated wind, solar and rain water harvester (WSRH) system analysed by
nonlinear static pushover analysis. The results are compared in terms of energy performance and its effects on
the structural system

INTRODUCTION
The primary energy supply is not limitless and the energy cost is increasing day by day determined the reduction
of the energy requirements to the residential buildings (Koroneos and Kottas, 2007). Apartment buildings are the
most common residential buildings in Rize. Residential building’s annual power consumption is approximately 37
2 2
kWh/m (Choi, 2012). In this study, 165 m four flat and 17 stories 50 meter building have analyzed. This building
have 415.140 MWh annual power consumption.

Alternative energy resources such as wind energy and solar energy have extensively invested interest by the
researchers. The common drawbacks to wind and solar energy options are their unforeseeable nature and they
rely on the climate changes (Kimura et al., 1996). Both wind and solar energy sources have strong relations with
the environment and climate conditions. That’s why to get a stable power output it would be necessary to
construct a hybrid power system (Yang et al., 2009). But, the problems can be partially overcome by integrating
two or more resources in a proper combination to form a hybrid system. For certain locations, the hybrid solar–
wind power generation systems with storage banks offer a highly reliable source of power, which is suitable for
electrical loads that need higher reliability.

Mostly, electricity is produced in rural power plants and transferred to urban areas. However, in developing
countries, grid extension is impractical because of dispersed population, rugged terrain or some more other
reason. Thus, on-site renewable energy systems have an important role for narrowing the electricity gap in rural
areas (Saheb-Koussa et al., 2009). The concept of on-site renewable energy generation is to extract energy from
renewable sources close to the populated area where the energy is required. The proposed wind–solar hybrid
renewable energy system in this paper is a new, feasible and compact design which can be built on the top of
high-rise buildings.

In the other point, the researches on these kinds of systems mostly focus on only energy economic issues and
skip disadvantages. In this paper, an innovative building integrated wind, solar hybrid system and rain water
harvester(WSRH) analyzed by nonlinear static pushover analysis. The results are compared in terms of energy
performance and its effects on the structural system.

As for static analyses, It is possible to use linear or nonlinear methods to determine seismic behavior of
structures. However, linear analysis uses the methods of the elastic solution; inelastic behavior includes to
solution by specific coefficients. But linear analyses results are lower realistic than nonlinear analyses. It is need
to include inelastic behavior of structural elements for more realistic results. Nonlinear time history (TH) analysis
which earthquake loads are applied to the building directly is the represents the most actual behavior of the
structure. Nevertheless, TH analyses need to long time period (Dervisoglu, 2007). Earthquake data should be
selected carefully. Past studies show that nonlinear pushover (NSP) analysis is suitable alternative to TH analyze.
(Kappos and Kyriakaki, 2000; Bhatt and Bento, 2012). Studies of (Saiidi and Sozen, 1981) provided static
pushover analyses occurrence and improvement. In addition, (ATC40, 1996)], (FEMA 356, 2000), (FEMA 440,
2004) and (TEC, 2007) procedures were introduced.

In this study, NSP analyze applied to a reinforced concrete residential building with 17 floors. Later, weight of the
wind, solar and rainwater harvester system added on the roof and analyze repeated. Energy studies are
calculated for 4 different forms.

Different forms are:

- Existing building integrated solar panels,

- Existing building integrated wind turbine,
- Existing building integrated rainwater harvester system
- Existing building with solar, wind and rainwater harvester hybrid system.

154
Additionally; water storage tank of rainwater collector placed on the roof to deal unfarobile situation. Results were
discussed in terms of base reactions, lateral displacements and energy performances.

2. SOLAR, WIND AND RAINWATER HARVESTER SYSTEMS

2.1 Solar Energy System

Based on the solar radiation data of Rize (GEPA, 2013), the solar energy generation is estimated using Eq. 1.

(1)

where, is annual mean daily global irradiation, is module conversion efficiency and is solar power loss.
Table 1 lists the value for each coefficient in Eq. 1. Thus, the estimated daily solar energy generated is
145 kW h/day and 52.925 MWh/year.

Table 1 Estimated solar energy generated in Rize

-2
Annual mean daily global irradiation, Gs (kWhm /day) 3.36
2
Solar panel area above the PAGV (m ) 520
2
Estimated solar cell active area, As (m ) 450
Efficiency of solar panel ɳPS 0.12
Estimated power loss (electric transmission), K 0.8
Estimated solar energy generated, Esolar (kW h/day) 145

2.2 Wind Energy System

In this study, hub height of the VAWT is assumed as 50 m. The average wind speed data of Rize for 50m is
used to calculate the wind power extraction of the VAWT. The free-stream wind speed is 4 m/s at Rize
(REPA, 2013).

PW can be calculated using Eq. 2,

(2)

where is power coefficient, is efficiency due to bearing and transmission loss, is generator
efficiency, is efficiency due to power-augmentation-guide-vane (PAGV) loss, is air density, is
swept surface area of wind turbine and is wind speed (Chong et al., 2012).

With the estimated efficiency of sub-systems or components , the wind energy generation could be
estimated as shown in Table 2. For a system 24 m diameter and 12 m high VAWT installed on top of a
building of 50 m height, the wind energy generated is approximately 103.842 MW h/year or 284.5 kW
h/day.

Table 2 Estimated parameters (efficiency, losses, etc.) of various components in wind energy system.
Generator efficiency, ɳG 0.7
Efficiency due to bearing and transmission loss, ɳTR 0.9
Efficiency due to PAGV loss, ɳPAGV 0.8
Air density, ρ (kg/m3) 1.225
Frontal swept surface area, AWT (=12×24 m2) 288
Power coefficient, Cp 0.05

Table 3 Summary of estimated annual energy savings.

Estimated energy source Energy generated/saved
Wind energy generated 284.5 kW h/day
Solar energy generated 145 kW h/day
Total estimated annual energy generated or saved 157 MW h/year

2.3 Rainwater Harvester System

The amount of rain water collected ( ) is estimated by using the Rational Equation as in Eq. 3

(3)
2
where assumed rain catchment area and runoff coefficient is 720 m and 0.75 respectively. Rainfall
intensity (I) is 2245 mm/month.

155
3. NONLINEAR STATIC PROCEDURE

Seismic analyses of structures may perform by linear or nonlinear methods. Linear analysis are based on first
order linear elastic theory and inelastic behavior includes to the solution by specific coefficients. Elastic analyses
results are lower realistic than inelastic analyses. It is need to include inelastic behavior of structural elements for
more realistic results.

3.1 Nonlinear Static Pushover Analyze

Nonlinear static pushover (NSP) analysis is credible way to analyze seismic behavior of structural elements of
buildings(Saidi and Sozen, 1981). In addition, (ATC 40, 1996), FEMA 356, 2000), (FEMA 440, 2004) and Turkish
Seismic Code(TSC, 2007) procedures were introduced NSP analyses.

Capacity curves of the structure were obtained at the end of the analysis. FEMA 356 values have been used for
cracking stiffness of the structural system elements. Plastic deformations where in plastic sections and building
material behaves linear elastic in other areas have been adopted. In the other hand, plastic deformations form by
bending moment in beams, bending moment and axial forces in columns have been adopted. (ATC 40, 1996)
values were used for plastic rotations and plastic hinge values.

3.2 Modeling Issues

In this study, NSP performed on 4 different forms of 17-story RC building integrated WSRH using Sap2000
structural analyses software. Results were discussed in terms of base reactions, lateral displacements and top
displacement rates.

All NSP analyses performed by using Sap2000 nonlinear analysis program. Columns of building are defined as
reinforced concrete elements which work for axial load and moment. As for beams, they defined reinforced
concrete element which working to moment. C20 Concrete and S420 still classes used according to Turkish
Standard (TS500). Sap model of the building is shown in figure 1

Figure 1. 3D View of the building for analyses

4. CASE STUDY

Our sample building discussed at four different forms. These are:

4.1. Existing Building

17 storey reinforced concrete building which designed according to the 2007 Turkish Seismic Code selected for
this study. All of the floors have the same height of 2.80 m but it of base floor is 5.00 m. The building has 30x60

156
cm beams. Columns are 35x90 cm on base story and all of other column are 30x60cm. The stirrups have a
diameter of 8 mm with 20cm spacing constant along the height. There is no confinement for stirrups. The mass of
17th story is calculated to be 268 tons, base story calculated to be 520 tons and all of other stories are calculated
tons. The building plan is shown in Figure 1.

(a) (b)
Figure 2 (a) Plan view (cm); (b) lateral view (m)

4.2. Existing Building Integrated Solar, Wind Hybrid System

100kW vertical axes wind turbine(VAWT) selected for simple building. The turbine specifications are shown in
table 4.

Table 4. 100 kW wind turbine specification

Manufacturer Astralux Wind Turbine and Power Equipment
Co. Ltd. – Genova / Switzerland
Rotor height 12 m
Rotor diameter 24 m
Total height 22m
Minimal working speed 1.5 m/s
Power on wind speed 4 m/s- 15 kW
Nominal working speed 9.5 m/s
Area occupied will be design to fit the roof.
Nominal working speed 9.5 m/s
Constraction materials Aluminum, steel, concrete and composite plastics
Total weight 35t

These types of wind turbines are suitable for buildings owing to slow and silent working abilities. They can be built
on roof reinforced concrete buildings because of they are made of safe and durable materials. Otherwise,
architectural view of building is important on turbine type. VAWT can be made of same material with the roof and
be integrated to the building easily.

Cut-in wind speed 1.5 m/s and nominal working speed 9.5 m/s and the turbine rotates slowly and silently.
2
Additionally, 400 pieces 1580x808x45mm solar cells putted on the system. 520m solar cells area obtained
totally. Solar cell specifications are shown in table 5. Total weight of 400 pieces solar cells are 6.2 tons.

157
Table 5 Solar cell specification
Materyal Mono-SiliconMono-Silikon
Maximum Power 180 W ± 3 % 180 W ± 3 %
Voltajı (V) 44.4 V / 5,08
44.4
A V / 5,08 A
Dimentions (H x W x D) 1580 x 8081580
x 45 mm
x 808 x 45 mm
Weight ≈ 15,5 kg ≈ 15,5 kg
Work Temperature -45 +85 °C-45 °C den +85 °C ye kadar

Figure 3. Sample building with roof type wind turbine

4.3. Existing Building With Solar, Wind And Rainwater Harvester Hybrid System.

Rainwater harvester(RWH) system designed on 2 different ways in the study. Water storage tank placed at the
base floor in the first design and water has distributed by a water pump. At the second design; water storage tank
placed on the roof.

Table 6. Annual Mean Rainfall Rates in Turkey

Average Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sept. Oct. Nov. Dec. Annual
rainfall total
period rainfall
1970-2010 77.6 71.2 64.4 60.0 48.3 32.6 18.9 18.2 27.6 59.7 75.3 88.8 642.8

Figure 4 Annual total rainfall in Turkey (1970-2010)

158
 Table 7. Rize, Annual Mean Rainfall Rates
Average Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sept. Oct. Nov. Dec. Annual
rainfall total
period rainfall
1970- 210.4 179.2 144.4 92.2 99.7 135.8 148.4 182.6 249.6 300.4 256.2 246 2244.9
2010

Figure 5 Annual total rainfall in Rize (1970-2010)

However, total annual mean rainfall 642.8mm in Turkey(Figure 4), This rate increases to 2244,9mm for Rize
City(Figure 5). In this study, rain collector area is 24x28m2 on the roof. In other words; 1508m3(tons) rain falls to
the roof area in a year. This proportion corresponds to over 4 tons of water per day. According to the results of
Municipal Water Statistics Survey, water abstraction per capita is about 217 liters per day in Turkey(TSI, 2012).
RWH can supply all water demand of 19 persons for sample building in this study. This amount is equal to %7.5
of the whole demand for the building.

A water tank with a storage capacity of 30 tons has placed on the roof for first design. Then, water thank has
designed on the base floor and added a water pump as an alternative solution.

6. RESULTS AND DISCUSSION

The results of the building are presented in terms of base reaction, lateral displacement, hinges results and
strains on the roof story. The capacity curve which obtained as a result of the NSP analyze is shown in Figure 6.
NSP analyze was perform on weak direction of the building(Y direction). Building has a symmetric columns and
rectangular diameters. It is expected that the building is weaker on Y direction about seismic capacity.

4500
4000
3500
Base Shaer (t)

3000
2500
2000
1500
1000
500
0
0 50 100 150 200 250 300 350
Top Displacement (mm)

EB WSH WSRH

Figure 6 Capacity Curves for Weak Direction

When this 3 capacity curves were compared, It was observed that the EB has high capacity as expected. It is
shown that; differences of the results are not as high as it effects of seismic performance. WSH and WSRH
results in agreement with EB results for inelastic part of the graphic. As for elastic part of graphi c, difference of
EB results with WSH and WSRH results grew a bit.

159
Residential buildings must be “Life Safety”(LS) or better performance to Turkish Earthquake Code 2007.(TEC) In
our study, hinge results shows that; all forms of the building provide “LS” level. Figure X-x and x shows hinges
results for EB – WSH and WSHR

Figure 7 Hinges results for EB

Even columns have larger area on the base story; the biggest damage was observed on the “A” axes columns of
base story. Because, this columns are higher than other stories and they compensate much more moment
values. Plastic hinges usually occurred on the edge of columns at base story. Other critical elements were in
middle stories. There are rarely “LS” level hinges and mostly “IO” level hinges occured in middle stories.

Figure 8 Hinges Results for WSRH

Hinges results for WSRH are shown in figure. Extra loads on the roof changed the building period and base story
damage were decreased. Cracks on the middle stories absorbed some of energy and lower energy transferred to

160
base stories. But middle stories had more damage than EB. Usually LS level of plastic hinges occurred on middle
story columns.

7. CONCLUSIONS

In this study, we have presented comparison of the energy performance and static loads on a building integrated
WSRH. Databases of REPA, GEPA and TSMS (Turkish State Meteorological Service) are used in the energetic
analysis and some concluding remarks from this study as follows:

(a) Electricity generated from wind is 104 MW h /year.

(b) Electricity from solar cell is 53 MWh/year.
(c) Total electricity from the WSRH is 157 MWh/year.
3
(d) Water collected by the rain harvester is 1178 m /year.
(e) Annual electric consumption of the building is 415 MW h /year.
3 3
(f) Annual water consumption of the building is 12240 m /year (15m /month.flat)
(g) Electricity compensation ratio from renewable energy is %37.8.
(h) Water compensation ratio from rain harvesting is %9.6.

The results of the building are presented in terms of base reaction, lateral displacement, hinges results and
strains on the roof story. The capacity curve which obtained as a result of the NSP analyze performed on weak
direction of the building(Y direction When this 3 capacity curves were compared, It was observed that the EB has
high capacity as expected. Differences of the results are not as high as it effects of seismic performance. WSH
and WSRH results in agreement with EB results for inelastic part of the graphic. As for elastic part of graphic,
difference of EB results with WSH and WSRH results grew a bit.

REFERANCES

Koroneos, C. and Kottas, G. (2007) ‘Energy consumption modeling analysis and environmental impact
assessment of model house in Thessaloniki—Greece’, Building and Environment Vol. 42, pp. 122–138

Choia, I.Y., Choa, S.H., Kimb, J.T. (2012) ‘Energy consumption characteristics of high-rise apartment buildings
according to building shape and mixed-use development’, Energy and Buildings, Vol. 46, pp. 123–131.

Kimura, Y., Onai, Y., Ushiyama, I. (1996) ‘A demonstrative study for the wind and solar hybrid power system’,
Renew Energy Vol. 9, pp. 895–898.

Yang, H., Zhou, W., Lou, C., (2009) ‘Optimal design and techno-economic analysis of a hybrid solar–wind power
generation system’ Applied Energy Vol. 86, pp. 163–169.

Saheb-Koussa, D., Haddadi, M., Belhamel, M., (2009) ‘Economic and technical study of a hybrid system (wind-
photovoltaic-diesel) for rural electrification in Algeria’, Applied Energy Vol. 86 pp. 1024–1030.

Dervişoğlu, Z. (2007) ‘Comparison Of Performance Evaluation Methods Of Reinforced Concrete Buildings Under
Earthquake Effects Within Non Linear Theory Framework, Master Thesis, Department of Civil Engineering,
Balıkesir University, Balıkesir

Kappos, A. J. and Kyriakakis, P. (2000) ‘A re-evaluation of scaling techniques for natural records’, Soil Dynamics
and Earthquake Engineering Vol. 20, pp. 111-123.

Bhatt, C., Bento, R., (2012) “Comparison of Nonlinear Static Methods for the Seismic Assessment of Plan
Irregular Frame Buildings with Non Sesismic Details”, Journal of Earthquake Engineering, Vol. 16 pp. 15-39.

Saiidi, M. and Sozen M.A. (1891) ‘Simple nonlinear seismic analysis of R/C structures’, Journal of the Structural
Division, ASCE Vol. 107, pp. 937–51.

Chong, W.T., Fazlizan, A., S.C. Poh, S.C., Pan, K.C., Ping, H.W. (2012) ‘Early development of an innovative
building integrated wind, solar and rain water harvester for urban high rise application’, Energy and Buildings
Vol.47 pp. 201–207

MUNICIPAL WATER STATISTICS, 2010, T.C Turkish Statistical Instute, Press Release, No:10753, 24/02/2012,
10:00

161
Repa website: Http://www.eie.gov.tr/YEKrepa/RIZE-REPA.pdf, 2013.

ATC-40, Seismic evaluation and retrofit of concrete buildings vols. 1–2. California. Applied Technology Council,
1996.

FEMA-356, Prestandard and commentary for seismic rehabilitation of buildings, Federal Emergency
Management Agency, Washington, 2000.

FEMA-440, Improvement of Nonlinear Static Seismic Analysis Procedures, Washington, 2004.

TSC-2007, Specifications for buildings to be built in seismic areas. Turkish Seismic Code 2007. Ministry of Public
Works and Settlement. Ankara. (Turkey).

162
 INVESTIGATING UTILITY OF LATENT HEAT STORAGE SYSTEMS FOR GREEN BUILDING
APPLICATIONS
a b a
D. Aydin , Z.Utlu , O. Kincay
a
Yildiz Technical University, Beşiktaş, 34349 İstanbul, Turkey
b
İstanbul Aydın Üniversity, Florya, İstanbul Turkey
devrimayd@yahoo.com, zaferutlu@aydin.edu.tr, okincay@yahoo.com

ABSTRACT
In green building applications,highest energy demands are needed for heating and cooling to provide
comfort conditions. This study is done by using real datas obtained from a prototype structure, built as part
of a project. Necessity of heat storage systems, during the usage of solar energy and ground sourced heat
pump systems for heating are investigated. By this way different techniques used for heat storage are
researched and compared. Efficiency of latent heat storages constituted by utilizing latent heat of phase
change material is analysed. Thermodynamic analyses of the heating system works with solar panels,
ground sourced heat pump and a latent heat storage which uses paraffin as a phase change material are
done.Lastly for various ground temperature and solar radiation, changes in amount of heat storage in heat
storage system is analysed. For six month heating period, analyses show that especially in winter,energy
consumption in compressor shows a sharp rise.For this reason utilizing period from heat pump with low
energy consumption is too short.This study focuses on energy generation from solar panels due to time.
Thereby, supply ratios of hourly energy demand of the inspected building are investigated.
Key Words: Energy storage, Renewable energy, Green building
1.INTRODUCTION
Due to the technological and industrial development and the increase in human population, environmental
pollution getting worser. This problem increase tendency to the renewable energy sources.Because of the
public buildings and industrial corporations are generally located at city centres, migration to the city centres
increase, for this reason air pollution and unplanned urbanisation occurs.The main results of unplanned
urbanisation are, structuring without any planning about energy systems.This situation ends up with sharp
rises in energy consumptions.On the contrary, with recent studies, building energy consumpion rates were
tried to get undercontrol. Concordantly ‘energy performance certificate’ given to the buildings, by this way
reducing energy consumption and increasing utility of renewable energy systems are aimed and tried to
encourage.
In addition to this, rise of the amount of greenhouse gas and price of fossil fuels are being the driven force on
the way of benefitting from various renewable energy resources[Benli at al., 2009]. For most people all
around the world, solar energy is evaluated as a primary energy source for the future. Scientists are
maintaining their research about new and renewable energy resources.But improving energy storage
systems is as important as using new energy resources. Storage of energy in an appropriate way and
devoloping systems to become it available in the case of necessity is the key misson of todays’ researchers.
Storing energy in a suitable way not only providing an equilibrium between supply and demand also by
increasing performance and utility of energy systems contribute to energy conservation. Energy storage
systems providing an economical usage of fuels, as a result of this reducing the amount of waste energy.
Thus this systems improve utility of energy systems in an economical way[Sharma at al., 2009].
Nowadays fossil fuels, that national energy systems dependent to, are limited and predicted to be an
insufficient and expensive source in the future.In addition to this, common energy demand rises day by day.
Beside,, depends on the rise in energy consumption, environmental impacts; such as climate change and
atmospherical pollution are increasing.Greenhouse emissions have been attracking attention as primary
factor of climate change. Within this context, various international organisations such as Kyoto Protocol
studying on this problem[Zhou at al., 2012].
Developing heat storage systems have a vital importance on achieving advanced level and efficient energy
systems.
2.ENERGY CONSUMPTION AND GREEN BUILDINGS
Global energy generation and consumption amounts have a vital importance in each step of economical
development, for energy conservation and using it efficientlly. Global energy consumption related with

163
economical and environmental improvement. For this reason energy consumption factor necessitate
considering design of energy-saving and efficient systems as well as environmental factors. Depends on the
economical development, main consumpion fields of energy are structuring, environmental improvement,
commercial, industrial, transportation and buildings.In recent years demand of structuring has been heavily
increasing. Huge buildings constructed for meeting this demand, has an important effect on energy
consumption. In developed countries hıghest energy consumption is in building sector with %27 (Fig 1). Also
%70 consumption of electricity energy and atmospherical emmisions including greenhouse gas sourced
from building sector.Determination of energy consumption rates and energy loses are two important
necessities for designing energy efficient buildings [Paremeshwaran at al., 2012].

Fig. 1. Energy consumption break-up by sector wise [Paremeshwaran at al., 2012]

Green Buildings are known as sustainable buildings. Green building applications aimed to meet the %75-80
of thermal and electrical demand from renewable energies. Whereas in classical structures %15-20 percent
of energy demand supplied from renewable energies. Nowadays green building applications have been
maintaining in a rising trend. This applications paid attention to not only active technologies (mechanical
cycle) but construction architecture, solar radiation, green fields and consturuction materials for decreasing
energy consumption and greenhouse gas emmisions. In addition to this, renewable energy sources using
for supplying energy to buildings are cyclical.For this reason, this systems sometimes cannot meet the
energy demand of building or sometimes provide more energy then the needing.When all these causes
taking into consideration, it is clear that improving energy storage systems will increase utility of renewable
energy sources.
Recently, ın order to providing usage of energy in an efficient way, smart building management systems had
been developed. In addition to this various disciplines of engineering in such fields of lighting, air
conditioning, cooling, vantilating and heating have been improved. By this way suitable thermal conditions,
hıgh quality indoor air and sufficient lighting provided. Thus appropiate comfort conditions achieved. Despite
usage of energy efficiency enhancement techniques increase capital costs, in a long time period with the low
energy consumption amounts, it meets the capital costs and provides profitability. Briefly a well designed
building is a construction which meets the comfort demand of people living in it with minimum energy
consumption[Paremeshwaran at al., 2012].

Fig. 2.Avarage break-up of energy usage and losses in buildings[Paremeshwaran at al., 2012.]

164
3. LATENT HEAT STORAGE SYSTEMS IN BUILDINGS
Recently, due to the increasing energy demand and environmental factors interest and request on energy
storage systems have been enhanced. As it can be seen from the diagram above, building sector is one of
the leading sectors in energy consumption. Although there are lots of ways for decreasing energy
consumption in buildings, a system which has a potential of compensating peak thermal load demand of a
building is a necessity. Energy storage systems store heat during the poriod which building has a low heating
or cooling energy demand, for use it during peak thermal load conditions. Heat storage systems integrated
with thermal systems provides usage of thermal energy in an efficient way as well as compensating unstable
energy demands[Koca at al., 2008].
3.1. Thermal Energy Storage Methods In Buildings
Heat storage systems, depends on their dimensions, applied method and heat storage material, can applied
in low temperature for a short time or a long time period in any application. In case of the difference between
solar energy and energy demand of building is low , short time heat storage applied. Using long time period
heat storage to supply seasonal energy demand with solar energy, increase contribution of solar energy in
the process of compensating overall energy demand. By considering thermal energy generating prices and
utilization from solar energy , will be decided to use whether short time storage or long time storage[Arda.,
2006].
Selection of heat storage method mainly depends on heat storage period, economical applicability and
operating conditions. For any application efficiency and economy of heat storing not only depends on heat
storage method but also design of the system. For a heat storage system design, these factors must be
considered;[Öztürk at al., 2007]
1.Heat storage capacity of unit mass and volume of heat storage material
2.Heat storage, recovery temperature and method.
3. Temperature gradient.
4. Dimensions of system components.
5. Material, shape and dimensions of heat storage tank.
6. Properties of the environment that the stored energy used.
7. Total cost of storage system.

Mainly there are three types of heat storage system. These are sensible, latent and chemical energy storage
systems(Figure 3). Sensible heat storage systems store energy by changing of the temperature of heat
storage material. In thermochemical heat storage systems energy stored with fission reaction.Recovering
energy from the storage takes place with an inverse reaction[Dinçer., 2002].
In latent heat storage systems storing heat and recovering it takes place with a fussion reaction
(melting/solidification) during phase change period. Latent heat storage systems which use phase change
materials allow storing heat in high density as well as in high amounts with a low temperature and volume
changes[Zhou at al., 2012].
Benefiting from heat storage systems allow usage of thermal systems efficiently with conservation of energy
and thermal comfort. In this study utility of latent heat storage systems which use paraffin as a heat storage
material are investigated. With an application, the effect of heat storage systems to the energy consumption
in buildings analysed.

165
Fig. 3.Heat Storage Methods[Sharma at al., 2009]

3.2. Latent Heat Storage Materials

3.2.1. Classification
Depends on the phase change type ,phase changed materials classified in three groups. These groups are
solid-solid PCM’s , solid-liquid PCM’s and liquid-gas PCM’s . The most suitable phase change materials for
thermal energy storage are solid-liquid one’s. Solid-liquid PCM’s investigated in three groups as organic
PCM’S , inorganic PCM’s and eutectics[Zhou at al., 2012].

3.2.2 PCM Selection Criterias

Although selection of PCM for any application is a complex problem, it is important for heat storage
effectivity.Potential PCM’ s ,depends on the application, must have appropriate melting temperature, hİgh
fussion(phase change) heat and thermal conductivity. Expected properties of PCM’s are given on the table
below(Table1). Most suitable PCM’s, using in heat storage systems in buildings, in terms of melting
temperatures are paraffins, fatty acids, salt hydrates and eutectic mixtures

Table 1. Expected Properties of Phase Change Materials[Zhou at al., 2012]

Appropriate melting temperature scale

High fussion latent heat for per unit volume

High thermal conductivity

Thermodynamic Properties
High specific heat and density.

For preventing storage problem ;low volume changes and vapour

pressure at phase change duration.

166
 Suitable melting.

High nucleation rate to prevent rapid cooling.

Kinetic Properties
High crystallisation rate to meet heat demand from heat storage.

Inverse melting/solidification cycle

Chemical stability.

Chemical Properties To not have any detorioration after high number of phase change cycle.

To not be corrosive

To not be a toxic ,flammable and explosive material.

Suitable price
Economic Properties
Wide range of application area .

4.EXPERIMENTAL STUDY
In this study utility of thermal storage for heating of Energy Building, which is located at Davutpaşa Campus
of Yildiz Technical University, with solar collectors investigated.
4.1. System Description
Solar energy is a cyclical energy source for this reason storing excess energy during daytime and using it at
2
night time will increase utilization of solar energy. Energy building is a duplex structure which has 46 m floor
area[Kincay at al., 2009].Building heated with the pipe coils furnished inside the wall.Heating demand
2
compensate with 3 solar collectors that has 4.86 m area.The heating system diagram of energy building
given below(Fig. 4). During the experiment, paraffin which melts between the range of 42-44° C and easy to
3
find in Turkey was used. The shall that Paraffin was stored in has a volume of 0,063 m and the paraffin shell
3
is located in an outer shell which is 0,22 m . Heat transfer fluid circulates inside the gap between the inner
and outer shall.There is also a 0,1 m diameter gap in the middle of the inner shall to increase heat transfer
surface area .

167
 Figure 4. Heating System Diagram of Yildiz Renewable Energy Building

4.2.Evaluation of Experimental Measurements

4.2.1 Determination Heat Loses of Building and Heat Gain of Solar Collectors
In this study 6 months of heating season between october and march was investigated. For analysing the
system, hourly avarage solar irridation and temperature for each day was used[10].By using this datas,
hourly avarage building heat load and heat gain of solar collectors were calculated. Accumulated excess
energy during daytime analysed and storable thermal energy for night time were determined(Fig 5-10).
The rate of heat gain of the solar collectors was calculated by using Eq(1):

Qu: Ac.FR[Ht.( τα)n –Uc(Tf,i-Ta)] (1)

The rate of heat loss from each surface of the experimental building was calculated by using Eq(2) :
Qc:U.A.(Ti-Ta) (2)
By collecting heat loss from each surface, overall heat loss of building investigated depends on time.The
difference between heat gain from heat collection unit and overall heat loses of the building Eq(3), gives the
amount of storable energy:
Qs= Qu- Qc (3)

168
 6000
6000

4000
4000

2000
2000

0
0
10000

5000

0
07:00-08:00 07:00-08:00
08:00-09:00 08:00-09:00 07:00-08:00
09:00-10:00 09:00-10:00 08:00-09:00
10:00-11:00 10:00-11:00 09:00-10:00
11:00-12:00 11:00-12:00 10:00-11:00
12:00-13:00 12:00-13:00 11:00-12:00
13:00-14:00 13:00-14:00 12:00-13:00
14:00-15:00 13:00-14:00
14:00-15:00
15:00-16:00 14:00-15:00
15:00-16:00
16:00-17:00 15:00-16:00
16:00-17:00
17:00-18:00 16:00-17:00
17:00-18:00
18:00-19:00 17:00-18:00

169
18:00-19:00
Fig. 5. October, Building Heat Lose-Collector Heat Gain Diagram

18:00-19:00
19:00-20:00

Fig. 7. December, Building Heat Lose-Collector Heat Gain Diagram

Fig. 6. November, Building Heat Lose-Collector Heat Gain Diagram
19:00-20:00
19:00-20:00
20:00-21:00
20:00-21:00 20:00-21:00

Hourly Heat Loses of Building

Hourly Heat Gain of Solar Collectors

Hourly Heat Loses of Building

Hourly Heat Loses of Building

Hourly Heat Gain of Solar Collecors

Hourly Heat Gain of Solar Collectors

 5000
10000

0
6000
5000
4000
3000
2000
1000
0
07:00-08:00
6000
5000
4000
3000
2000
1000
0
08:00-09:00 07:00-08:00
08:00-09:00 07:00-08:00
09:00-10:00 09:00-10:00 08:00-09:00
10:00-11:00 10:00-11:00 09:00-10:00
11:00-12:00 10:00-11:00
11:00-12:00
12:00-13:00 11:00-12:00
12:00-13:00 13:00-14:00 12:00-13:00
13:00-14:00 14:00-15:00 13:00-14:00
14:00-15:00 15:00-16:00 14:00-15:00
16:00-17:00 15:00-16:00
15:00-16:00
17:00-18:00 16:00-17:00
16:00-17:00
18:00-19:00 17:00-18:00
17:00-18:00 19:00-20:00 18:00-19:00

170
18:00-19:00

Fig. 10. March, Building Heat Lose-Collector Heat Gain Diagram

20:00-21:00
Fig. 8. January, Building Heat Lose-Collector Heat Gain Diagram

19:00-20:00

Fig. 9. February, Building Heat Lose-Collector Heat Gain Diagram

19:00-20:00
20:00-21:00
Hourly Heat Loses Of Building
Hourly Heat Loses of Building

Hourly Heat Gain of Solar Collectors

Hourly Heat Loses of Building

Hourly Heat Gain of Solar Collectors

Hourly Heat Gain of Solar Collectors

This study carried out in Yildiz Renewable Energy Building, located at Davutpaşa Campus,Yıldiz Technical
University. Building Heat Lose-Collector Heat Gain graphics given above ,shows hourly changes in heat
loses of building against thermal energy absorbed by collectors. Concordantly, due to this results; between
the months December-February , the three month period, because of the heat load of the building rose
sharply, storing energy gained by collectors seems not presentable. On the other hand in spring and winter
period (December,November,March) especially between the hours 10:00-15:00 utilizing from thermal storage
seems possible. During this period storable excess thermal energy is available. Stored heat can bu used at
night time while the solar energy run out.

4.2.2. Determining Amount of Excess Storable Energy

Daily amount of storable, excess energy shown in the diagram below( Fig. 11). It is clearly shown from the
figure that especially in December, October and March high amount of storable energy is available. In the
last decade energy demand have been increasing from day to day however energy supply decrease in the
similar trend.For this reason utilizing from renewable energies such as solar energy and wind energy is a
necessity. Studies in the way of increasing efficiency of energy systems are important for improving utility
of this sources. However storing energy and using it in the time while energy demand increases, will
countibute economy also provide demand-supply balance. Concordantly it seems worth to do research and
development studies about heat storage systems to increase their efficiency.

Fig. 11. Daily Storable Excess Energy for 6 Month Heating Period

RESULTS AND DISCUSSION

This study carried out in Yildiz Renewable Energy Building located at Davutpaşa Campus,Yıldiz Technical
University. Utility of thermal storage in the process of heating the energy building was investigated. Heat
storage systems; depends on their dimensions, applied storage system and heat storage material, can be
applied for short time or long time .In the condition of the difference between solar energy and energy
demand is low ,short time storage can be applied. For meeting the seasonal energy demand, long time
storage applied .This type of application allows a rise in utility of solar energy.
Due to study results;
*With this system between the months December-February , the three month period, because of the heat
load of the building rise sharply, storing the energy gained by collectors seems not presentable.
*In Spring and Autumn period (December,November,March) especially between the hours 10:00-15:00
utilizing from thermal storage seems possible. During this period storable excess thermal energy is
available. Stored heat can bu used at night time while the solar energy run out.

171
CONCLUSIONS
The results of this study show that during day time, especially in Autumn and Spring, because of the heat
load of the building is low, most of the collected thermal energy cannot be used. Heat storage sytems allow
storing this excess energy for anytime while the heat loads getting higher.
Research and development studies are important on the way of increasing efficiency and utility of latent heat
storage systems. By the way rate of utilisation from solar energy will get higher.

Nomenclature

Q heat, w
2
A area, m
FR exchanger Heat Efficiency Coefficient
2
H global Radiation, w/m
2
U heat Transfer Coefficient, w/m °C
T temperature, °C

Greek Letters
τ diffusion Coefficient
α absorption Coefficient

Subscripts
u useful
c collector
t total
f,i fluid inlet
a ambient
i indoor

REFERENCES
[1]Benli H., Durmuş, A. 2009. Performance Analysis of a latent heat storage system with phase change
material for new designed solar collectors in greenhouse heating,Fırat Üniversitesi. Solar Energy 83:2109-
2119.
[2]Sharma A., Tyagi V., V., Chen R., C., Buddhi D. 2009. Review on thermal energy storage with phase
change materials and applications. Renewable and Sustainable Energy Reviews. 13:318-345.
[3]Zhou, D., Zhao, Y., C., Tian Y. 2012. Review on thermal energy storage with phase change materials in
building applications. Applied Energy. 92:593-605.
[4]Paremeshwaran, R., Kalaiselvam S., Harikrishnan, S., Elayaperumal, A. 2012. Sustainable thermal
energy storage technologies for buildings; A review. Renewable and Sustainable Energy Rewievs.16:2394-
2433.
[5]Koca, A., Oztop.,F.,H., Koyun T., Varol Y. 2008. Energy and exergy analysis of a latent heat storage
system with phase change material for a solar collector. Renewable Energy. Vol. 33, Issue 4, 567-574.
[6]Koray Arda. 2006. Güneş Enerjisinin Depolanması ve Isıl Analizi, Master’s Thesis,Erciyes University.
[7]Öztürk,H.,H. 2005. Experimental evaluation of energy and exergy efficiency of a seasonal latent heat
storage system for greenhouse heating. Energy Conversion and Management. 46:1523-1542
[8]Dinçer, İ. 2002. Thermal energy storage systems as a key technology in energy conservation.
International Journal of Energy Research 26:567-588.
[9]Kıncay O., Utlu Z., Ağustos H., Akbulut U., Açıkgöz Ö. 2009. Yenilenebilir Enerji Kaynaklarında Birleşme
Eğilimi. Sigma: National Journal of Engineering and Science. Vol.27, Issue 2, 60-82.
[10]http://re.jrc.ec.europa.eu/pvgis/

172
 ENVIRONMENTAL EFFECT EVALUATION OF TOPOGRAPHY AND NATURAL GAS USAGE
ON AIR QUALITY: A CASE STUDY OF K.MARAŞ

M. Imal, C. Karapınar
University of K.Maraş S.I. Faculty of Engineering. Mechanical Eng. Dept., Kahramanmaraş
muharremimal@ksu.edu.tr

ABSTRACT: In this study, sulfur dioxide (SO₂) and particulate substance (PM) density rate have been
measured through analyzing the various factors affecting the air quality in Kahramanmaraş. It has been
aimed to point out how the air pollution could be cut down to minimum with these parameters.
For this purpose SO₂ and PM density rates between the years 2007–2012 from Provincial Environmental
Directorate of Kahramanmaraş, temperature and wind range values from Provincial Meteorology Directorate
of Kahramanmaraş, the increase of population from Turkish Statistical Institute, volume of consumed natural
gas and the increase rate in the number of subscribers between the years 2007–2012 from ARMADAŞ
(incorporation of distribution) have been obtained.

Keywords: Natural Gas, Air Pollution, SO₂, PM

INTRODUCTION
Air, just like water and soil, is an environment which can be polluted. Air pollution can be described as the
presence of one or more pollutants in air at a certain level and period of time that will harm human life,
habitat, both articles of trade and personal belongings and environmental quality. (Muezzinoglu, A., 1987).

The source of pollutants leading to pollution in the atmosphere is varied. Beyond doubt primary particles
leading to air pollution are fossil fuels which are used as a source of energy. Also some particles released as
gas by chemical industry, some kind of metallurgical dust, the dust emitted by cement chimneys, and other
industrial activities are the causes of pollution (Akman, Y.,2004).

One of the most important negative factors of modern life in human life is air pollution. Air pollution, in
th
general environmental, pollution began to come into question as from the 19 century when industrialization
had started developing rapidly. Case of air pollution indicates a complex structure with its social, technical
and economical dimensions, and every country makes its legal and technical regulations in accordance with
its structure. In these regulations, air pollution is tried to be controlled within the harmless limits in terms of
environmental health (Alp., K., 2010).

In Turkey, depending on the meteorological conditions especially in winter in some urban areas, air pollution
is observed. Main causes of air pollution resulting from heating in winter can be listed as the use of poor
quality fuels without overhauling process, wrong combusting techniques and lack of maintenance in
combustion systems. But in heating, as a result of the use of natural gas and high-quality fuels, especially in
big cities, air pollution has been reduced compared to the 1990s. While building industrial plants,
inappropriate locating also increases negative effects of air pollution. Because of high sulfur content of
domestic lignite used in coal based thermal power plants and lack of treatment systems in some facilities;
sulfur dioxide (SO ₂) emissions pose a problem.

MATERIALS AND METHODS

Air pollution can be described as a mixture of pollutants, enough to harm people and the environment, into
air (Muezzinoglu, A., 1987). But there are particulate substances in fresh air, named as fog, smoke, vapor
etc. Particulate emission may result from human doings (fuel combustion, grinding stones and minerals),
natural resources (volcanoes, wind erosion, sea waves, natural corrosion) and atmospheric reactions.

173
Table 1. Clean and Polluted Settlements
Pollutant Clean Air Polluted Air
PM₁₀ 10-20 µg/m3 260-3200 µg/m3
SO₂ 0.001-0.01 ppm 0.02-3.2 ppm
CO₂ 300-330 ppm 350-700 ppm
CO 1 ppm 2-300 ppm
NOx 0.001-0.01 ppm 0.3-3.5 ppm
T. Hydrocarbons 1 ppm 1-20 ppm
T. Oxidants 0.01 ppm 0.01-1.0 ppm

In our country, SO2 and PM rates are measured in air pollution measurement stations. Among the gaseous
pollutants sulfur oxide is the most common nonflammable and colorless air pollutant. Its retentively in the
atmosphere lasts for about 40 days. Often it arises from the combustion of fossil fuels. More than 80 per cent
of anthropogenic sulfur oxide is estimated to have arisen from industrial resources (Agren, C., 1991).
Particulate substance is the mixture of solid particles suspended in the atmosphere with liquid droplets.
Particle size is very wide in range. Some particles such as dust and smoke are big enough to be seen
through naked eye. Along with this, there are also microscopic particles (Ministry of Health., 2010).

Effect of Topography on Air Quality

In this study, how the natural gas use and topography change the amount particulate substance SO2 is
studied.
Especially in cities where air pollution is very high, as a result of failure to take necessary measures there
has been air pollution problem and in case critical meteorological conditions occur, air pollution starts to
threaten the environment and human health (Ministry of Environment and Forestry., 2004).
Kahramanmaras has been settled in around the mount of Ahir. As the population has increased, the city has
expanded upward, to the foothills of the mountain. That the city has expanded in this direction has had a
positive impact on the air quality of the city. Especially in winter when viewed from a vantage point, it can be
seen the pit area of the city, which is called “Ancient Maraş”, is under a cloud of smoke. But in newly
urbanized higher part of the city, this cloud of smoke is less.

It will be helpful to study the reports of Provincial Directorate of Environment in order to more clearly observe
this circumstance.

In the center of Kahramanmaraş, there is one air quality monitoring station. Until 2009 March, this station
had been located in the yard of Governor’s Office, which is the pit area of the city, then has been moved to
the yard of City Cultural Center, which is in newly urbanized and higher part of the city.

174
Fig. 1. Air Quality Monitoring Station

Based on this information, comparison between the parameters about the air quality of the periods 2007-
2008 and 2009-2012 will provide us with the information about whether topography has an effect on air
quality or not.

250

200
μg/mᶾ

150

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 MONTHS

2007 PM 2008 PM

Fig. 2 . The PM parameters measured monthly for the years 2007 to 2008.

175
 180

160

140

120

100
μg/mᶾ

80

60

40

20

0
1 2 3 4 5 6 7 8 9 10 11 12 MONTHS

2007 SO₂ 2008 SO₂

Fig. 3. The SO2 parameters measured monthly for the years 2007-2008

180
μg/mᶾ

160

140

120

100

80

60

40

20

0
1 2 3 4 5 6 7 8 9 10 11 12
MONTHS
2008SO₂ 2009 SO₂

Fig. 4. The SO2 parameters measured monthly for the years 2008-2009

176
 250

200

150
μg/mᶾ

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 MONTS

2008 PM 2009 PM

Fig. 5. The PM parameters measured monthly for the years 2008 to 2009.

250

200

150
μg/mᶾ

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 MONTHS

2009 PM 2010 PM
Fig. 6. The PM parameters measured monthly for the years 2009 to 2010.

177
 40

35

30

25

20
μg/mᶾ

15

10

0
1 2 3 4 5 6 7 8 9 10 11 12
MONTHS

2009 SO₂ 2010 SO₂

Fig. 7. The SO2 parameters measured monthly for the years 2009-2010

When the figure 4., figure 5., figure 6. and figure 7. above are analyzed, it can be seen the PM and SO2
parameters of the periods 2007-2008 and 2008-2009 are in the same line with each other. Also, the PM and
SO2 graphs of 2009-2010 are in the same line with each other. But a significant change stands out in the
graphs starting from 3rd month of 2009. This coincides with the date when the location of air quality

In order to see this change more clearly, we have to examine the following analysis where PM and SO2
parameters of the years from 2007 to 2011 are supplied in a single graph.

178
 250

200

150
μg/mᶾ

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12
MONTHS
2007 PM 2008 PM 2009 PM 2010 PM 2011 PM

Fig. 8. The PM parameters measured monthly for the years 2007 to 2011.

When the PM parameter of five years shown in the figure 8. are analyzed it can be easily seen that these
parameters have changed significantly, especially starting from 2009. As mentioned earlier, the location of
air quality monitoring station was replaced in February 2009. As a result of this replacement, there has been
a significant reduction in PM parameters.

179
 180

160

140

120

100

80
μg/mᶾ

60

40

20

0
1 2 3 4 5 6 7 8 9 10 11 12

2007 SO₂ 2008 SO₂ 2009 SO₂ 2010 SO₂ 2011 SO₂ MONTHS

Fig. 9. The SO2 parameters measured monthly for the years 2007 to 2011.

The monthly change of SO2 parameters in 5 years is given in the graph shown in the figure 9.
When the PM parameters in this graph are analyzed, it can be seen that SO2 change over the years is very
sharp, especially in particular periods. From 2009 on SO2 parameters have declined in comparison with
those of 2008 and 2009. The reason for this decline is that the air quality monitoring station, that was earlier
located in the yard of Governor’s Office, the pit area of the city, was moved to the yard of City Cultural
Center, the higher part of the city.

Furthermore, that the location of Governor’s Office is in the pit area of the city directly affects the velocity of
wind in this area. The average velocity of the wind here is about 0.6 m/s. Low velocity of the wind hampers
the dispersion of pollutants collected over the city.

Low velocity of the wind does not only result from the topography. Along with that the Governor’s Office is
located in the pit area of the city; this is the area where there is the highest density of buildings. In this area,
the buildings are contiguous and 10-storeyed approximately. These contiguous buildings hinder the flow of
wind and significantly reduce the air current. Thus, it takes a while for pollutants to disperse and move away
from the area.

Based on these results, it can be said that the expansion of the city must be towards to the higher and plain
part where there is fluent air current rather than the pit area where there is less air current. Furthermore, the
buildings must be built in a way that they don’t hamper the air current.

In the light of the data above, it is clear that the topography of the city directly affects its air quality.
The Effect of Fuels Burned for Heating
One of the main sources of air pollutants in Kahramanmaraş is the fuel used for heating. Coal and natural
gas are used for heating in dwellings and workplaces for heating. In 2011, 145.000 tons of coal was sold for
heating by the vendors of imported coal and 31.000 tons of domestic coal was distributed by the Social Aid
3
Association in Kahramanmaraş. In the same period 81.650 million m of natural gas was consumed in the
city.

180
Natural gas began to be used in 2006 in Kahramanmaraş and the number of subscribers is increasing day
by day. Even if it was not at desired level in 2007-2009 when the economic crisis was influential, in 2010 and
2011 the sector recovered from this recession and the number of subscribers began to increase. According
to the data received from ARMADAŞ, incorporation of distribution, the number of subscribers in years 2007
to 2012 is as follows in the Table 2 (Armadas., 2012).

Table 2. Number of Subscribers in Kahramanmaraş in Years.

THE NUMBER OF SUBSCRIBERS DECLARED BY
ARSAN MARAŞ NATURAL GAS DISTRIBUTION INC.

DATE NUMBER OF SUBSCRIBERS

31.12.2007 7887

31.12.2008 14460

31.12.2009 19835

31.12.2010 30545

31.12.2011 42229

01.05.2012 43779

TOTAL TARGET 100.000

That natural gas is a clean fuel makes it advantageous in terms of boiler maintenance and performance.
When fuel oil or coal is burned, ash and soot accumulated on the inner layer of boiler both corrode it and
lower the efficiency hindering heat conduction. So the boiler pipes must be cleaned at least twice a week,
whereas this is not the case in natural gas (Cagisitmaklima.,2005).

Furthermore, natural gas, emission standard of which is quite low in comparison with other types of fuels
used, is crucial in terms of air quality. In the table 3. Some emission standards of coal fuel oil and natural
gas, each of which is equivalent to 1 ton of oil (Eren Z., et al.,2007).
Table 3. Emission coming out as a result fuels, equivalent to 1 ton of oil.
Fuel Emission (kg)
Coal Fuel Oil
Natural Gas
%sulphur %10 ash %1 sulphur
PM 100 1.8 0.1-0.3
Sulphur Oxide 29.2 20 -
Carbon Monoxide 1.5 0.7 0.3
Nitrogen Oxides 11.3 8.2 2.3-4.3

About forty percent of objected natural gas user number has now been achieved and these subscribers have
been using natural gas since April 2012. This situation is supposed to affect the air quality. In order to
observe the effect of increased natural gas use, the air quality reports of Kahramanmaraş in years 2007 to
2011 are examined in graph.

181
 180 50,000

160 45,000

140 40,000

35,000
120
SO₂ μg/mᶾ

Natural gas users

30,000
100
25,000
80
20,000
60
15,000
40 10,000
20 5,000

0 0,000
2007 2008 2009 2010 2011 2012 YEARS
NATURAL GAS THE NUMBER OF SUBSCRIBERS

Fig. 10. SO2 parameters measured monthly for the years 2009 to 2010.

In the figure 10. the change in SO2 parameters in February compared to the number of natural gas users in
Kahramanmaraş is observed. It is observed that there is a sharp decline in the years 2008 to 2009. This is
because the location of air quality monitoring station was replaced in 2009. Besides, measured SO2
parameters continued to decline in 2010, 2011 and 2012. It has been observed that the more natural gas
use resulted in the less SO2 in the air.

As seen in the table 3. while there is 29.2 kg sulphur ingredient in 1 TEP coal, it is about zero in natural gas.
Coal use will decline in proportion to the increased use of natural gas. Therefore, the amount of SO2 emitted
into the atmosphere will decrease.

In the figure 11. the change in the number of natural gas users is indicated. A sharp decline is observed in
the years 2008 to 2009. This is because the location of air quality monitoring station was replaced in 2009.
Besides, measured PM parameters continued to decline in 2010, 2011 and 2012. It has been clearly
observed that the more natural gas use resulted in the less PM in the air.

As seen in the table 3. while there are 100 kg PM ingredients in 1 TEP coal, it is about 0.1-0.3 in natural
gas. Coal use will decline in proportion to the increased use of natural gas. Therefore, the amount of PM
emitted into the atmosphere will decrease.

182
 250 50,000

45,000

200 40,000

35,000
PM μg/mᶾ

Number of natural gas users

150 30,000

25,000

100 20,000

15,000

50 10,000

5,000

0 0,000
2007 2008 2009 2010 2011 2012 YEARS

NATURAL GAS THE NUMBER OF SUBSCRIBERS

Fig.11. Measured PM parameters for February in years and the change in the number ofnatural gas users in
the city.

CONCLUSION

The only air quality monitoring station in the city center of Kahramanmaraş is located in the pit area of the
city, the yard of Governor’s Office. In 2009, it has been moved to the yard of City Cultural Center, which is in
the higher part of the city and where there is more fluent air current. It has been concluded that the
topography is one of the main factors affecting the air quality of the city according to the measurements
made before and after the year 2009. When constructing new neighborhoods in the city, choosing areas
where there will be more air current and wind will improve the air quality. That the buildings are contiguous
and 10-storeyed approximately in the area of Governor’s Office adversely affects the air current and
hampers the dispersion of pollutants in the atmosphere. While constructing buildings in newly urbanized
areas if not to cut off the air current is taken into account, the air quality improves.

Natural gas has improved the air quality in Kahramanmaraş. For this study, temperature and wind statistics
over the years from the Meteorological Service, PM and SO2 statistics from the Directorate of Environment
and Urban Planning, the number of subscribers and the amount of natural gas consumed over the years
from ARMADAŞ (incorporation of distribution) were collected. Using this statistics, the graphs of “SO2 – the
number of natural gas users” and “PM –the number of natural gas users” were formed. It has been observed
in these graphs that the more natural gas use resulted in the less SO2 and PM in the air, especially in recent
years when the number of natural gas users has increased.

The use of coal decreases in inverse proportion to the increase in the use of natural gas. This decline in the
use of coal directly affects the air quality and leads to more clean air. Thus, the municipality should
cooperate with the incorporations of natural gas distribution to help the natural gas be serviced to people,
and the use of natural gas should be supported.

183
Burning equipment should be maintained regularly and renewing the economically worn-out burners will
increase the efficiency. Due to the boilers running more efficiently will emit less pollutant to environment, air
quality will improve. Because the boiler is renewed in the buildings that shifted to natural gas from coal, its
efficiency directly increases. Besides, the companies that take on the conversion to natural gas and boiler
maintaining services inform people, thus the average boiler efficiency increases in proportion to the natural
gas use.

REFERENCES

[1] Agren, C., 1991., EMEP Report, MCS-W 1/91 Norway.

[2] Akman, Y., Ketenoğlu, O., Kurt, L., Evren, H., and Düzenli, S., 2004., Environmental Pollution,
Palme Press., Ankara.

[3] Alp., K., 2010., Environmental Specialist Education Meeting, Antalya.

[4] Armadas., 2012., Arsan Maraş Natural Gas Distribution Inc., The Bulletin of ARMADAŞ.,
Kahramanmaras

[5] Eren Z., Turan T., 2007., With the transition from fossil fuels to natural gas in the Province of
Erzurum Weather Change pollution, TBMMOB 7. Chamber of Environmental Engineers, National
Environmental Engineering Congress, Izmir.

[6] http://www.cagisitmaklima.com/faydali4.htm., 2005.

[7] Muezzinoglu, A., 1987., Fundamentals of Air Pollution and Control, Dokuz Eylul University,
Press, 0908.87.DK.006.042, Izmir.

[8] Ministry of Health of The Republic of Turkey., General Dir. of Primary Health Care Press., 2010.,
No.811, Ankara.

[9] Ministry of Environment and Forestry of The Republic of Turkey.,2004.,, General Dir. Of
Environmental Man., Circular Letter of Air Pollution., B.18.O.ÇYG.0.02.00.02/3888., Ankara.

184
 THERMODYNAMIC ANALYSIS OF GEOTHERMALLY DRIVEN
HIGH-TEMPERATURE STEAM ELECTROLYSIS SYSTEM FOR HYDROGEN
PRODUCTION
a b c
M.T. Balta , I. Dincer , A. Hepbasli
a
Department of Mechanical Engineering, Faculty of Engineering,
Aksaray University, 68100 Aksaray, Turkey
b
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology (UOIT),
2000 Simcoe Street North, Oshawa, ON L1H 7K4, Canada
c
Department of Energy Systems Engineering, Faculty of Engineering,
Yaşar University, 35100 Bornova, Izmir, Turkey

ABSTRACT
Hydrogen production by high-temperature steam electrolysis (HTSE) has been receiving increasing
attention worldwide due to its high efficiency and carbon-free operation. Although it is still considered in
its early developmental stage, it offers a promising solution for highly efficient hydrogen production.
From the thermodynamic viewpoint of water decomposition, it is more advantageous to electrolyse
water at high temperatures because the energy is supplied in mixed form of electricity and heat. In this
study, a HTSE process coupled with and powered by a geothermal power plant is considered for
analysis and assessment and as a case study. In this regard, its thermodynamic analysis through
energy and exergy is conducted for performance evaluation and comparison purposes. As a result,
compared to conventional water electrolysis, the operation at high temperatures reduces the electrical
energy requirement for the electrolysis and also increases the efficiency.

Keywords: Energy, Exergy, Efficiency, Geothermal, Hydrogen, High temperature steam electrolysis.

INTRODUCTION
Increasing population, rapid urbanization and development lead to high and continuous increase in
energy demand. Nowadays world’s energy demands are almost supplied by carbon-containing fossil
sources such as natural gas, oil and coal, which have been formed during many millions of years from
plant biomass. Global consumption of these non-renewable resources for energy demand is resulting in
critical environmental issues, such as greenhouse effect, climate change, ozone depletion and global
warming. In addition to these environmental issues, there are other critical concerns, such as energy,
economic and political crises. New and more efficient energy conversion systems are required in the
near future, due to the increase in oil prices and demand as well as global warming (Laguna-Bercero,
2012).
Primary energy resources, such as fossil fuels, are diminishing and must be regarded as a finite
resource. Therefore, water may be considered to be the cleanest fuel source instead of fossil fuels. By
developing hydrogen energy system, one can transfer energy consumption to the water cycle rather
than the hydrocarbon-based cycle. It is important to highlight that hydrogen has substantial
environmental benefits if it is produced through renewable energy sources.
Although many scientists propose a large variety of solutions, one of the important solutions is
implementing the hydrogen economy. Hydrogen is probably the preferred energy carrier for a future
zero-carbon economy, but several research efforts are required in order to supply inexpensive and
plentiful amounts of fuel. Hydrogen economy offers a compelling vision of an energy future for the
nation and the world that is at once abundant, clean, flexible and secure (ANL, 2004).
Hydrogen is usually found as a compound combined with other elements, and thus, the
production of hydrogen always requires energy (Laguna-Bercero, 2012). Currently the most mature and
inexpensive way to produce hydrogen is through steam reforming, partial oxidation of heavy
hydrocarbons, gasification and conventional alkaline water electrolysis. Figure 1, illustrates that 96%
hydrogen is produced directly from fossil fuels, while about 4% is produced indirectly by electricity
utilization.
The conventional hydrogen production methods are not considered renewable since these
processes consume natural resources and also release high levels of CO 2. Therefore, it will not be a
potential solution to combat with environmental issues (Balta et al., 2010a; 2011). Clearly, we urgently
need to find alternative, eco-friendly ways for hydrogen production technologies to ensure the future
generations receive a cleaner and sustainable energy-based economy and society. Several
researchers have been investigating a wide range of technologies to produce hydrogen economically
from a variety of resources in environmentally friendly ways (Kanoglu et al., 2011). In this regard,

185
renewable energy resources have attracted great interest in recent years. Alternative clean and efficient
pathways for the production of pure hydrogen are water electrolysis and thermochemical water-splitting
cycles with renewable energy sources.
Hydrogen can be produced from water or steam electrolysis using much simpler technology
through the following reaction:
H2O + Energy => H2 + ½ O2 (1)

Other abbbbbbbbbbbb Electrolysis

0.1% 3.9% Coal
Gasification
18.0%

Methane Steam
Reforming Oil Reforming
48.0% 30.0%

Fig. 1. Feedstock used in the global hydrogen production (based on the data taken from Ewan and Allen (2005)).

Water electrolysis process is one of the environmentally benign hydrogen and oxygen
production methods without emitting greenhouse gases if the required electricity is supplied from
renewable energy sources. Electrolytic hydrogen production has many advantages such as it yields the
highest-purity hydrogen (up to 99.999%). This technology is well developed, but overall system
efficiency is too low, about 27%. The main disadvantage of water electrolysis is that electricity is an
expensive “fuel”. The concept of integrating renewable energy with hydrogen production systems was
given some serious consideration in the 1970s (Eisenstadt and Cox, 1975; Costogue and Yasui,
1977).The aim of the above mentioned concept was to improve the performance of existing hydrogen
production systems, but also to integrate the hydrogen production systems effectively with renewable
energy sources. When considering the use of renewable energy for hydrogen production, geothermal
resources seem to be an important option (Balta et al. 2009).
Geothermal is treated as a reliable and promising renewable energy. The use of geothermal
energy through hybrid or integrated systems for sustainable hydrogen production appears to be an
environmentally benign and sustainable option for the countries having abundant geothermal energy
resources. In this study, a HTSE process coupled with and powered by a geothermal power plant is
considered for analysis and assessment and as a case study. In this regard, its thermodynamic analysis
through energy and exergy is performed for performance evaluation and comparison purposes.

HIGH-TEMPERATURE STEAM ELECTROLYSIS

High temperature steam electrolysis (HTSE) is still considered in the early developmental stage. The
HTSE offers a promising method for highly efficient hydrogen production. From the thermodynamic
viewpoint of water decomposition, it is more advantageous over electrolyse water at high temperature
o o
(800 C -1000 C) because the energy is supplied in a mixed form of electricity and heat. The steam to
be dissociated enters on the cathode side. After the steam has been divided into hydrogen gas and
oxygen ions, the oxygen ions are transported through the ceramic material to the anode where they
discharge and form oxygen gas. The most common ceramic material is zirconia, ZrO2 (Arnason and
Sigfusson, 2003). Compared to conventional water electrolysis, operation at high temperatures reduces
the electrical energy requirement for the electrolysis and also increases the efficiency. The effect of
temperature on the energy requirement for HTSE is shown in Fig. 2, as related to the enthalpy of the
water splitting process.The enthalpy is given as the sum of the minimum work needed and the
temperature-entropy term:
ΔH = ΔG + TΔS (2)
where ΔH is the enthalpy change or total energy demand, ΔG is the Gibbs free energy or the minimum
work, T is the absolute temperature and ΔS is the entropy change. The term TΔS can be considered as
the total amount of thermal energy needed to split water (Arnason and Sigfusson, 2003). The electrical
energy demand decreases with increasing temperature leading to increased direct heat requirements,

186
the decrease in electrical energy demand drives to thermal to hydrogen energy conversion efficiency to
higher values. This is one of the advantages of HTSE. The temperature dependence of the
thermodynamic functions for splitting of the fluid is shown in Fig. 2.

4
Energy Demand (kWh/Nm3

3
Total Energy Demand
H2 )

2
Electrical Energy
Demand
1
Heat Demand

0
473 573 673 773 873 973 1073 1173
Temperature (K)

Fig. 2. Energy demand for high temperature steam electrolyses (Balta et al., 2009).

The energy and exergy efficiencies of the overall system are generally defined as follows:
 energy in products 
 overall    (3a)
 total energy input 
 exergy in products 
 overall    (3b)
 total exergy input 
Here, the overall efficiencies of the HTSE system were analyzed by Mingyi et al. (2008). A
thermodynamic model of the HTSE system efficiency was well-established, and the effects of electrical,
thermal and overall efficiencies of the HTSE system were investigated. The overall efficiencies of the
HTSE system under different conditions ranged from 33% to a maximum of 59% at a temperature of
o
1000 C, which was over two times higher than that of the conventional alkaline water electrolysis.
Note that analyses of experimental and computational model of high-temperature steam
electrolysis were presented by Herring et al. (2007). Computational model and experimental results
were compared favorably.Hydrogen production efficiency of high temperature electrolysis coupling with
high temperature gas cooled reactor was evaluated by Fujiwara et al. (2008). Hydrogen production
o
efficiency at high temperature electrolysis for operating temperature of 800 C was calculated over 53%.
A research and development program has been underway at the Idaho National Laboratory (INL) on
high-temperature electrolysis for hydrogen production from steam (O’Brien et al. 2005). The electrolysis
was coupled with an advanced reactor and power cycle, and the efficiency of the system was calculated
45- 50%, respectively.
Energy and exergy analyses of a water-electrolysis process for producing hydrogen were
investigated by Rosen and Scott (1992) and Rosen (1995). Three driving energy inputs were
considered: (i) electricity, (ii) the high-temperature heat used to generate the electricity, and (iii) the heat
source used to produce the high-temperature heat. The analyses indicated where the losses were
occurred.
Electrolysis currently accounts for a very small portion of the hydrogen generated in developed
countries that have a commercial hydrogen infrastructure. However, electrolysis can be economic for
small-scale generation in areas with inexpensive electricity which was generated by renewable such as
geothermal resources.

SYSTEM DESCRIPTION
Many studies have recently appeared concerning future hydrogen demand; for example, one was
presented by Sigurvinsson et al. (2006), considering a HTSE process coupled with a geothermal source
o
at Nasjavellir site in Iceland. The geothermal source was about at 230 C and 15 bar. In this section,

187
energy and exergy analyses of a high-temperature electrolysis are performed for comparison purposes,
of which detailed description is given in Sigurvinsson et al. (2006). The schematic flow diagram of the
HTSE process along with its main components is shown in Fig. 3 while their thermodynamic properties
are listed in Table 1.

Table 1. The HTSE system data, thermodynamic properties and energy and exergy values.

State
Substance Description T P m h s ex ph E x ph E
No (K) (kPa) (kg/s) (kJ/kg) (kJ/kg.K) (kJ/kg) (kW) (kW)
0 H2O Reference State 298 101.325 - 104.8 0.3669 - - -
0' H2 Reference State 298 101.325 - 0 64.82 - - -
0'' O2 Reference State 298 101.325 - 0 6.407 - - -
1 H2O H2O inlet LT- H2 503 1500 11.400 2874 6.613 908 10350 32764

H2O outlet LT- H2 /

2 H2O 753 1500 11.400 3429 7.513 1195 13619 39091
H2O inlet MT- H2

H2O outlet MT- H2 /

3 H2O 978 1500 11.400 3932 8.096 1524 17373 44825
H2O inlet HT- H2

4 H2O H2O outlet HT- H2 1185 1500 11.400 4421 8.549 1878 21408 50399
5 H2O H2O inlet LT- O2 503 1500 4.000 2874 6.613 908 3631 11496

H2O outlet LT- O2 /

6 H2O 693 1500 4.000 3299 7.332 1119 4474 13196
H2O inlet MT- O2

H2O outlet MT-O2 /

7 H2O 983 1500 4.000 3944 8.108 1532 6129 15776
H2O inlet HT- O2

8 H2O H2O outlet HT- O2 1185 1500 4.000 4421 8.549 1878 7512 17684
9 H2O Electrolyser inlet 1185 1500 15.400 4421 8.549 1878 28920 68083
10 H2 H2 inlet HT- H2 1223 10000 1.147 13653 66.51 13149 15076 15653
10' H2O H2O inlet HT- H2 1223 10000 5.082 4486 7.731 2187 11113 22798
11 H2 H2 outlet HT- H2 987 10000 1.147 10056 63.24 10527 12069 11529
11' H2O H2O outlet HT- H2 987 10000 5.082 3903 7.202 1761 8951 19835
12 H2 H2 inlet MT- H2 987 2000 1.147 10056 69.88 8548 9800 11529
12' H2O H2O inlet MT- H2 987 2000 5.082 3950 7.982 1576 8009 20074
13 H2 H2 outlet MT- H2 782 2000 1.147 7016 66.43 6536 7494 8044
13' H2O H2O outlet MT- H2 782 2000 5.082 3488 7.457 1270 6456 17726
14 H2 H2 inlet LT- H2 782 7000 1.147 7016 61.27 8074 9257 8044
14' H2O H2O inlet LT- H2 782 7000 5.082 3432 6.826 1402 7127 17441
15 H2 H2 outlet LT- H2 577 7000 1.147 4030 56.84 6408 7347 4620
15' H2O H2O outlet LT- H2 577 7000 5.082 2854 5.958 1083 5504 14504
16 O2 O2 inlet HT- O2 1223 10000 9.200 955.3 6.634 888 8166 8789
17 O2 O2 outlet HT- O2 987 10000 9.200 725.9 6.428 720 6621 6678
18 O2 O2 inlet MT- O2 987 2000 9.200 725.9 6.816 604 5557 6678
19 O2 O2 outlet MT- O2 782 2000 9.200 406.6 6.474 387 3557 3741
20 O2 O2 inlet LT- O2 782 7000 9.200 406.6 6.149 483 4448 3741
21 O2 O2 outlet LT- O2 569 7000 9.200 258.2 5.917 404 3719 2375
*State numbers refer to Fig. 3

188
 Fig. 3. A schematic of the HTSE system studied, as proposed by Sigurvinsson et. al (2006).

ANALYSIS
Before conducting energy and exergy analyses, the following assumptions are made:
 The values for the reference environment (dead state) temperature (T 0) and pressure (P0) are 25 °C
and 100 kPa, respectively.
 All processes are considered steady-state and steady-flow with negligible potential and kinetic
energy effects in an adiabatic form.
For a general steady-state, steady-flow process, the three balance equations, namely mass, energy
and exergy balance equations, are employed to find the work input, the rate of exergy destruction and
overall energy and exergy efficiencies.
Since mass is conserved in chemical reactions, the mass of products and reactants are equal
and in general, the mass balance equation can be expressed in the rate form as
∑m in  ∑m
 out or ∑m R  ∑m P (4)

where m is the mass flow rate, and the subscript in stands for inlet and out for outlet.
The general energy balance is written as
E in  E out  E s (5)

which becomes

Q W   n h
P
o
f h ho    n h
P R
o
f h ho 
R
(6)

and the exergy balance for the process, involving chemical reactions, becomes

 Exin -  Exout  Exd  E x s (7)

For a steady-state system, E x s is zero. The exergy associated with a process at a specified state is
the sum of two contributions: physical and chemical.
Thus, the specific exergy of the process is calculated by

ex  (h - h0 ) - T0 (s - s0 )  ex ch (8)
where h is enthalpy, s is entropy, and the subscript zero indicates properties at the reference (dead)
state of P0 and T0.
The exergy rate is then calculated by

E x  m ex (9)
Combining Eqs. (7) and (8) yields

189
   
ex d   (h - h0 ) - T0 (s - s 0 )  ex ch   (h - h0 ) - T0 (s - s 0 )  ex ch 
out
 We, in (10)

After writing mass, energy and exergy balances for the system, enthalpy values of H 2, O2 and H2O are
evaluated with Shomate equations as follows:

T2 T3 T4 1
h  h0  A * T  B * C*  D*  E*  F H
2 3 4 T (11)
where T is 1/1000 of the specified temperature (in K) of compound and A, B, C, D, E, F,G and H are
constants, as given in Table 2 for H2, O2 and H2O.

Table 2. Enthalpy of formation and Shomate constants for H2, O2 and H2O.
h fo
Compound A B C D E F G H
(kJ/kmol)
H2O (g) -241830 30.0920 6.832514 6.793435 -2.534480 0.082139 -250.881 223.3967 -241.8264
O2 (g) 0 29.6590 6.137261 -1.186521 0.095780 -0.219663 -9.861391 237.9480 0
H2 (g) 0 33.0661 -11.36340 11.432816 -2.772874 -0.158558 -9.980797 172.7079 0
Source: Adopted from (NIST, 2013)

The chemical exergy based on a typical exergy reference environment exhibiting standard
values of the environmental temperature T 0 and pressure P0 such as 298 K and 100 kPa is called
standard chemical exergy. The values of the chemical exergies for the reactants and products are taken
from the literature (EP, 2013), as listed in Table 3.

Table 3. Gibbs free energy and standard chemical exergy of H 2, O2 and H2O.

Specific Gibbs free energy of formation Standard chemical exergy

Compound
g of (kJ/kmol) ch
ex (kJ/kmol)
H2O (g) -228638 9437
H2 (g) 0 236090
O2 (g) 0 3970

The exergy destructions in the heat exchangers and electrolyses are calculated as
E xd , HE  E xin - E xout (12)

E xd, el  W e - Exout - Exin  (13)

The exergy efficiency of the heat exchanger is determined by the increase in the exergy of the
cold stream divided by the decrease in the exergy of the hot stream on a rate basis as follows:
E xcold ,out - E xcold ,in m cold (ex cold ,out - ex cold ,in )
 HE   (14)
E xhot ,in - E xhot ,out m hot (ex hot ,in - ex hot ,out )
The expressions of energy and exergy efficiencies for the geothermal assisted HTSE system are based
on the above definitions. The theoretical overall energy and exergy efficiencies of the system can be
expressed as
HHV
 overall 
G /  el es  Q /  th (15)

Ex H 2
 overall 
G /  el es  Ex Q (16)

190
  T  Q
Ex Q  1  0 
 T   th
(17)
Table 4. The assumed typical values.
Parameter ηel (%) ηes (%) ηth(%)
Typical values 10-20 80-100 80-100
Adopted from Zhang et al. (2012).

Here, the theoretical overall efficiency can be affected by three key parameters. These parameters are
the efficiency of the high temperature steam electrolysis cell (ηes), the electricity generation efficiency by
the geothermal power plant (ηel) and thermal efficiency (ηth) and also these parameters are closely
related to the temperature.

Heat exchanger networks

In the study of Sigurvinsson et al. (2006;2007), counter current heat exchangers were used while the
o
inlet temperature of the electrolyses was kept as 950 C. The temperatures in the geothermal case
ranged from 200 to 950 °C. The heat exchangers were classified into three groups according to the
ranges of the temperatures since this temperature range could not be covered with one type of heat
exchangers. Therefore,
o
 Low temperature (LT): stainless heat exchanger, T < 600 C and 7 MPa,
o o
 Medium temperature (MT): nickel based heat exchanger, 600 C < T < 850 C and 7 MPa,
o
 High temperature (HT): ceramic based heat exchanger, T > 850 C and 10-50 MPa.
The fluid in the first heat exchanger network is only oxygen. In the second heat exchanger, the fluid is a
steam and hydrogen mixture. This mixture depends on the recycling ratio, which is taken to be 0.33 in
these analyses. The mass flow rates of hydrogen and oxygen are calculated by
M H2
m out , H 2  (1  r )  m H 2O   r  m H 2O (18)
M H 2O

(1  r ) M O2
m out ,O2   m H 2O  (19)
2 M H 2O
RESULTS AND DISCUSSION
The temperature, pressure, and mass flow rate data on water, hydrogen and oxygen are given in Table
1 according to their state numbers specified in Fig. 3. The specific physical exergy and energy rates are
calculated for each state, as presented in Table 1. In this study, the reference state is taken to be 25°C
at the pressure of 100 kPa. The thermodynamic properties of water, hydrogen and oxygen are obtained
using Engineering Equation Solver (EES) software package program. Note that state 0 indicates the
restricted dead state for the water, hydrogen and oxygen.
The chemical exergy is associated with the departure of the chemical composition of a system
from that of the environment. For the simplicity, the chemical exergy considered in the analysis is rather
a standard chemical exergy, based on the standard values of the dead state temperature of 25°C and
pressure of 100 kPa. Generally, these values are in a good agreement with the calculated chemical
exergy, relative to alternative specifications of the environment. In the analyses, the values of the
chemical exergies of the reactants and products for electrolyser are taken from Orhan et al. (2008a,b).
The energy and exergy efficiencies of the high temperature steam electrolyser unit are also
calculated, depending on the given parameters. It is found that energy and exergy efficiencies of the
HTSE are 87% and 86%, respectively. These are considerably consistent with the practical efficiencies.
One of the main factors affecting the hydrogen production cost is the temperature of electrolyser. Figure
4 illustrates the effects of the electrolyse temperature on the energy and exergy efficiencies.
The energy efficiency values for the HTSE unit vary between 80% and 87% while exergy
efficiency values for that range from 79 to 86%. Also, it is clear from this figure that energy and exergy
efficiencies increase with the electrolysis temperatures ranging from 473 to 1173 K. The details about
the system studied here for the performance analysis may be obtained from Sigurvinsson et al.
(2006,2007).
The overall theoretical energy and exergy efficiencies of the HTSE system are calculated and
given in Tables 5 and 6.

191
 89

Efficiency (%)
87
85
83
81 Energy Efficiency
79
Exergy Efficiency
77
75
473 573 673 773 873 973 1073 1173
Temperature (K)

Fig. 4. Energy and exergy efficiencies of the HTSE unit.

Table 5. The overall theoretical energy efficiency values of various conditions for HTSE.

ηel (%) ηes (%) ηth (%)

80 90 100 80 90 100
10 14.10 15.82 17.53 17.41 17.48 17.53
11 15.48 17.36 19.23 19.09 19.17 19.23
12 16.85 18.89 20.92 20.76 20.85 20.92
13 18.21 20.42 22.61 22.41 22.52 22.61
14 19.57 21.94 24.28 24.05 24.18 24.28
15 20.92 23.44 25.95 25.69 25.83 25.95
16 22.27 24.95 27.60 27.31 27.47 27.60
17 23.61 26.44 29.25 28.92 29.10 29.25
18 24.95 27.93 30.89 30.52 30.72 30.89
19 26.28 29.41 32.51 32.11 32.33 32.51
20 27.60 30.89 34.14 33.69 33.94 34.14

Table 6. The overall theoretical exergy efficiency values of various conditions for HTSE.
ηel (%) ηes (%) ηth (%)
80 90 100 80 90 100
10 11.87 13.34 14.82 14.79 14.80 14.82
11 13.05 14.67 16.29 16.26 16.27 16.29
12 14.23 15.99 17.76 17.72 17.74 17.76
13 15.40 17.32 19.22 19.18 19.21 19.22
14 16.58 18.64 20.69 20.64 20.67 20.69
15 17.76 19.96 22.15 22.10 22.13 22.15
16 18.93 21.27 23.61 23.56 23.59 23.61
17 20.10 22.59 25.07 25.01 25.05 25.07
18 21.27 23.91 26.53 26.46 26.50 26.53
19 22.45 25.22 27.99 27.91 27.95 27.99
20 23.61 26.53 29.44 29.35 29.40 29.44

Tables 5 and 6 show the calculation results of the ηoverall and ψoverall of different ηel, ηes and ηth. As given
in this table, the ηoverall and ψoverall are increased by increasing ηel, ηes and ηth. The overall theoretical
energy and exergy efficiencies of the system are calculated as 14.10-34.14%, and 11.87-29.44%
respectively, compared to the energy efficiency of the well-developed conventional alkaline water
electrolysis, which is about 27%. In the system studied, exergy destruction and exergy efficiency of heat

192
exchangers were also calculated. The values for exergy destruction and exergy efficiency of heat
exchangers are given in Table 7. The results in this table indicate that heat exchangers operate at
relatively high exergy efficiencies and that most exergy destructions in the overall system are due to
heat exchanger network. It is seen from this table that the maximum exergy destructions in the heat
exchanger system occur at HT-H exchanger as 1133 kW.

Table 7. Exergy efficiency and exergy destruction rate values of heat exchangers.

Heat Exchanger E xdest, Hex (kW)  Hex (%)

LT-H2 263 92.55

MT-H2 106 97.26
HT-H2 1133 78.08
LT-O2 114 86.51
MT-O2 345 82.75
HT-O2 163 89.43

Table 2 lists the enthalpy of formation values and the Shomate constants while Table 3 shows
the Gibbs free energy and standard chemical exergy of H 2, O2 and H2O. The thermodynamic
parameters in the range of 473-1173 K were basically calculated using Eq. (2), as shown in Fig. 5.

6
Energy Demand (kWh)

5
4 Total Energy Demand
3
2 Electrical Energy
Demand
1
Heat Demand
0
473 573 673 773 873 973 1073 1173
Temperature (K)

Fig. 5. Thermal and electricity supply to the studied system.

Here, the results show that the total energy demand increases slightly with the increase in
temperatures while electrical energy demand decreases, due to increasing the heat demand. The
decrease in the demand of electrical energy increases the thermal to hydrogen energy conversion
efficiency. The exergy demand of the system is illustrated in Fig.6.

5
Exergy Demand (kWh)

4
Total Exergy Demand
3

2 Electrical Exergy
Demand
1
Exergy inlet
0
473 573 673 773 873 973 1073 1173
Temperature (K)

Fig. 6. Exergy demand of the studied system.

193
 The total exergy demand of the system is constant with an increase in the temperature while
electrical demand decreases, due to the exergy inlet of the steam. The decrease in the electrical
demand increases the thermal to hydrogen exergy conversion efficiency, as illustrated in Fig. 6.

CONCLUSIONS
We have evaluated the overall theoretical energy and exergy efficiencies of the HTES process coupled
to the geothermal power plants to produce hydrogen in this study. We can expect a higher hydrogen
production yield at the high operation temperature. The promising research results of HTSE in
developed countries highlight the fact that it can be a suitable process for the next decades to consider
massive production of hydrogen. A well-developed HTSE might also be a promising potential
technology to resolve the long-term energy requirements in developed countries.
From this study, we can extract the following concluding remarks:
 Geothermal steam assisted high temperature electrolysis in countries, which are abundant due to
geothermal sources could possibly reduce the hydrogen production cost.
 The energetic and exergetic efficiencies of the overall system are obtained to be 87 and 86%,
respectively.
 The overall theoretical energy and exergy efficiencies of the system are calculated as 14.10-
34.14%, and 11.87-29.44% respectively, compared to the energy efficiency of the well-developed
conventional alkaline water electrolysis, which is about 27%.
 The maximum exergy destruction rate occurs within the HT-H2 exchanger with 1133 kW.
 The overall system exergy destruction rate is found as 9008 kW.
o
 For a geothermal temperature range of 500-600 C, the system consumes less electricity.
 It is confirmed from the exergetic point of view that the electrolyse temperature is one of the basic
parameters that affects the efficiency. Increase in the electrolyse temperature leads to an increase
in the exergy efficiency and a decrease in the exergy destruction rate.

References
ANL. 2004. Basic research needs for the hydrogen economy, report of the basic energy sciences
workshop on hydrogen production, storage and use, May 13-15, 2003. p.3.
Arnason B., Sigfusson T.I. 2003. Application of Geothermal Energy to Hydrogen Production And
nd
Storage, 2 German Hydrogen Congress, Essen 02.2003, Published in Proceedings.
Balta M.T., Dincer I., Hepbasli A. 2009, Thermodynamic assessment of geothermal energy use in
hydrogen production, International Journal of Hydrogen Production 34:2925-2939.
Balta M.T., Dincer I., Hepbasli A. 2010a. Thermodynamic performance comparison of some renewable
and non-renewable hydrogen production processes. 18th World Hydrogen Energy Conference 2010-
WHEC 2010, Parellel sessions book 2: Hydrogen Production Technologies-Part 1, Proceedings of the
WHEC, May 16-21, pp.331-341, Essen.
Balta M.T., Dincer I., Hepbasli A. 2011. Energy and exergy analyses of Mg-Cl thermochemical cycle,
International Conference on Hydrogen Production, 19-22 June 2011, Thessaloniki, Greece.
Costogue E.N., Yasui R.K. 1977. Performance data for a terrestrial solar photovoltaic/water electrolysis
experiment. Solar Energy 19:205–10.
Eisenstadt M.M., Cox K.E. 1975. Hydrogen production from solar energy. Solar Energy 17:59–65.
EP, The Exergoecology Portal, http://www.exergoecology.com/excalc/ Access date: May 14, 2013.
Ewan B.C.R, Allen R.W.K. 2005. A figure of merit assessment of the routes to hydrogen, International
Journal of Hydrogen Energy 30: 809 – 819.
Fujiwara S., Kasai S., Yamauchi H., Yamada K., Makino S., Matsunaga K., Yoshino M., Kameda T.,
Ogawa T., Momma S., Hoashi E. 2008. Hydrogen production by high temperature electrolysis with
nuclear reactor, Progress in Nuclear Energy 50: 422-426.
Herring J.S., O’Brien J.E., Stoots C.M., Hawkes G.L., Hartvigsen J.J., Shahnam M. 2007. Progress in
high-temperature electrolysis for hydrogen production using planar SOFC technology, International
Journal of Hydrogen Energy 32: 440–450.
Kanoğlu M., Ayanoğlu A., Abusoğlu A. 2011. Exergoeconomic assessment of a geothermal assisted
high temperature steam electrolysis system, Energy 36:4422-4433.

194
Laguna-Bercero M.A. 2012. Recent advances in high temperature electrolysis using solid oxide fuel
cells: A review, Journal of Power Sources 203: 4–16.
Mingyi L., Bo Y., Jingming X., Jing C. 2008. Thermodynamic analysis of the efficiency of high-
temperature steam electrolysis system for hydrogen production, Journal of Power Sources 177: 493–
499.
NIST, National Institute of Standards and Technology, http://webbook.nist.gov/
chemistry/ Access date:May 14, 2013.
O’Brien J.E., Stoots C.M., Herring J.S. 2005. High-Temperature Electrolysis for Hydrogen Production
from Nuclear Energy, INL research and development, 05-GA50193-19,
http://www.inl.gov/technicalpublications/Documents/3394826.pdf, Access date: May 14, 2013.
Orhan M.F., Dincer I., Naterer, G.F. 2008b. Cost analysis of a thermochemical Cu–Cl pilot plant for
nuclear-based hydrogen production, International Journal of Hydrogen Energy 33: 6006–6020.
Orhan M.F., Dincer I., Rosen M.A. 2008a. Energy and exergy assessments of the hydrogen production
step of a copper–chlorine thermochemical water splitting cycle driven by nuclear-based heat,
International Journal of Hydrogen Energy 33: 6456–6466.
Rosen M.A., Scott D.S. 1992. Exergy analysis of hydrogen production from heat and water by
electrolysis, International Journal of Hydrogen Energy 17(3): 199-204.
Rosen MA. 1995. Energy and exergy analyses of electrolytic hydrogen production. International Journal
of Hydrogen Energy 20(7):547–53.
Sigurvinsson J., Mansilla C., Arnason B., Bontemps A., Maréchal A., Sigfusson T.I., Werkoff F. 2006.
Heat transfer problems for the production of hydrogen from geothermal energy, Energy Conversion and
Management 47: 3543–3551.
Sigurvinsson J., Mansilla C., Lovera P., Werkoff F. 2007. Can high temperature steam electrolysis
function with geothermal heat? International Journal of Hydrogen Energy 32: 1174 – 1182.
Zhang W., Yu B., Xu J. 2012. Efficiency evaluation of high temperature steam electrolytic systems
coupled with different nuclear reactors, International Journal of Hydrogen Energy 37: 12060–12068.

195
 ENERGY AND EXERGY ANALYSES OF A COMBINED MULTIGENERATION SYSTEM
J. Hogerwaard and I. Dincer
Faculty of Engineering and Applied Science
University of Ontario Institute of Technology (UOIT)
2000 Simcoe Street North, Oshawa, ON, Canada, L1H 7K4
Emails: Janette.Hogerwaard@uoit.ca, Ibrahim.Dincer@uoit.ca

ABSTRACT
A renewable energy based system is proposed in the present study which will produce electricity, hot water,
space heating and cooling, fresh water, and hydrogen. Thermodynamic analyses – including energy and
exergy details – are performed to evaluate the feasibility of the system under the proposed operating
conditions, and parametric study is performed to assess the system for varying conditions. The baseline
system (as defined in the results section) has modeled energetic and exergetic efficiencies of 0.28, and 0.25,
and for the multi-generation system, values of 0.69 and 0.54 are achieved. In a parametric study performed
to reflect Canadian offshore wind speed and ocean current conditions, the multi-generation efficiencies for
the system ranged between values of 0.32 – 0.48 energetically, and 0.22 - 0.35exergetically.

Keywords: Combined system, multigeneration, energy, exergy, exergy.

INTRODUCTION
Renewable energy systems are a necessary part of the solution to reducing dependence on fossil fuels.
Resources such as solar, wind, biomass/biofuel, and ocean, are widely available world-wide, as clean
energy (fuel) suitable for various systems. There are many challenges that must be addressed in order to
make renewable energy systems comparable to fossil fuel-based systems in terms of providing consistent,
reliable power production, and economic feasibility. To maintain the necessary power demand, renewable
energy systems are often arranged in combination with fossil fuel systems, to either augment or offset
production by the fossil fuel based system.

One such system is proposed by Mousavi (2012), combining and offshore wind turbine, ocean tidal turbine,
and a natural gas fuelled microturbine for electricity production and battery storage. The study performed in
the research focuses primarily on the power management and control aspects of the combined system, and
addresses generation of electricity only. The system offers a good platform for integrating additional
processes to produce valuable commodities from the waste heat produced by the microturbine.
Multigeneration of both power and commodities through system integration is an effective method of meeting
demands, and can improve resource utilisation efficiency when compared to simpler, stand-alone systems.

Mini- and microturbine units offer a great deal of flexibility in design, with a wide range of generation
capacities(Pilavachi, 2002), and fuel/heat source flexibility. Various fuel combinations are possible, such as
combined gas and concentrated solar heating, and various alternate fuels such as biofuel(s) or landfill
gas(Reddy et al, 2012; Poullikkas, 2005; Muis et al, 2010), reducing the overall consumption of the primary
fuel. Exhaust heat from these turbines may be utilised in many different applications, depending on the
quality of the waste heat. Most gas turbine units incorporate preheating of air, pre-combustion, from
recuperated heat of the turbine exhaust gases. If this heat is not used, it represents a major exergy – and
energy – loss for the system. The exhaust gas heat energy can also be applied as process heat for
multigeneration processes – for example, low-to-mid temperature organic Rankine cycles, or to direct
heating processes such as desalination, or space heating and cooling systems (Dincer and Zamfirescu,
2012; Rubio-May et al, 2011; Chicco, 2010)

Wind turbines installed in offshore locations have certain advantages over inland installations. Firstly,
offshore wind velocity tends to be higher than inland wind velocities. The relationship between the electricity
3
produced is proportional to the cube of the velocity – that is, ∝V – thus, a small increase in the velocity

196
can have a large effect on the amount of electricity generated. Secondly, the wind velocity over the ocean
tends to be more constant than that of inland winds. Tidal stream turbines generate electricity from the
kinetic energy of the tidal currents. These types of turbines generate power from streams at much lower
velocities – in the range of 0.8 m/s to 3 m/s (NRC, 2010).

In this study, the system will be adapted to include multi-generation of several commodities, specifically:
electricity, hot water, space heating and cooling, fresh water, and hydrogen. An adapted system is proposed
and objectively assessed in the following sections through the application of energyand exergy analyses. A
parametric study of the system and subsystems is conducted to observe the performance under varying
conditions, based on Canadian resource data.

SYSTEMS DESCRIPTION
The Reference System
The investigated system is based on the renewable energy system described by Mousavi (2012)(Figure 1).
There are three main subsystems considered; an offshore wind turbine, a tidal turbine, and a microturbine. In
the reference case, all of the turbines are assumed to operate at rated capacity. The turbines are connected
to a main AC bus where the converted power can be distributed to consumers, or to a specific application.
The various control systems of the reference system are not described, but can be found in the reference
work.

Vwind PWT

Vwater POT

Ocean Tidal Turbine Offshore Wind Turbine

AC

AC Bus

CC

GT

a ex

(Gas) Micro-Turbine

Figure 1: Original system (adapted from Mousavi, 2012).

The Adapted MultigenerationSystem

For this work, the addition of sub-systems for hot water, space heating and cooling, fresh water, and
hydrogen production are considered as shown in Figure 2, which provide multi-generation of power and heat
for the required commodity production.Power generated from the system is shared between direct electricity
conversion, and electricity provided to the electrolysis of water for hydrogen (H 2) production. Hot exhaust
gases from the microturbine pass through a heat exchanger, transferring heat to incoming salt water for a
single stage flash desalination process for fresh water production. The heated salt water is then flashed to a
pressure at, or just below the saturation pressure of water at the inlet temperature, and passed into a
chamber at the same pressure. Water vapour and brine mist rise in the chamber, where demister baffles
remove any salt water droplets suspended in the rising vapour. Pure water vapour ( s) is condensed on
cooling tubes, and collected for both fresh water storage and hot water production. The water for hot water
production is heated in the secondary exhaust gas heat exchanger, and stored in an insulated tank. Warm

197
waste brine leaves the desalination unit, where it is then coupled in a heat exchanger with the condenser
and evaporator of vapour-compression units for cooling and heating, respectively.

Vwind PWT

Vwater POT

Ocean Tidal Turbine Offshore Wind Turbine

AC

AC Bus

(Gas) Micro-Turbine el,H2

f
H2
CC st,H2
Electrolysis

GT

4
a ex HX1 HX2
sw,o
10
6
SW 7 8 9
ex,o
12
FW 5
Heat Exchange
(Space Heating) s
sw,i
12a 12b

14 13 11 FWS HW
B
Flash FWS HW
EV Cond
desalination

HEAT COOL
COPH COPC
Figure 2: Proposed integrated multi-generation system.

ANALYSIS
The thermodynamic analysis; energy, entropy, and exergy balance equations, and efficiency definitions
isdescribed in this section. The analysis of auxiliary components (pumps, valves, etc.) is not included in the
analysis.

Offshore Wind Turbine:

The optimum power output and torque are defined by Mousavi (2012) by the following, with the variable CP
as the power coefficient of the turbine, which is the overall efficiency of the turbine, product of the
mechanical and energetic efficiencies of the turbine.

(1)

where the ratio of the optimal rotational velocity, ωopt, to the tip-speed-ratio, λopt, is equal to the ratio of the
wind velocity, Vw, to the radius of the turbine blade.

(2)

198
The a relationship for power is given by Mousavi (2012) as a function of the wind velocities, by

(3)

where the subscripts ra, c.in, and c.off, refer to rated, cut-in, and cut-off velocities of the turbine. The energy
and exergy balances for the wind turbine are given by Dincer and Rosen (2007) in the following set of
equations, where the subscripts 1 and 2 refer to the inlet and outlet sides of the wind turbine.

(4a)

(4b)

In the energy equation, WTis equal to the change in the kinetic energy is due to the change in the wind
velocity across the turbine, defined as .The exit velocity, V2, and pressure on either
side of the wind turbine, pi, are equal to

(5)

(6)

The energy and exergy efficiencies of the wind turbine are defined based on the useful output – in both
efficiencies, this is the power generated –over the total change in energy or exergy across the unit.

(8)
(9)

From the equations, it is clear that the way in which the reference state is defined is important in assessing
the performance of the turbine, since this is a main source of loss in the efficiency of the turbine. The
efficiency of the conversion process from generated power to user-ready electricity, ηel is addressed in the
efficiency of the overall system.

Tidal Turbine:
The balances for the tidal turbine are much the same as those of the wind turbine, replacing the properties
such as density, ρ, and specific heat, cp, with those for water, and defining the velocity of the flow through
the turbine as the velocity of the tidal current, Vt.

(10)

Typical values of CP for tidal turbines are in the range of 0.35 – 0.5, which is higher than the typical values
for wind turbines, which range from 0.25 – 3 (Elghali et al, 2007). The balances for energy, entropy, and
exergy are given by the same equations for the wind turbine, using the appropriate fluid properties of water
in place of air properties. The energy and exergy efficiencies of the tidal turbine aresimilar to those for the
wind turbine, for the associated tidal turbine values.

199
Microturbine:
The energy balances for the main components of the microturbine system; compressor, combustion
chamber, and gas turbine, are given by the following set of equations, for which the mass balance
applies. No heat losses are considered in the analysis, but should be included in practical
applications.

(11a)
(11b)
(11c)

The entropy and exergy balances for the microturbine system are defined for eachcomponent by the
equation sets for exergy destruction, which account for internal and external losses of the components.

(12a)

(12b)

(12c)

The energy and exergy for the microturbine system is defined below for the initial case of single-generation.
It will be redefined for multi-generation in a later equation to include the various additional outputs.

(13)
(14)

Energy and exergy efficiencies of any individual component is defined as the ratio of the rates of change in
useful energy for that component (i.e. – heat, work), to that of the input energy.

Electrolysis:
The hydrogen production process is based ontheprocess defined by Leveneet al, (2007) for hydrogen
production from wind-source electricity as input for an electrolyser. According to the authors, the process
requires 53kWh per kg of hydrogen (H2) produced, and has aspecific (per kg) energy efficiency of
0.74,based on a higher heating value forH2 of 141.8 MJ/kg (39.3 kWh/kg)(Cengel and Boles, 2006)
Therefore, the equation for the production of H2( H2) is defined for this system by thespecified efficiency
ofηH2 = 0.74, as

(15)

where elrefers to the portion of the electricity produced by the turbines (wind, tidal, and/or gas) directed to
the electrolysis process.

Desalination Process:
Fresh water produced from desalination of sea water is used for both fresh water storage and hot water
production. The overall energyand exergy balance equations are given for the desalination process, with the
boundary defined as the low-pressure chamber, by Eq. (17), and mass balance defined for fresh water
production rate, FW.

(16)
(17a)

200
 (17b)

Heat is transferred to the portion of the salt water stream to be flashed, sw,i, in the exhaust gas heat
exchanger. The process is described by the following equations for energy and exergy.

(18a)

(18c)

The rates of production of fresh water for storage, and for hot water production, are defined by the mass
balance of the total fresh water production rate,

(19a)
(19b)

From the equation, water for storage collects at a rate of (1- α)∙ FW in the storage tank, and the mass
fraction of fresh water for hot water production ( HW) is defined as α∙ FW, which passes through the second
exhaust gas heat exchanger.

(20a)
(20b)

Space cooling and heating:

The condenser of the cooling cycle is defined by energy balance of with the exiting waste brine streamfrom
the desalination process, by the equation

(21a)
(21b)

The evaporator of the heat pump is similarly defined by the following

(22a)
(22b)

For the system both the cooling cycle and the heat pump are designed for a specified coefficient of
performance (COP), where the useful energyfor the cooling cycle and heat pump are the evaporator and
condenser quantities for the refrigerant, respectively.

(23)

The overall efficiency for the entire system is given based on the same principle of a ratio of useful outputs to
total inputs.Net electrical output for the system is given by the following equation, which includes a term for
the conversion from DC to AC, ηem, which is an assumed value of 0.95.

(25)

(26)

201
RESULTS AND DISCUSSION

The characteristics of the wind turbine, tidal turbine, and microturbine are given in the following tables
(Tables 1-3). These data are used in the modeling of the system as described in the Analysis section. The
fuel used in the microturbine is assumed to be CH4, with a lower heating value of 50 MJ/kg, but could be any
gaseous fuel (for internal combustion), or solid fuel for an external combustion application.

a
Table 1: Wind and tidal turbine characteristics
Wind turbine Tidal turbine
Model VestasWT Generator Type V-25-200 Verdant power KHPS
Rotor diameter, d (m) 21 5.0
Swept area, A (m2) 415.5 19.6
RPM (rated) 14 ~35
Tip speed, Vtip (m/s) 16.75 9
Cut in speed, Vc.in (m/s) 3.8 0.8
Cut off speed,Vc.off (m/s) 25 -
Rated velocity, Vi (m/s) 13.8 2.1
Rated power, Pra (kW) 200 35
Drive train (mechanical) eff., ηm - 0.86
Total height, Ht (m) 41.5 6
Total lengthb, Lt (m) - 4.82
System mass (in air), (kg) - 3629
a. Turbine data obtained from Mousavi (2012),Kalantar (2010) for wind, Natural Resources Canada (2010) for tidal.
b. Total length refers to the distance from inlet to outlet across the turbine blades

c
Table 2: Microturbinecharacteristics
Microturbine
Rated power, Pra (kW) 250
Rated voltage, V (V) 660
Frequency, f (Hz) 60
Friction factor 0
c. Turbine data obtained from Mousavi (2012),Kalantar (2010).

Baseline modelresults
The wind and tidal turbinesare assumed to operate under the defined rated conditions. The rated values are
used to determine characteristic values for the energy efficiency, ηi, mechanical efficiency, ηm. The product
of these two efficiencies is the ratio of the converted power (PiT), to the total change in energy (or exergy) for
that stream across the turbine blades, and equal to the power coefficient, CP.

Table 3: Characteristic efficiency values for wind and tidal turbines

Wind turbine (WT) Tidal turbine (OT)
Energy eff., ηi 0.33 0.45
Drive train (mechanical) eff., ηm 0.91 0.86
Power coefficient, CPi 0.30 0.39
Exergy eff., ψi (rated conditions) 0.15 0.33

The power generated is plotted as a function of velocity, giving the characteristic curve for the turbines. The
velocity ranges are from the cut-in to the cut-off velocity of the wind turbine (Figure 3), and for the tidal
turbine, from the cut-in velocity to a reasonable high-end velocity of 5 m/s (Figure 4).

202
 250 0.5

h WT
200 0.4
y WT
CP

150 0.3
PWT (kW)

h, y
100 0.2

50 0.1

0 0.0
0.0 5.0 10.0 15.0 20.0 25.0 0.0 5.0 10.0 15.0 20.0 25.0
Vw (m/s) Vw (m/s)
(a) (b)
Figure 3: (a) Turbine power, and (b) Energy and exergy efficiency,of offshore wind turbine power generation
as a function of wind velocity.
50.0 0.5

CP
h OT
40.0 0.4
y OT

30.0 0.3
POT (kW)

h, y

20.0 0.2

10.0 0.1

0.0 0
0.0 1.0 2.0 3.0 4.0 5.0 0 1 2 3 4 5
Vw (m/s) Vw (m/s)
(a) (b)
Figure 4: (a) Turbine power, and (b) energy and exergy efficiency,of ocean tidal turbine power generation as
a function of water current velocity.

The energy and exergyefficiency,using thecaptured wind energy as the usefuloutput, is equal to the ratio of
power (calculated from the velocity relationships given in Eq. (4)) to the change in energy of the air passing
through the turbine sweep. The efficiencies in both turbines drop off drastically after the rated conditions are
surpassed, since the operational characteristics do not allow the turbines to generate more than the
specified capacity (for mechanical/functional reasons), while the kinetic energy available in the flow
continues to increase.

Microturbineand Subsystems
In order to determine the required state conditions for the microturbine system, the loads to be met are
defined. These details are provided in Table 4, and include; required fresh water and hot water supply rates,
hot water storage temperature, heat required for desalination of seawater (to meet required total fresh water
production), and the space heating and cooling requirements. For the analysis, auxiliary components of the
subsystems (such as pumps, vapour-compression cycle expansion valves, fans) are not included, and are
considered in terms of the major heat loads, and major work quantities (i.e. – compressors, turbines).
Furthermore, heat losses are limited to those associated with the streams leaving the system – that is, the
subsystems are assumed to have no external heat loss.

Table 4: Fresh water production and sensible heating and cooling load data
Fresh WaterProduction (Desalination) Space Heating and Cooling
HW FWS Heating Cooling
Temperature, T (°C) > 70°C To COP 2.5 2.5
3 3
Total daily prod. 5000 L/day (5 m /day) Volume (m ); ACH 200; 6 50; 0.5
HW Volume Fr., αHW 0.75 ΔT (Supply-air) To - 40°C To - 0°C

203
The required loads for the water and sensible heating and cooling applications are determined from basic
thermodynamic analyses, based on the assumed fresh water production requirements, and space (room)
conditions for the heating and cooling applications. These loads are used to determine minimum
performance characteristics for the microturbinesystem, which operates at the rated output of 250kW.

Table 5: Operating conditions, microturbine system

Parameter Value
Fuel (gas) Methane, CH4
Microturbine, MT (kW) 250
Air flow rate, a (kg/s) 0.985
Fuel flow rate, f (kg/s) 0.015
Exhaust flow rate, ex (kg/s) 1.0
Exhaust temperature, T3 (°C) 800
Lower heating value, LHV (MJ/kg) 50
Higher heating value, HHV (MJ/kg) 55.4
Isentropic efficiency (ηs,C = ηs,T) 0.85

Hydrogen is produced by the electrolysis of steam, requiring heat input to produce saturated steam, and
electricity input (supplied from wind or tidal power, as mentioned). These loads are determined based on the
hypothetical hydrogen electrolysis process discussed in the analysis section. The resulting output loads
(heat, energy and net power) are shown with the flow pathsfor the systems in Figure 5 for the rated cases of
the wind, tidal and micro-turbines.

CC = 568 kW

OT = 35 kW WT = 200 kW GT = 250 kW EX = 330 kW

HX1 = 163 kW DES

HX2 = 7.22 kW
st,H2 = 2.56 kW

el,H2 = 25 kW
C = 163 kW Cond = 6.22 kW EV = 61.15 kW
C,H = 40.8 kW

C,C = 1.78 kW
FWS = 1.125 m3/day

el,NET = 242 kW C = 4.44 kW H = 102 kW HW = 3.375 m3/day H2 = 11.27 kg/day

Figure 5: Flow path of energy, with rated case values.

In both cases, the fuel is considered the only external input, since the potential power of the wind and tidal
currents not captured by the turbines. Thus, for the initial case, the three systems are given their own
efficiency values. For the rated case, these are the CP values for the wind and tidal turbines. The efficiencies
for single and multi-generation cases are shown graphically in Figure 6.

An overall energetic and exergetic efficiency for the initial case is given as the total power generated divided
by the fuel input, however this value is more of an evaluation of valuable outputs to valuable (in terms of
cost, environmental, and exergetic value) input. These values are η = 0.28, and ψ = 0.25.

204
 1.0

0.8
hII

0.6 y II
Efficiency

CPOT
0.4
CPWT

0.2 hMT

0.0
Figure 6: System efficiencies and multi-generation efficiencies (rated case).

Parametric study
Performance of the system is tested under Canadian environmental conditions. Typical values are used as
inputs to the model, and the system performance is observed. Offshore wind speed and wind energy data
are generated using the Canadian Wind Energy Atlas tool (Environment Canada, 2008) for 30m
elevation.The data for wind speed for the offshore region of eastern Canada show speeds ranging from 7-10
m/s. Ocean current velocities are obtained from (NASA, 2012). Surface currents are considered, since the
depth is not specified. The data of interest is for the northwest region of the Atlantic, around the Canadian
east coast. For the same geographical region as that used for the wind values the surface ocean current
ranges from 0.3–1.0 m/s, however, mean current velocities at maximum passage depth(s) for high tidal
power density regions such as the Bay of Fundy are reported in the range of <1.0–5.25 m/s (Tarbotton and
Larson, 2006)

The efficiencies of the tidal and wind turbines are plotted against varying tidal current up and wind speeds,
and shown graphically in Figure 7.The tidal current speed is considered up to the rated value for the turbine.
The plot shows, as expected, that the performance of the wind turbine at varying velocities has a more
significant impact on the system performance due to its capacity within the system. Because the wind
speeds encountered in the considered region are generally below the rated wind speed for the turbine
considered, the efficiency is lower for both the single and multigeneration cases.

1 0.75
hI hI
hII hII
0.8 yI 0.6 yI
y II y II

0.6 0.45
h, y
h, y

0.4 0.3

0.2 0.15

0 0
0 0.4 0.8 1.2 1.6 2 0 2 4 6 8 10
Tidal current velocity, Vt (m/s) Wind velocity, Vw (m/s)
(a) (b)
Figure 7: System energy and exergy efficiencies vs. (a) tidal current, (b) wind velocity.

CONCLUDING REMARKS
In this paper, a renewable energy power generation system comprised of, offshore wind and ocean tidal
turbines, and a gas microturbine unit was investigated, and further integrated with commodity production
systems for sensible heating and cooling, desalination of seawater for fresh water storage and hot water,
and hydrogen production. The baseline system, operating at full rated capacity,results inenergetic and

205
exergetic efficiencies of 0.28 and 0.25. For the multi-generation case with commodity production, values of
0.69 and 0.54 are achieved.
NOMENCLATURE
2
A area, m
cp specific heat, kJ/kg∙K
COPi coefficient of performance (i=H for space heating, i=C for cooling)
CP power coefficient, ~
d rotor diameter, m
ex specific exergy, kJ/kg
x exergy rate, kW
h specific enthalpy, kJ/kg
flow enthalpy, kW
KE kinetic energy
mass flow rate, kg/s
p pressure, kPa
P power, kW
heat rate, kW
T temperature, K
V velocity, m/s; voltage, V
shaft work rate, kW

Greek Letters
α mass fraction
η efficiency
λ tip-speed-ratio
3
ρ density (kg/m )
ψ exergy efficiency
ω rotational speed (rad/s)
Subscripts
a air
C Compressor
CC Combustion chamber
C,C Cooling cycle compressor
C,H Heat pump compressor
Cond Condenser
d destruction
DES Desalination
el Electric
em electromechanical
EV Evaporator
ex exhaust
f fuel
FW fresh water
FWS fresh water storage
GT Gas (micro) turbine
H2 hydrogen
HW hot water
i{i = 1,2,3,…,n} state points (or inlet)
opt optimal
OT Ocean tidal turbine
st steam
SW salt water
sw fraction salt water
t tidal
WT Wind turbine
W wind

206
w water
Superscripts
a air
Q heat (exergy)

REFERENCES
Mousavi, G. "An autonomous hybrid energy system of wind/tidal/microturbine/battery storage."International
Journal of Electrical Power & Energy Systems 43, no. 1 (2012): 1144-1154.
Pilavachi, P. A. "Mini-and micro-gas turbines for combined heat and power." Applied Thermal Engineering 22, no.
18 (2002): 2003-2014.
Siva Reddy, V., S. C. Kaushik, and S. K. Tyagi. "Exergetic analysis of solar concentrator aided natural gas fired
combined cycle power plant." Renewable Energy 39, no. 1 (2012): 114-125.
Poullikkas, Andreas. "An overview of current and future sustainable gas turbine technologies."Renewable and
Sustainable Energy Reviews 9, no. 5 (2005): 409-443.
Muis, Zarina A., HaslendaHashim, Z. A. Manan, F. M. Taha, and P. L. Douglas. "Optimal planning of renewable
energy-integrated electricity generation schemes with CO2 reduction target." Renewable energy 35, no.
11 (2010): 2562-2570.
Dincer, I., and C. Zamfirescu."Renewable‐energy‐based multigeneration systems."International Journal of Energy
Research (2012).
Rubio-Maya, Carlos, Javier Uche-Marcuello, Amaya Martínez-Gracia, and Angel A. Bayod-Rújula. "Design
optimization of a polygeneration plant fuelled by natural gas and renewable energy sources." Applied
energy 88, no. 2 (2011): 449-457.
Chicco, Gianfranco. "Sustainability challenges for future energy systems." JOURNAL OF SUSTAINABLE
ENERGY 1, no.March (2010): 6-16.
Canada. Natural Resources Canada.Marine Renewable Energy – Wave, Tidal, and Water Current: Canadian
Technology Status Report. Ottawa: CanmetENERGY, 2010. ISBN: M154-40/2010E-PDF 978-1-100-
17618-5.
Kalantar, M., and S. M. Mousavi G. "Dynamic behavior of a stand-alone hybrid power generation system of wind
turbine, microturbine, solar array and battery storage."Applied Energy 87, no. 10 (2010): 3051-3064.
Dincer I., Rosen M. Exergy, Environment, and Sustainable Development. Elsevier: Oxford, UK, 2007.
Ben Elghali, S. E., M. E. H. Benbouzid, and Jean FrédéricCharpentier. "Marine tidal current electric power
generation technology: State of the art and current status." In Electric Machines & Drives Conference,
2007.IEMDC'07. IEEE International, vol. 2, pp. 1407-1412. IEEE, 2007.
Levene, Johanna Ivy, Margaret K. Mann, Robert M. Margolis, and AneliaMilbrandt."An analysis of hydrogen
production from renewable electricity sources."Solar energy 81, no. 6 (2007): 773-780.
Cengel, Yunus A., and Michael A. Boles.Thermodynamics: an engineering approach. McGraw-Hill Higher
Education, 2006.
Environment Canada.Canadian Wind Energy Atlas. (web). www.windatlas.ca. 21 August 2008. Accessed: 5
November 2012.<www.windatlas.ca/en/maps.php>.
NASA.Ocean Motion and Surface Currents: Drift Model. (web). oceanmotion.org. N.D. Accessed: 12 November
2012. <http://oceanmotion.org/html/resources/drifter.htm>.
Tarbotton, Michael, and Max Larson."Canada Ocean Energy Atlas (Phase 1) Potential Tidal Current Energy
Resources Analysis Background."Report to Canadian Hydraulics Centre (2006).

207
 PERFORMANCE ANALYSIS OF A SOLAR POND FOR DIFFERENT DIMENSIONS
1 2 2 3
I. Bozkurt , A. Atiz , M. Karakilcik and I. Dincer
1
Department of Mechanical Engineering, Faculty of Engineering, University of Adiyaman,
Adiyaman, 02040, Turkey
2
Department of Physics, Faculty of Sciences and Letters, University of Cukurova, Adana 01330, Turkey
3
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology (UOIT)
2000 Simcoe Street, North Oshawa, Ontario, L1H 7K4, Canada
ismail_bozkurt44@yahoo.com, ayhanklicc@hotmail.com, kkilcik@cu.edu.tr, ibrahim.dincer@uoit.ca

ABSTRACT
In this study, we develop of a theoretical solar pond in different geometries for the Adiyaman region in
Turkey and analyze its performance assessment through efficiency. Common insulating materials are
considered. In the parametric studies, the dimensions of the zones (e.g., upper convective zone, non-
convective zone and low-convective zone) are varied to investigate their effects. Although, the density
and temperature gradient of the inner zones are considered similar to each other, the energy
efficiencies of the storage zones become different due to varying dimensions of the pond. The shading
area decreases by increasing the surface area of the inner zones. The amount of incoming solar
energy to the upper surface is increased with the sizes of the solar pond. The efficiency of the solar
pond is determined to be maximum 31.76% for case4a, to be minimum 16.80% for case1b in August.
As a result, through careful determination of the dimensions and insulation parameter, incoming solar
radiation reaching the storage zone increases the efficiency of the solar pond.

Keywords: Solar energy, solar pond, heat storage, shading effect, energy, efficiency.

INTRODUCTION
Solar energy appears to be an important alternative energy source that will more likely be utilized in
the future. One main factor that limits the application of solar energy is that it is a cyclic, time-
dependent energy resource. Therefore, solar energy systems require energy storage to provide energy
during the night and overcast periods (Dincer and Rosen, 2003). One of the thermal energy storage
system is solar pond. Solar pond was discovered as a natural phenomena around the turn of the last
century in the Medve Lake in Transylvania in Hungary. In this lake, temperatures up to 70°C were
recorded at a depth of 1.32 m at the end of the summer season. Nowadays, mini model solar ponds
are also being constructed for various thermal applications (El-Sebaii et al.,2011). The solar pond
energetic and exergetic performance was studied experimentally and theoretically by Karakilcik and
Dincer (2008), Bozkurt and Karakilcik (2012), Karakilcik et al.(2013a). In the literature, some studies
have been undertaken on shading effect in solar ponds by various researchers (Hassab and El-Masry,
1991; Jaefarzadeh, 2004; Mazidi et al., 2011; Karakilcik et al., 2013b). The sunny area of the model
solar ponds are effected by the wall shading. So, when a model solar pond is constructed the sunny
area should be considered.In this paper, we present the development of a theoretical solar pond in
different dimensions for Adiyaman region in Turkey and study its performance assessment through
efficiency. Thus, the efficiency of the solar pond is predicted acording to the surface area and height.

SYSTEM CONSIDERED
A model solar pond is described as consisting of three layers. The top surface layer is known as upper
convective zone (UCZ). The first layer UCZ is formed due to upward salt transport, surface heating
and cooling and wave-action. The second layer is the non-convective zone (NCZ) acts as an insulating
layer of the pond. The density in NCZ increases with increasing depth of the gradient layer. The
bottom layer is a high temperature layer known as the heat storage zone (HSZ). This layer has a
constant temperature and salinity. Useful heat is usually extracted from this layer; its thickness
depends on the temperature and the amount of the thermal energy to be stored (El-Sabii et al., 2011).
The vertical walls in the inner zones of the pond are caused a shading area. The shading area is
defined as the area where solar radiation does reach or which is under shade. Thus, the incident solar
energy on surface of the pond decreases the pond’s efficiencies by shading from vertical side walls.
Therefore, it is important to better understand of the shading effect on performance of the solar pond
(Karakilcik et al., 2006). So, we study now, in order to determine the shading effect on efficiencies of
different dimensions’ solar pond for Adiyaman region in Turkey.

208
ANALYSIS
Sunny area of inner zones
In this section, we determine the sunny area for different surface and height. The angle of refraction,
which is the angle that the refracted ray makes with the normal line, is defined by the Snell law as:

θ r  sin
1
1.33sin θi  (1)

where iis called incidence angle and is defined as:

θ i  cos
1
cosδ d cos cosθh   cosδ d cos  (2)

where dis declination angle which can generally be calculated by the Cooper equation as
 
δ d  23.45sin 360(284  n)/365 . n is any day of the year as (1n365) (Duffie and Beckman,1991). 
is calculated according to latitude angle. (90  +90) is latitude degree which is find out solar
elevation angle and this is plus (+) for the Northern hemisphere and minus () for the Southern
hemisphere. We use 38°23' East longitude and 37°75' North latitude for the location of Adiyaman
University in Adiyaman, Turkey.θh is the hour angle which θh is equal to zero while the sun is at the
00
highest point in the sky at 12 hour and its value in the morning becomes positive (+) and also in the
00
afternoon becomes negative (). According to per 15° of longitude for per hour, for example; at 11
00
am, +15° but for it is 15° at 13 pm and for a special case, z = 90°, hour angle is defined as
θ h  cos
1
 tan(  )tan(δd ). The day length (the smallest sunshine duration) is defined as
N  2/5θ h as given by Karakilcik (1998). The sunny area of HSZ for model solar ponds are defined as

A sa.  A sur. - A sh. (3)

where Asa. is the sunny area of HSZ, Asur is the surface area of the model solar pond, Ash.is the shading
area of HSZ. Ash.is defined as

A sh.  I wdIHSZ (4)

Here, Iwd is the width of the model solar pond.lHSZ is the length of the shading area in HSZ’s middle point
from the surface. lHSZ is defined by Karakilcik et al. (2006) as

IHSZ  hHSZ tanθr (5)

where hHSZ is the height of HSZ’s middle point from the surface.
Note that χ sa. is the ratio of the sunny area of the model solar pond and determined for for NCZ and
UCZ as follows:

A sa.
χ sa.  (6)
A sur.

The refractive index of the salty water and water were assumed to be about the same. The thickness
of NCZ and UCZ was assumed to be fixed. NCZ’s thickness was taken as 0.5 m. UCZ’s thickness was
taken as 0.1 m. HSZ’s thickness varies with depth. The thicknesses of the solar pond zones for
different dimensions cases are given in Table 1.

209
Table 1.The thicknesses of the solar pond zones for different dimensions cases.
Surface
Case 2 Depth (m) HSZ (m) NCZ (m) UCZ (m)
Area (m )
(a) 1.50 0.90 0.50 0.10
1 11
(b) 2.00 1.40 0.50 0.10
(a) 1.50 0.90 0.50 0.10
2 22
(b) 2.00 1.40 0.50 0.10
(a) 1.50 0.90 0.50 0.10
3 33
(b) 2.00 1.40 0.50 0.10
(a) 1.50 0.90 0.50 0.10
4 44
(b) 2.00 1.40 0.50 0.10

Energy analysis of model solar pond

Fig. 1 illustrates the inner zones of the solar pond and the sunny area of the zones. Inner zones
consist of the salty water layers with various densities.
ce
en
id
c
in
ar
ol
S

Normal

Ɵi

UCZ

NCZ
Ɵr

HSZ
ea
ar
ng

ea
di

ar
ha

ny
S

un
S

Figure 1. The inner zones of the solar pond and the sunny area of its zones.

It is very important to determine the energy efficiency of the solar pond. To calculate the energy
efficiency, we used the temperature distribution, the incident radiation reaching the heat storage zone,
the ambient temperature and the insulation parameters. Solar energy is transmitted through UCZ and
NCZ, after attenuation, to HSZ. Part of the transmitted solar energy from NCZ to HSZ is reflected from
the bottom and the majority of the solar energy is absorbed in HSZ. Therefore, the solar energy is
collected and stored in HSZ. The radiation heat losses of the solar pond are neglected. Because, solar
ponds are working at low temperature. At the same time, the convection heat losses from the solar
pond are prevented by NCZ, with no heat losses. The energy efficiency of the solar pond can be
defined as

Q stored Q bottom  Q up  Q side 

η  1 (5)
Q in,HSZ Q in,HSZ

where Qstored is stored heat energy in HSZ of the solar pond. Qin,HSZ is the amount of net solar energy
absorbed by HSZ. Qbottom is the total heat loss to the bottom wall from HSZ. Qup is the heat loss from
HSZ to the above zone. Qside is the total heat loss to the side walls of HSZ. Substituting equations for
each parameter in Eq. (5) provides us with the following energy efficiency of the solar pond:

 k w A sur. k A k A 
 THSZ  Ta   s sur. THSZ  TNCZ   w HSZ THSZ  Ta 
η  1
 Δx bottom Δx HSZ-NCZ Δx side  (6)
βEA sa. 1  F  hx  δ 

210
where E is the total solar energy reaching the solar pond surface, A HSZ is the side area of HSZ, F is the
fraction of energy absorbed at a region of δ-thickness, h is solar radiation ratio; Asa. is the sunny area
of HSZ, Ta is the ambient air temperature, k w is the thermal conductivity of the side and bottom walls,
ks is the thermal conductivity of the salty water, Δx bottom is the thickness of the bottom wall, Δxside is the
thickness of the side wall, ΔxHSZ-NCZ is the thickness of HSZ’s middle point and NCZ’s middle point, β is
the fraction of the incident solar radiation that actually enters the pond and is given by Hawlader
(1980).

2 2
 sinθ i  sinθ r   tanθ i  tanθ r 
β  1  0.6   0.4  (7)
 sinθ i  sinθ r   tanθ i  tanθ r 

where θi and θr are the incidence and refraction angles. h represents the ratio of the solar energy
reaching the depth in the layer I to the total solar incident falling on to the surface of the pond and is
given by Bryant and Colbeck (1977) as

 (x I  δ) 
hI  0.727  0.056ln  (8)
 cosθ r 

where xI is the thickness of the layer,  is the thickness of the layer in the UCZ where long-wave solar
energy is absorbed. The following equation is employed to estimate the temperature distributions in
HSZ, theoretically as follows:

  k w A sur.  
βEA sa. (1  F)h(x  δ) -  (THSZ,K - Ta,K ) 
Δt   Δx bottom  
THSZ,K 1  THSZ,K   
m HSZ C HSZ   k s A sur. (T   k w A HSZ  (9)
  Δx HSZ,K - TNCZ,K ) -  (THSZ,K - Ta,K )

  HSZNCZ   Δx side 
where T(HSZ,K+1) is the temperature of HSZ at time (K+1). T(HSZ,K) is the temperature of the HSZ at time
(K). T(a,K) is the of air at time (K). T(NCZ,K)is the temperature of NCZ at time (K). mHSZ is the mass of
HSZ. CHSZ is the heat capacity of HSZ.

Figure 2. Comprasion of the sunny area ratio of HSZ for 1.50 m depth and various surface area.

211
RESULTS AND DISCUSSION
The efficiency of the solar pond depends not only on thermal energy flows (e.g., heat losses and/or
heat gains in the zones), but also on incoming solar radiation (e.g., accounting for reflection,
transmission and absorption). Moreover, shading decreases the performance of the solar pond
(Karakilcik et al., 2013a). In this study, we present the results of the theoretical calculations for the
sunny area of the different dimensions model solar ponds. At the same time, the model solar pond
energy efficiencies were determined. These results show how sunny area effect the efficiency of the
solar pond. Fig. 2 shows the variations of the sunny area according to the surface area during a year.
As seen in Fig. 2, the sunny area is effected by different surface area cases. As apparent,
there are significant differences between sunny area. The sunny area distributions are calculated by
using Eq. 4 during a year. In Fig.2, the theoretical sunny area ratio according to the surface area is
observed to be maximum 33.21%, 64.75%, 76.50% and 82.37% for case1a, case2a, case3a and
case4a in June, respectively. The amount of the sunny area for HSZ is increased with surface area of
the solar pond. The highest sunny area is observed for case 4a. The sunny area of HSZ for 1.5 m
depth is determined to be minimum 3.33%, 48.38%, 65.59% and 74.19% for case1a, case2a, case3a
and case4a in December, respectively. As seen in Fig.2, very little solar energy reaches to HSZ for
case1a. The selection of solar pond dimensions is very important for energy efficiency. This sunny
area can be used to determine of the solar pond dimensions.
Fig. 3 shows the variations of the sunny area ratio for HSZ during a year for 2.00 m depth.As
seen in Fig.3, the theoretical sunny area ratio of HSZ is observed to be maximum 25.89%, 56.35%,
70.90% and 78.18% for case1b, case2b, case3b and case4b in June, respectively. The amount of the
sunny area for HSZ is decreased with depth of the solar pond. The sunny area of HSZ is determined to
be minimum 0%, 36.09%, 57.39% and 68.04% for case1b, case2b, case3b and case4b in December,
respectively. As seen in Fig.3, solar energy does not reach to HSZ for case1b during four months
(November, December, January, February) and this profile shows that the dimensions is not suitable
for solar ponds. So, we can select other cases according to the size of the application area.

Figure 3. Comprasion of the sunny area ratio of HSZ for 2.00 m depth and various surface area.

Fig. 4 shows the variations of the sunny area ratio for NCZ during a year for different surface
area cases. As seen in Fig.4, the theoretical sunny area ratio of NCZ is observed to be maximum
76.50%, 88.25%, 92.17% and 94.12% for case1, case2, case3 and case4 in June, respectively. The
amount of the sunny area for NCZ is increased with the surface area of the solar pond. The sunny
area ratio of NCZ is determined to be minimum 66.23%, 83.11%, 88.74% and 91.56% for case1,
case2, case3 and case4 in December, respectively. The sunny area ratio of NCZ is changed a little
during a year because NCZ is close to the surface.

212
Figure 4. Comprasion of the sunny area ratioof NCZ for various surface area.

Fig. 5 shows the variations of the sunny area ratio for UCZ during a year for different surface
area cases. As seen in Fig.5, the theoretical sunny area of UCZ is observed to be maximum 96.64%,
98.32%, 98.88% and 99.16% for case1, case2, case3 and case4 in June, respectively. The amount of
the sunny area ratio for UCZ is increased with the surface area of the solar pond. The sunny area ratio
of UCZ is determined to be minimum 95.08%, 97.54%, 98.36% and 98.77% for case1, case2, case3
and case4 in December, respectively. The sunny area of UCZ is effected a little because UCZ is the
surface zone of the solar pond.

Figure 5. Comprasion of the sunny area ratio of UCZ for various surface area.

The temperature of the model solar pondswere determined by using Eq. 9. Fig. 6 shows
average temperature distributions inside the pond during the months of June, July and August. As
shown in Fig. 6, the maximum average temperature of HSZ is determined as 87.12 °C in August for
case4a. The reasons of the low temperature in HSZ for especially case1a-1b are that small surface
area, shading by the side vertical walls, high depth. So, the reaching solar energy of HSZ is decreased
by these negative effects.

213
 100
Case1a Case2a
Case3a Case4a
90
Case1b Case2b
Case3b Case4b
80
Temperature (°C)

70

60

50

40

30
June July August
Month

Figure 6. The temperatures distributions of the model solar pond for the various cases.

In a solar pond, the efficiencies are low since the reaching solar energy of HSZ is much smaller than
the incident solar radiation on the zone surfaces. To increase performance of the solar pond, the
reaching solar energy of HSZ should be increased selecting the correct size of the solar pond before
construction. The solar pond efficiencies are calculated by using Eq. 6 for different cases. As seen in
Fig. 7, the efficiency of the solar pond to be maximum 31.76 % for case4a to be minimum 16.80 % for
case1b in August. Fig. 7 shows that the sunny area of the solar pond is an important effect on
performance of the solar pond. Increasing the sunny area of the pond allows much more solar
radiation to reach HSZ and increase the thermal potential of the pond and hence its performance. We
analyze the energy efficiency of solar pond for different cases. The maximum efficiencies of the solar
pond are seen to occur for case4a.

0.35
Case1a Case1b Case2a Case2b
Case3a Case3b Case4a Case4b
0.3

0.25
Efficiency

0.2

0.15

0.1

0.05

0
June July August
Month
Figure 7. The efficiencies of the model solar pond for various cases.

CONCLUSIONS
In this study, we carried out the effect of the solar pond’s size to see their effect on the energy
efficiency. The solar energy reaching depends on the zone thicknesses, shading area of the solar
pond. So, to increase the efficiency of the solar pond, the sizes of the solar pond should be predicted
to achieve higher efficiency. Meteorogical data are used to determine the efficiencies for the various

214
sizes. As expected, the energy efficiencies of the solar pond are affected by the sizes of the solar
pond, strongly. This result helps to select the sizes of the solar pond before construction.

NOMENCLATURE
2
A area (m )
C heat capacity (J/kg°C)
E total solar energy reaching to the pond (J)
F absorbed energy percentage at a region of -thickness
HSZ heat storage zone
k thermal conductivity (J/m°C h)
l length (m)
m mass (kg)
n number of day
N day length
NCZ non-convective zone
T temperature (°C)
Q heat energy (MJ)
UCZ upper convective zone

Greek letters
 thermal energy efficiency
 thickness where the long wave solar radiation is absorbed (m)
 incident beam entering rate into the water
 latitude angle (rad)
 ratio
 angle (rad)
x thickness of horizontal layers (m)

Subscripts
a air
d declination
h hour
i incident
r refracted
sa. sunny area
side side wall
stored heat stored
s salty water
sur. surface
sh. shading
up just above zone
w wall
wd width

References
Bozkurt, I. and Karakilcik, M.,2012.The daily performance of a solar pond integrated with solar
collectors.Solar Energy86:1611-1620.
Bryant, H.C. and Colbeck, I. 1977. A solar pond for London. Solar Energy19:321.
Dincer, I., Rosen M.A.2003.Thermal Energy Storage: Systems and Applications. New York:John Wiley
and Sons.
Duffie, J.A., Beckman, W.A.1991.Solar engineering thermal processes (3rd ed.). New York:John Wiley
and Sons.
El-Sebaii A.A., Ramadan, M.R.I., Aboul-Enein, S. and Khallaf A.M.2011. History of the solar ponds: A
review study. Renewable and Sustainable Energy Reviews15:3319-3325.

215
Hassab, M.A., El-Masry, O.A.1991.Effects of edges on solar-energy collection in small solar ponds.
Applied Energy 38/2:81-94.
Hawlader, M.N.A. 1980. The influence of the extinction coefficient on the effectiveness of solar ponds.
Solar Energy 25:461-464.
Jaefarzadeh, M.R.2004. Thermal behavior of a small salinity-gradient solar pond with wall shading
effect. Solar Energy77:281-290.
Karakilcik, M.1998. Determination of the performance of an insulated prototype solar pond.Ph.D.
Thesis, Cukurova University, Adana-Turkey (in Turkish).
Karakilcik, M., Kıymaç, K., Dincer, I. 2006. Experimental and theoretical distributions in a solar pond.
Int. Journal of Heat and Mass Transfer 49:825-35.
Karakilcik, M. and Dincer, I. (2008). Exergetic performance analysis of a solar pond. International
Journal of Thermal Sciences47:93-102.
Karakilcik, M., Bozkurt, I. and Dincer, I. 2013a. Dynamic exergetic performance assessment of an
ıntegrated solar pond. International Journal of Exergy12:70-86.
Karakilcik, M., Dincer, I., Bozkurt, I., Atiz, A.2013b. Performance assessment of a solar pond with and
without shading effect. Energy Conversion and Management 65:98-107.
Mazidi, M., Shojaeefard, M.H., Mazidi, M.S., Shojaeefard, H.2011. Two-dimensional modeling of a
salt-gradient solar pond with wall shading effect and thermo-physical properties dependent on
temperature and concentration. Journal of Thermal Science 20:362-370.

216
 EFFICIENCY ASSESSMENT OF CRUDE OIL DISTILLATION SYSTEMS
1 1
S. Zafar and I. Dincer
1
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology,
2000 Simcoe Street North, Oshawa, Ontario L1H 7K4, Canada
Emails: sayem.zafar@uoit.ca and ibrahim.dincer@uoit.ca
*Corresponding author: sayem.zafar@uoit.ca Phone: +1 416-358-7006

ABSTRACT
The assessment of energy and exergy efficiencies is conducted on a crude oil distillation system for three
different cases. Each case defines efficiency in a unique way with different inputs and outputs. First case
treats the heat transfer rate as useful output and the heat gained from the heaters as input. For the second
case, we consider the heat provided by the heaters as input and the heat transfer rate in the distillation
system as output. The third case treats the heat provided by the heaters as input and heat transfer rate along
with the heat of exhaust gases as useful output. The system efficiencies are studied parametrically by
changing the amount of heat transfer rate and the ambient temperatures. The results show that case 1 has
the baseline energy and exergy efficiencies at 53% and 25.3% respectively. Case 2 has the efficiencies at
40% for energy and 23.3% for exergy efficiency. Case 3 efficiencies are at 72% for energy while exergy
efficiency is at 65%. Case 1 has the highest efficiencies, followed by case 3 then the least efficient was case
2. Utilizing the unused energy, as a useful input to some other system, improves the overall efficiency of the
plant and saves operating cost while making the system more environmentally friendly.

Keywords: Energy, exergy, efficiency, crude oil, distillation.

INTRODUCTION
Energy is defined as the maximum work producing ability by a system or a flow of matter as it comes to
equilibrium. The exergy of an energy form or a substance is a measure of its usefulness or quality or
potential to cause change (Rosen and Dincer, 1997). Exergy analysis is an effective thermodynamic method
for design and analysis of thermal systems while it is an efficient technique for revealing the improvement
capacity of thermal systems (Al-Muslim et al., 2003). With energy and exergy precisely known, it becomes
easier to evaluate their efficiency for any given system.

In thermodynamics, efficiency is a concept utilized to specify the effectiveness of the energy conversion
process. Efficiency is often misused and is a cause of confusion (Kanoglu et al., 2007). This confusion exists
because efficiency is often used without being properly defined first (Cengel and Boles, 2006). Basically,
efficiency can be described as the ratio of output against the input. This definition of efficiency holds true for
all thermodynamic systems and is clearly understood. However the output and input parameters are specific
to a system or a component which needs to be specified clearly for efficiency assessment.

Although, efficiencies change with the operating and environmental conditions, the overall efficiency can also
be improved if the exhaust energy is used as an input to some other system. For large plants and systems,
the recovered energy can add up to large values and can cause significant increase in the system’s overall
efficiency (Farooque et al., 2008). Even economically, it is feasible to recover a large proportion of the waste
heat generated, rather than reject it to the atmosphere (Talbi and Agnew, 2002).

Improving the energy efficiency of crude oil distillation plant is of great interest since they consume almost
2% of the total crude processed energy (Bagajewicz and Ji, 2001, Errico et al., 2009). Crude oil distillation
system is a part of crude oil refining process where crude oil compounds convert to gases when heated at
high temperature. These gases become cooler as they move along the vertical column. When gaseous
compounds cool below their boiling point, they condense to be drawn off the distillation column at various
heights (Dincer and Rosen, 2007). The crude oil is processed in two towers for which the atmospheric tower
separates the light hydrocarbons while the vacuum tower separates the heavier hydrocarbons. The products
of the crude oil distillation plant can be either final products or feedstock to other plants for further processing
(Al-Muslim et al., 2005).

Numerous researchers have studied the efficiencies of crude oil distillation systems. Al-Muslim et al. (Al-
Muslim et al., 2003) have conducted an energy and exergy study to precisely evaluate the exergy losses in a
one-stage and a two-stage crude oil distillation unit composed of a heating furnace and atmospheric
distillation column and to study their energy and exergy efficiencies. Their crude oil distillation system was
further evaluated to see the effects of reference states on exergy and energy efficiencies (Al-Muslim et al.,
2005).

217
The present paper conducts a combined energy and exergy study to propose three different efficiencies at
different case scenarios and make evaluations and comparisons. Each case represents the possibility of
using the input and output in different perspectives. Anything exiting or entering the distillation system cannot
be treated as useful output or input, respectively, without proper assessment. Efficiency is seen through
usefulness-perspective. Multiple exergy and energy definitions, for the overall distillation system, is described
in the paper and evaluated for a range of surrounding conditions. Exergy destruction for each component in
each case is also presented.

SYSTEM DESCRIPTION
The crude oil distillation system, considered for a comprehensive efficiency assessment, is taken from Al-
Muslim et al. (2003; 2005), as a part of the petroleum refining process. The system taken from this literature
has two columns, namely, atmospheric distillation unit (ADU) and vacuum distillation unit (VDU). ADU
separates the lighter hydrocarbons under atmospheric pressure. Before being fed to ADU, crude passes
through Heater 1 (station 1 to 2) where its temperature increases. Superheated steam enters from the bottom
of the column (station 18) to reduce the partial pressure in the column and thus enhancing vaporization and
separation of the crude oil. The gaseous hydrocarbons get extracted at their corresponding trays while the
heavier hydrocarbons get collected at the bottom of the column as residue. The atmospheric residue (station
10) passes through Heater 2 before entering the VDU (station 11). Steam is also added in the VDU (station
20) for the same purpose as in ADU. Fig. 1 shows the schematic of the studied system. Further details are
available elsewhere (Muslim et al., 2003; 2005).

Fig. 1: A schematic diagram of the two-stage crude oil distillation system (crude heating furnace, E1, is heater 1 and E2 is heater 2. Atmospheric
distillation unit (ADU) is shown as T1 and vapor distillation unit (VDU) is shown as T2. PA is pump-around circuits which are used to control the
temperature profile) (Muslim et al., 2005).

218
ANALYSIS
A study is conducted to assess the efficiencies of overall crude oil distillation system by considering different
scenarios as input and output. The first case represents the heated crude oil as input while the heat
consumed in the distillation process, to convert the compounds to gases, as output. It is assumed that the
unabsorbed heat from the heater is used as an input to some other system hence does not get wasted. The
heat transfer rate, in the distillation system, is considered useful output whereas the heat from exhaust gases
is considered useless. The energy efficiency for case 1 can be defined as follows:

ηC1= (1)

where is the energy rate which is the total enthalpy rate and the numbers in subscript correspond to the
station numbers shown in Fig. 1 and it values tabulated in Table 1. The heat transfer rate, , represents the
heat rate utilized in the distillation process to convert the compounds to gases.
The equation describing the exergy efficiency is given as follows:

ΨC1= (2)

where the exergy rate is denoted by x and the total exergy destruction, in the system, is denoted by .
The surface mean temperature is defined as T while T0 is the ambient temperature.
The second case treats heat provided by the heaters as input while the heat transfer rate in the
distillation process as output. This case assumes the unabsorbed heater heat gets wasted hence not
considered as a useful output. The energy efficiency of case 2 can be written as follows:

ηC2= (3)

where the subscripts ADU stands for atmospheric distillation unit and VDU for vacuum distillation. The
subscripts H1 and H2 represent the heater before ADU and heater before VDU, respectively. The exergy
efficiency equation can be described as follows,

ΨC2= (4)

The third case presents in which all the exhaust is considered as useful output while heater heat is assumed
to be the input. This case treats the exhaust gases as well as the unused heat from the heaters as a useful
input to some other system. The energy efficiency equation for case 3 is described as:

ηC3= (5)

where represents the sum of energy from all the exiting gases, defined as follows:

(6)

The exergy of exhaust gases, , is the defined as the sum of exergy of all exhaust gases below:

(7)
Hence, the exergy efficiency equation becomes

ΨC3= (8)

The corresponding thermodynamic properties of each station in the distillation system are tabulated in Table 1. The
ambient conditions for pressure and temperature are 102 kPa and 298 K respectively.

219
 Table 1: Operating condition parameters at different stations (Muslim et al., 2003; 2005).
o
# T ( C) P (kPa) m (kg/s) s (kJ/kg K) h kJ/kg)
1 25 102 507.6221 4.440448 34.39055
2 352.3746 102 507.6221 6.615505 1040.125
3 68.00522 206.6044 1.09E-12 4.216201 760.3895
4 68.00522 206.8148 0.566983 2.509089 164.5537
5 68.91907 208.4696 15.56693 3.039461 348.8646
6 68.86612 211.7791 32.14957 2.323762 182.8285
7 91.4 215.91 75.812 4.2661 186.98
8 157.1815 219.2254 53.9682 5.015905 323.7495
9 261.4419 223.3622 58.0393 5.878275 555.9205
10 342.2222 228.8781 260.1879 6.557474 750.3401
11 408.8889 228.8781 260.1879 6.858651 945.6614
12 203.2595 11.10056 10.21269 6.830466 1125.868
13 203.2595 11.10056 0.834219 5.768256 398.4854
14 219.95 11.1 22.617 5.8602 438.4
15 286.3149 11.92794 39.7135 6.238811 602.1979
16 312.649 13.16899 46.74861 6.426736 666.1357
17 299.5011 15.23742 142.2034 6.386489 628.2944
18 353.3333 515.0106 5.415384 7.651651 3172.316
19 68.00522 206.8148 0.517354 0.96201 284.4449
20 768 413.6856 2.141962 8.831165 4075.276
21 0 0 0 0 0
22 1100 102 600 7.986952 1180.016
23 353.3333 102 600 7.100531 329.1276
24 1100 102 60 7.986952 1180.016
25 357 102 60 7.106715 333.0128

Table 2 shows the baseline values for enthalpy, , heat transfer rate in the distillation columns, cv, exergy,
x, and exergy destruction, . Heater 1 is the heater that provides the heat to the crude feed entering the
ADU while heater 2 heats the crude entering the VDU. These values are used as baseline figures to conduct
the parametric studies on the system.

Table 2: The baseline values used in the study (Muslim et al., 2003; 2005).

Heater1 ADU Heater2 VDU

Σ in (kW) 725467 545170 266030 287044

Σ out (kW) 725467 270692 266030 166148

cv (kW) 0 274478 0 120896

Σ xin (kW) 589496 500365 223817 248621

Σ xout (kW) 484221 216749 213858 124560.5

(kW) 105275 366121 9960 167033

220
RESULTS AND DISCUSSION
The exergy and energy efficiency assessment is conducted on crude oil distillation system. The efficiency
assessment is conducted on the overall system which consists of ADU and VDU together. The crude oil
distillation system is assessed based on the model presented by Al-Muslim et al. (2003; 2005). Throughout
the process, the crude oil feed gets heated twice by two heaters. During the process, some heat is utilized to
produce desired compounds and when the crude exits, it exits with high temperature and pressure. The
study is conducted to assess the efficiencies by comparing different energy utilization scenarios.

0,8

0,7

0,6

0,5
Efficiency

Energy
0,4

0,3

0,2 Exergy

0,1

0
395 400 405 410 415 420
Heat transfer rate [kW]

Fig. 2: The changes of energy and exergy efficiencies with respect to consumed heat rate for case 1.

Fig. 2 shows the change in exergy and energy efficiency for case 1. This case treats the heat
transfer rate in the distillation column as useful output and the heated incoming crude feed as input. Heated
crude feed included the absorbed heat from the heaters. As the heat transfer rate, in the distillation column
increases, the efficiency increases. Heat transfer rate can increase if distillation involves higher amount of
compounds that require higher heat for distillation. For the baseline conditions, as shown in Table 2, the
energy and exergy efficiency comes out to be 53% and 25.3% respectively. More heat transfer rate in the
distillation plant means more heat is being used to separate compounds. The trend shows the linear increase
in the efficiencies with energy efficiency increases faster than exergy efficiency. Exergy efficiency change is
slower than energy efficiency since consumed exergetic heat rate is compared with the ambient temperature.

0,8
0,7
0,6
0,5
Efficiency

Energy
0,4
0,3
Exergy
0,2
0,1
0
395 400 405 410 415 420
Heat transfer rate [kW]

Fig. 3: The changes of energy and exergy efficiencies with respect to consumed heat rate for case 2

221
 Exergy and energy efficiency for case 2 are shown in Fig. 3. In this case, the heat transfer rate in the
distillation column is considered as useful output. The heat, provided by the two heaters, is treated as input.
The input is higher in this case as compared to case 1 since in this case, the provided heat is considered as
input whereas in the case 1, just the absorbed heat is input. As the system heat rate consumption increases,
the efficiency increases. More heat transfer rate in the distillation plant means more heat is being used to
separate compounds. For the baseline conditions, energy efficiency is 40% while exergy efficiency is 23.3%.
Case 2 has lower efficiencies as compared to case 1 since case 2 assumes heaters’ input unlike case 1
which assumes only the absorbed heat from heaters as input. The trend shows the linear increase in the
efficiencies with energy efficiency increases faster than exergy efficiency.

0,8
0,7
0,6
0,5
Efficiency

Energy
0,4
0,3
0,2 Exergy
0,1
0
395 400 405 410 415 420
Heat transfer rate [kW]

Fig. 4: The changes of energy and exergy efficiencies with respect to consumed heat rate for case 3.

Exergy and energy efficiency for case 3 are shown in Fig. 4. In this case, the heat provided by the
two heaters is treated as input. The heat transfer rate in the distillation column and heat from exhaust gases
is considered as useful output. The graph in Fig. 4 shows that as the heat transfer rate increases, the
efficiency increases. For the baseline conditions, energy efficiency is 72% while exergy efficiency is 65%.
The trend shows the linear increase in the efficiencies with energy efficiency increase faster than exergy
efficiency. Case 3 has highest efficiencies as compared to all the other cases. The efficiencies are higher
because the useful output is not just the heat transfer rate but it also includes the exhaust gas heat.

0,8

0,7

0,6
Exergy efficiency

Case 1 - Exergy
0,5
Case 2 - Exergy
0,4
Case 3 - Exergy
0,3
Case 1 - Energy
0,2 Case 2 - Energy
0,1 Case 3 - Energy
0
290 300 310 320 330 340
Ambient temperature, T0 [K]

Fig. 5: The changes of energy and exergy efficiencies with ambient temperature.

Fig.5 shows the energy and exergy efficiency for all three cases when ambient temperature
changes. The ambient temperature changes with respect to season, time of day and location of the

222
distillation plant. Ambient temperature does not have a direct effect on the energy efficiency hence the
energy efficiencies remain constant at 53%, 40% and 72% for case 1, case 2 and case 3 respectively.
Ambient temperature directly affects the exergy efficiency because exergetic heat is related to ambient
temperature as exergy efficiency is defined. The baseline exergy efficiencies are 25.3%, 23.3% and 65% for
case 1, case 2 and case 3 respectively. When compared with energy efficiencies, case 1 has the highest
difference whereas case 3 has the lowest. As the ambient temperature increases, the exergy efficiency
decreases linearly. This decrease in exergy efficiency is understandable since higher ambient temperature
means lower amount of useful heat can be extracted from the output. The ambient temperature affects the
case 1 efficiency the most as it has the highest decline for the studied temperature. Case 1 gets affected the
most since its output solely relies on exergetic consumed heat in the system.

Exergy destruction

Heater1
VDU
16%
26%

Heater2
2%

ADU
56%

Fig. 6: Exergy destruction of the crude oil distillation system.

Fig. 6 shows the exergy destruction of each component of the distillation system. Atmospheric distillation unit
(ADU) has the highest exergy destruction, followed by vacuum distillation unit (VDU), then heater 1 and heater
2. Heater 1 is the heater that heats the feed before it enters the ADU while heater 2 heats the feed before it
enters the VDU. ADU has higher exergy losses since it has greater number of trays compared to VDU. Since
the oil feed interacts with more number of trays and deals with heavier compounds, the exergy destruction is
greater. Heater 1 has higher exergy destruction since it has higher mass flow rate to heat and secondly the
feed to be heated is at a lower temperature as compared to what heater 2 secures.

CONCLUSIONS
The paper introduces three different exergy and energy efficiencies and discusses them for various inputs
and outputs. For all the different cases of inputs and useful outputs, a parametric study is conducted to show
the change in efficiencies when consumed heat and ambient temperature changes. Case 3 has the highest
baseline energy and exergy efficiencies at 72% and 65%, respectively. Case 1 efficiencies are the second
highest at 53 % for energy while exergy efficiency is at 25.3%. Case 2 has the lowest baseline efficiencies at
40% for energy and 23.5% for exergy efficiency. Case 3 has the highest baseline efficiencies since this case
treats the heat transfer rate as well as the exhaust gases as useful output. The heater heat is considered as
the only input. Case 2 treats the heater heat as the input while heat transfer rate, in the distillation columns,
as the output. This case has low magnitude output as the exhaust gases are not considered useful hence it
has the lowest efficiency. Case 1 also treats the heat transfer rate, in the distillation columns, as the only
output but it has a better efficiency compared to case 2 because it takes the absorbed heat by crude as
input, not the total provided heat by the heaters. The unabsorbed heat from the heater is assumed to be
utilized as a useful output. Overall, this paper concludes that by using the unabsorbed heat from the heaters,
the overall plant efficiency can be improved. The paper also suggests that by utilizing the exhaust gases as
useful input to some other system, the efficiencies can be further improved for the overall plant. The
improvement in efficiency helps reduce the energy wastage, in turn, reduces the operational cost of the
plant. The increase in efficiencies means reduction in dumped heat to the surrounding. This translates to
lowered adverse effects of heat dumping to the environment and is a step towards sustainability.

223
NOMENCLATURE
ADU Atmospheric distillation unit
VDU Vacuum distillation unit
Energy rate, W
Exergy rate, W
Exergy destruction rate
PA Pump-around circuit
Heat rate consumed in the distillation system, W
T Temperature, K
T0 Ambient temperature, K
Greek Letters
Ψ Exergy efficiency
η Energy efficiency
Subscripts
1-17 States
C Case for efficiencies
H Heater

REFERENCES
Al-Muslim H, I. Dincer, S.M. Zubair. 2003. Exergy analysis of single and two-stage crude oil distillation units,
ASME Journal Energy Recourse Technology 12 : 199–207

Al-Muslim H, I. Dincer, S.M. Zubair. 2005. Effect of reference state on exergy efficiencies of one- and two-
stage crude oil distillation plants. International Journal of Thermal Sciences 44: 65–73.

Bagajewicz M. and S. Ji. 2001. Rigorous procedure for the design of conventional atmospheric crude
fractionation units. Part I: Targeting. Industrial & Engineering Chemistry Research 40: 617–626

Cengel Y A, M. A. Boles. 2006. Thermodynamics: An Engineering Approach, 5th ed. McGraw-Hill: New York.

Dincer, I and M. Rosen. 2007. EXERGY Energy, Environment and Sustainable Development. London:
Elsevier.

Errico M., G. Tola and M. Mascia. 2009. Energy saving in a crude distillation unit by a preflash
implementation. Applied Thermal Engineering 29: 1642-1647

Farooque A.M. et al. 2008. Parametric analyses of energy consumption and losses in SWCC SWRO plants
utilizing energy recovery devices. Desalination 219: 137-159

Kanoglu M, I. Dincer and M. A. Rosen. 2007. Understanding energy and exergy efficiencies for improved
energy management in power plants. Energy Policy 35: 3967-3978

Rosen M.A. and I. Dincer. 1997. On exergy and environmental impact. International Journal of Energy
Research 21: 643–654

Talbi M. and B. Agnew. 2002. Energy recovery from diesel engine exhaust gases for performance
enhancement and air conditioning. Applied Thermal Engineering 22: 693-702

224
 THERMODYNAMIC ANALYSIS OF A SOLAR DRIVEN TRI-GENERATION SYSTEM FOR
BUILDING APPLICATIONS

H. Ozcan* and I. Dincer

Faculty of Engineering and Applied Science, University of Ontario Institute of Technology
2000 Simcoe St. N. L1H 7K4 Oshawa ON, Canada
Hasan.Ozcan@uoit.ca*, Ibrahim.Dincer@uoit.ca

ABSTRACT
Energy and exergy analyses and sustainability assessment of a conceptual solar driven tri-generation system with
thermal energy storage option for power, space and water heating and air cooling are performed. The present tri-
generation system includes parabolic trough solar collectors (PTSC), an organic Rankine cycle for power generation
(ORC), an absorption chiller for cooling (AC) and a thermal energy storage system (TES). The effects of solar collector
dimension variations, system parameters, environmental conditions and system integration on system energy and
exergy efficiencies are parametrically studied. The largest irreversibility occurs in solar collectors due to very high
exergy input to solar panels and less conversion of exergy to working fluid. Energy and exergy efficiencies of the
present system become 77.1% and 27% during day time and 37.6% and 18.7% during night time, respectively. The
proposed system performs better performance at lower ambient temperature and higher solar radiation and PTC
concentration ratio.

Keywords: Solar energy, tri-generation, storage, energy, exergy, efficiency.

1. INTRODUCTION
Environmental problems associated with energy use span a growing spectrum of pollutants, hazards, and accidents
and degradation of environmental quality and natural ecosystems. Over the past few decades, the increasing use of
energy has expanded our concerns from what were once primarily local or regional issues, to a growing awareness of
the international and global nature of major energy-related environmental problems (Dincer, 1998). Economic and
environmental problems related to unsustainable energy systems have been a growing concern particularly after Kyoto
Protocol in 1997 and Stern Review in 2006. Hence, relationship between the energy demands and environmental
impacts should be carefully analyzed (Ozbilen et al., 2012). It is reported by International Energy Agency (IEA, 2012)
that global energy supply by fossil fuels is 81.3% by 2010.

Renewable energy technologies have become increasingly important as environmental concerns increase.
Development of advanced renewable energy technologies can serve as cost-effective and environmentally responsible
alternatives to conventional energy generation (Dincer, 2000). There are many alternative energy sources which can
be used instead of fossil fuels. The decision as to what type of energy source should be utilised must, in each case, be
made on the basis of economic, environmental and safety considerations. Because of the desirable environmental and
safety aspects it is widely believed that solar energy should be utilised instead of other alternative energy forms, even
when the costs involved are slightly higher (Kalagirou, 2004).

In order to deliver high temperatures with good efficiency, high-performance solar collectors are required. Systems
with light structures and low cost technology for process heat applications up to 400ºC could be obtained with
parabolic through collectors (PTCs) (Kalagirou, 1997, 2004). Tyagi et al. (2007) performed a comprehensive exergy
analysis and parametric study on PTCs and investigated effect of PTC design and environmental conditions and
specified optimal value for concentration ratio of PTCs. Other parametric studies to optimize PTCs can be found in
(Manikandan et al., 2012) and (Huang et al., 2012). Garcia et al. (2011) investigated a new performance model for
PTC based power plants with thermal energy storage. Giostri et al. (2012) compared various solar plants driven by
PTCs.

225
Tri-generation usually refers to the simultaneous production of cooling, heating and power from a single energy
source. It is known as combined cooling, heating and power (CCHP). Sometimes combined heating and power (CHP)
refers to tri-generation. In other words, if the heat produced from CHP is used for cooling as well as heating, the plant
is called a tri-generation plant. CHP could refer to a cogeneration plant if it produced heat and power only (Al-Suleiman
et al., 2011). Zhai et al. (2009) proposed a novel tri-generation system to be used in remote areas producing power,
heating and cooling and carried out a comprehensive thermodynamic and economic analysis to validate applicability of
such system. Marques et al. (2010) presented a generic tri-generation scheme for office building and carried out first
law analysis. Al-Suleiman et al. (2011, 2012) performed a thermodynamic assessment of solar driven tri-generation
system with thermal energy storage and specified best system efficiencies considering system and environmental
parameters.

The main objective of this study is to perform a comprehensive thermodynamic analysis of a solar driven tri-generation
plant for power, cooling and heating. Energy and exergy analysis, irreversibilities of each sub-unit and components are
evaluated, best system efficiencies are determined by parametric studies. PTC, ORC and AC parameters and
environmental conditions are considered to optimize the proposed plant. A comparative study is performed for solar
mode and TES mode. Detailed explanation and discussion of thermodynamic analysis are presented in next sections.

2. SYSTEM DESCRIPTION
Proposed tri-generation plant is represented in Fig. 1. Therminol-66 which is suitable to work up to 345 ºC at liquid
phase is used as the working fluid for the solar cycle (Eastman, 2013). Working fluid is heated up to 250 ºC through
the PTCs and release its heat to TES heat exchanger, ORC evaporator and AC heat exchanger. To provide a steady
state working condition for the plant, heat is stored at TES to be utilized when the sun is out. Heat is initially stored in
hot storage tank, and after utilization of stored heat, it is stored in the cold storage tank. However, the performance
assessment is performed for solar system only to evaluate TES integration.
13
28
Pump 3 30
Condenser Generator ORC
21

ORC Evaporator
10 Turbine
11
AC-HEX

Cooling 18 17 POWER
Tower Cold Hot
22 12 Storage Storage 9 31
29 Regenerator
27 tank tank

14 4 8
19 16 2
TES-HEX 33
Expansion
Valves Pump 7 6
23 15 ORC
20 Pump
32 Condenser
Evaporator Absorber 5 1
24

Pump 27 26 35 34
Residential
36 Water Water
26 25
Tank
COOLING
SOLAR FIELD HEATING

Fig. 1. Schematic of the tri-generation plant.

Transport fluid pumped at 28-30 bar at ORC pump and gains heat at ORC evaporator from solar cycle. Evaporated
and superheated fluid expands at ORC turbine, and the power production is achieved. After condensing at ORC
condensing the ORC loop is completed (Liu et al., 2012). Here, we consider residential water at 20ºC as heat sink for
the ORC plant. Heat rejected from ORC condenser heats residential water up to 50ºC and water is stored in a water
tank for later use. R245fa is selected as the transport fluid for the ORC plant. R245fa, has a satisfactory thermal
efficiency and production of mechanical work and most important of all is ozone safe as well as it follows all the

226
environmental regulations and commercially produced (Kosmadakis et al., 2009). However, R123 is also considered
as an optional working fluid for a comparative assessment. Some thermophysical, thermodynamic and environmental
properties of R123 and R245fa is given in Table 1. Considering environmental indicators given in Table 1, R245fa is
considered as the working fluid for the baseline model, and R123 is used for comparison only.

Table 1. Some properties of R123 and R245fa (Modified from Kosmadakis et al., 2009)
Critical Critical Molecular Ashrae 34
Working GWP
Temperature Pressure weight ODP* Safety
Fluid (for 500y)**
(ºC) (Bar) (kg/mol) Group***
R123 183.7 36.6 152.9 0.02 36 B1
R245fa 157.6 36.4 134.05 0 300 A1
* ODP: Ozone Depletion Potential, expresses the contribution to ozone depletion, based upon R11=1.
** GWP: Global Warming potential, potential to warm to planet related to CO2 emission during production.
*** ASHRAE 34: (A) means low toxicity, (B) means high toxicity. (1) means no flammability propagation.

The weak Li-Br water solution is first heated up to desired temperature with the heat provided by water cycle and
higher temperature strong solution Li-Br water. The strong solution rejects its heat to weak solution in regenerator,
expands to low pressure side and turns back to absorber. Some of the water content evaporates in the generator and
condenses in the AC condenser. The heat rejected in the condenser and absorber is transmitted to cooling tower
water cycle (Sencan et al., 2005; Gebreslassie et al., 2010). The condensed water is than expanded into a lower
pressure with expansion valve and its temperature decreases up to 4°C. The performance investigation of the
absorption chiller model is performed with mass, energy, exergy and entropy balances and defined in the next
subsection.

3. SYSTEM ANALYSIS
In order thermodynamically model the present tri-generation system, mass, energy and exergy balance equations for
analysis and performance evaluation are presented in this section. The following assumptions are made for system
analysis and assessment:
 Steady-state steady-flow conditions are considered for overall system except for TES and water tank.
 Pressure drops through piping are negligible.
 System components work adiabatically.
 Changes in kinetic and potential energies and the pressure drops in the piping and components of the system
are negligible.
 Input parameters for the overall system and range of variations are given in detail in Table 2.

3.1. Energy and Exergy Analyses

The performance evaluation of all subunits and the overall system is performed using mass, energy, exergy and
entropy balances. The generalized balance equations for the system are defined with the following definitions (Dincer
and Rosen, 2007):
Mass : (1)

Energy : (2)

Exergy : (3)

Entropy : (4)

Here, indices i, o, des, gen, sys and r represents inlet, outlet, destruction, generation, system and control region.
represents specific exergy and s represents entropy. and are destructed exergy rate and exergy due to heat

227
transfer. As system is assumed as steady-state and steady-flow, mass, energy, entropy and exergy of control volume
of the components are equal to zero.

The specific exergy of a stream is comprised of two components, namely physical and chemical exergy, and defined
as follows:

(5)

(6)

(7)

(8)

(9)

(10)

where indices 0, ch and ph represent ambient conditions, chemical and physical, respectively. For further details on
system analysis through energy and exergy can be found elsewhere (Dincer and Rosen, 2007; Kotas, 1995).

Table 2. System operating conditions and parameters.

Parameter Unit Value
Dead state temperature (T0) °C 20-30
Dead state pressure (P0) bar 1
Heat Exchanger efficiencies (ηhex) % 85
PTSC
-2
Global Solar Radiation (St) kW m 0.4-0.8
2
Heat loss coefficient (UL) W/m K 8
Concentration ratio - 30-50
Absorptivity, Reflectivity and emissivity - 0.9
Receiver effectiveness ( ) % 90
Collector efficiency % 90
Sun Temperature K 5700
2 4 -8
Stephan Boltzmann Constant W/m K 5.67x10
ORC
Pressure ratio (PR) - 4.2
Working fluid - R245fa
Condensing temperature °C 68
Turbine isentropic efficiency (ηs) % 90
Turbine mechanical efficiency(ηm) % 90
Turbine electrical efficiency(ηel) % 95
AC
Chilled air inlet/outlet temp. °C 18/7
Cooling water inlet/outlet temp. °C 32/35
Condenser temp. °C 40
Evaporator temp. °C 7

3.2. Thermal Energy Storage

Thermal energy storage (TES) generally involves the temporary storage of high- or low-temperature thermal energy for
later use. TES systems achieve benefits by fulfilling one or more of the following purposes: Increasing generation
capacity, enable better operation of cogeneration plants, shift energy purchases to low-cost periods, increase system

228
reliability and integration with other functions (Dincer and Rosen, 2007). Energy balance for overall storage process
considering Fig. 1 is given as follows:

(11)

(12)

(13)

where represent total heat loss during charging, storage and discharging process, j=1,2 and 3 indicate charging,
storage and discharging processes. Energy balance for overall storage process considering Fig. 1 is given as follows:

(14)

(15)

(16)

(17)

where is exergy loss associated with heat transfer for all three processes and donate irreversibilities during
charging, storage and discharging processes.

3.3. Parabolic Trough Collectors

Actual useful heat gain from PTCs is defined with Hottel-Whillier equation considering radiation loses as follows
(Kalagirou, 2004; Tyagi et al., 2007):

(18)

Here, is collector heat removal factor, is absorber area, C is concentration ratio, , are reflectivity,
absorptivity and emissivity and Stephan Boltzmann constant, is overall heat loss coefficient, is global solar
radiation and Tcol is collector temperature. The collector temperature can be determined using numerical solutions
proposed in (Tyagi et al., 2007). Heat removal factor of the collectors and concentration ratio can be defined through
eqs. (19-20):

(19)

(20)

where F, and are collector efficiency, working fluid mass flow rate and specific heat and reflector area,
respectively. L is collector length; w is collector width and is collector cover diameter. Actual heat gain can also be
determined as follows:

(21)

For desired outlet temperature, solar parameters can be specified using Eqs. (18) and (21). Exergy gain from PTCs
and collector exergy associated with sun temperature is given as in (Petela, 2005) as

229
 (22)

(23)

3.4. System Performance

Considering efficiencies of all subunits and overall system, a general definition of rate of useful energy/exergy to inlet
energy/exergy can be used for determination of system performances. Energy and Exergy efficiencies for the subunits
shown in Fig. 1 are written as

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

where is net power produced is ORC plant, and are heat loads in AC evaporator and
generator, respectively. Subscripts en and ex defined energy and exergy. Finally, overall energy and exergy
efficiencies of the plant can be defined as follows:

(32)

(33)

4. RESULTS AND DISCUSSION

The results of the thermodynamic analysis of the proposed tri-generation plant are presented and discussed in this
section. EES (Engineering Equation Solver) software is used to study the performance of the system and its
components.

Energy and exergy efficiencies, as well as irreversibility rates of plant sub-units and overall plant are represented in
Fig. 2(a). The energy and exergy efficiencies of ORC are 12.27% and 52%, respectively. The energetic and exergetic
COPs of AC are 77.07% and 30.06%. The energy and exergy efficiencies for water heating are 86% and 24%. Also,

230
the energy and exergy efficiencies for TES are 60.1% and 35.2% and PTC are 6.7% and 29.4. The PTC energy and
exergy efficiencies are in agreement with works in the literature. Low energy efficiency is due to very high reflector
area need to provide tri-generation and thermal energy storage for continuous production. 11812 m2 of reflector area is
needed to provide tri-generation and TES with a concentration ratio of 40. When no TES is considered, the reflector
area needed is 3726 m2. When determining overall efficiencies, night time plant performance is also considered by
using TES energy as plant input. Overall energy and exergy efficiencies of solar mode and TES mode are 61% -
21.2% and 37.6 - 18.7%, respectively. The energy storage option through a TES system is considered as useful output
for the solar mode.

The largest irreversibility occurs in PTC due to very high input exergy associated with sun temperature as defined in
Eq. 23. Irreversibility rates of ORC, AC, water heating, TES, PTC, Solar mode and TES mode are 119, 78, 79, 639,
3420, 4335 and 796 kW, respectively. Energy stored during the day time in TES is considered as the energy input for
the tri-generation plant at night time. Thus irreversibility of TES mode is considerably lower than that of Solar Mode as
exergy of PTC associated with sun temperature is not used for TES mode.

Exergy;
90 Energy Efficiency
5000
13%
80 Exergy Efficiency Energy;
Irreversibility Rate
4000 9%

Irreversibility Rate (kW)

70
60
Efficiency (%)

3000 Exergy;
50 Power
18%
40 Cooling
2000
30 Energy;
Energy; Heating 40%
20 51%
1000
10
Exergy;
0 0
69%
ORC AC Heating TES PTC Overall Overall
(Solar (TES
mode) mode)

(a) (b)
Fig. 2. (a) Energy, exergy efficiencies and irreversibility rates of the plant sub-units and overall plant. (b) Energy and
exergy output ratios.

Energy and exergy output ratios are represented in Fig 2(b). Tri-generation plant produces 125.4 kW power, 557 kW
cooling load and 709 kW of residential water heating. As exergy of power is same as its energy, it remains the same in
exergy basis; however cooling and heating exergy corresponds to 33.6 and 24.3 kW. Considering energy output ratios,
power production generates only 9% of total production, and it increases up to 69% by considering exergy outputs.

A comparative energy, exergy and exergy destruction variations considering ORC parameters and ambient
temperature are represented in Figs. 3(a), 3(b). and 3(c). A pressure ratio increment slightly increases exergetic and
energetic efficiencies and has an influential effect on specific work of turbine. A 20ºC change in assumed pump inlet
temperature decreases exergy efficiency by 25% and brings a 6 kJ/kg increase in irreversibility, whereas no
considerable change occurs in energy efficiency. Ambient temperature increase also slightly increases exergy
efficiency of ORC plant and decrease total irreversibility. ORC plant working with R123 shows slightly better energy
and exergy performance and lower irreversibilities than that of R245fa. However, R245fa is the most promising
working fluid considering environmental and safety challenges. R245fa is selected as the transport fluid for the ORC
plant.

231
(a)

(b)

(c)
Fig 3. ORC pump pressure ratio (a), ORC pump inlet temperature (b) and ambient temperature (c) effect on
efficiencies, irreversibilities and specific work output of ORC plant.

232
Absorption chiller condensing temperature effect on energetic and exergetic COP for various evaporator temperatures
are illustrated in Fig 4. Since the solution fractions of LiBr-Water are strongly dependent on saturation pressures for
condensing and evaporating temperatures of AC, a more careful investigation should be performed to optimize the AC
performance. Energetic COP is higher at higher evaporating temperatures and exergetic COP is higher at lower
evaporating temperatures. Both energy and exergy COPs decrease at higher condensing temperatures.

Fig. 4. Condensing and evaporation temperature effect on COP performance of absorption chiller.

Global solar radiation is a strong influential environmental parameter on solar plants. Solar radiation effect on collector
and overall efficiencies are represented in Fig. 5. Solar global radiation changes in a sunny day from 0.4 to 0.8 kW/m 2.
In this study, instead of considering the solar data from a specific region, we parametrically illustrated the effect of
solar radiation on system performance. PTC energy efficiency increases up to 10%, even if solar global radiation is
used as a denominator for PTC energy performance.

Fig. 5. Solar radiation effect on PTC and overall energy and exergy efficiencies.

Ambient temperature effect on exergy efficiencies and irreversibility rates of generated energy products are illustrated
in Fig. 6. Exergy efficiencies of ORC and AC plant increases with ambient temperature increase whereas water
heating exergy efficiency decrease by almost 40%. Irreversibility rates of ORC and AC decrease proportional to their
exergy efficiencies with increasing ambient temperatures. Irreversibility rate of water heating increase at higher
ambient temperatures due to higher condensing energy input to residential water as mentioned in Fig. 1. Considering
exergy output ratios of energy forms, power ratio will increase among other forms of energy at higher ambient

233
temperatures. It can be concluded from the figure that, proposed system would show better performance with power
and cooling production especially in summer seasons.

Fig. 6. Ambient temperature effect on heating, cooling and power production.

Finally, exergy efficiency and sustainability index comparison of overall system considering solar and TES mode at
varying ambient temperature and global solar radiation conditions are illustrated in Fig. 7. As known, sustainable
development requires not just that sustainable energy resources be used, but that the resources to be used efficiently.
Exergy methods are essential in improving efficiency, which allows society to maximize the benefits it derives from its
resources while minimizing the negative impacts. By noting that energy can never be ‘‘lost’’ as it is conserved
according to the first law of thermodynamics, while exergy can be lost due to internal irreversibilities, the study
suggests that exergy losses, particularly due to the use of non-renewable energy forms, should be minimized to obtain
sustainable development. Thus, sustainability should be considered together with exergy analysis. Overall system
shows better sustainability and efficiency at higher solar radiation and lower environmental temperature. However,
these conditions cannot be optimized and this illustration is to show the impact of environmental conditions on system
performances.

Fig. 6. Effect of environmental parameters on plant exergy efficiency and sustainability.

Thermodynamic analysis results show that plant performance is strongly dependent on environmental conditions and
collector design conditions as well as plant parameters such as mass flow rate, pressure ratios and specified
temperature ranges.

234
5. CONCLUSIONS
Energy and exergy analyses and sustainability assessment of a conceptual solar driven multi-generation system for
power, space and water heating, air cooling and thermal energy storage are performed. The following findings are
obtained from the present study.
 Highest irreversibility occurs at solar collectors due to very high exergy input considering solar temperature.
 AC energy efficiency and ORC exergy efficiency appear to be highest among other system sub-units.
 Exergy output ratios of cooling and heating are considerably lower than that of energy output ratios due to low
exergy quality of corresponding energy forms.
 Energy and exergy performance of ORC plant using R123 as working fluid is slightly higher than that of
R245fa. However R245fa is selected for the baseline model due to toxicity and ozone depleting potential of
R123.
 It is of importance to select proper condensing and evaporation temperatures for absorption chiller, as these
values are strongly influencing fraction of Li-Br in water and effecting system performance.
 Environmental parameters such as ambient temperature and solar global radiation strongly influence system
performances and exergetic sustainability.
 Exergy efficiency of TES mode at night time is better than that of solar mode at day time.
 Energy stored at TES during day time is considered as useful energy output, thus energy efficiency of overall
system is comparatively higher than that of TES mode.
 TES integration brings around 8000 m2 more solar panel installation to provide a steady-state tri-generation,
thus a thermo-economic optimization may be performed for the proposed system to find out flow and
investment costs. However, TES option integration is inevitable to provide steady-state operation of the plant.
 Higher collector concentration ratio, ORC pump pressure ratio and pump inlet temperature and lower,
absorption chiller evaporating and condensing temperatures are recommended to provide better energetic
and exergetic performance and lower plant irreversibility.

NOMENCLATURE
exergy rate, kJ/s
Irreversibility rate, kJ/s
heat transfer rate, kJ/s
work rate, kJ/s
mass flow rate, kg/s
Aa absorber area, m2
Ar reflector area, m2
C concentration ratio
ex specific exergy, kJ/kg
Fr collector heat removal factor
SI sustainability index
St global solar radiation, kW/m2
Greek Letters
absoptivity
emissivity
efficiency
reflectivity
Boltzmann’s constant, 5.67x10-8, W/m2K4
collector width, m
Subscripts
ch chemical
col collector
des destruction

235
i inlet
o outlet
ph physical
Acronyms
AC absorption chiller
ORC organic rankine cycle
PTC parabolic trough collectors
TES thermal energy storage

REFERENCES
Al-Suleiman, F., Hamdullahpur, F., Dincer, I. 2011. Exergy modeling of a new solar driven trigegeneration system. Solar
Energy 85: 2228-2243.
Al-Suleiman, F., Hamdullahpur, F., Dincer, I. 2011. Trigeneration :A comprehensive review based on prime movers.
International Journal of Energy Research 35: 233-258.
Al-Suleiman, F., Hamdullahpur, F., Dincer, I. 2012. Performance assessment of a novel system using parabolic trough solar
collectors for combined cooling, heating, and power production. Renewable Energy 48: 161-172.
Dincer, I. 1998. Energy and environmental impacts: present and future perspectives. Energy Sources 20: 427-453.
Dincer, I. 2000. Renewable energy and sustainable development: a crucial review. Renewable and Sustainable Energy
Reviews 4: 157-175.
Dincer, I., Rosen, M.A. 2007. Exergy: Energy, Environment and Sustainable Development. Oxford: Elsevier.
Eastman Chemical Company. 2013. http://www.therminol.com/pages/products/66.asp
Garcia, I. L., Alvarez, J. L, Blanco, D. 2011. Performance model for parabolic trough solar thermal power plants with thermal
storage: Comparison to operation plant data. Solar Energy 85: 2443-24460.
Gebreslassie, B.H., Medrano, M., Boer, D. 2010. Exergy analysis of multi-effect water-LiBr systems: From half to triple effect.
Renewable Energy 35: 1773-1782.
Giostri, A., Binotti, M., Astolfi, M. Silva, P., Macchi E., Manzolini, G. 2012. Comparison of different solar plants based on
parabolic trough technology. Solar Energy 86: 1208-1221.
Huang, W., Hu, P., Zeshao, C. 2012. Performance simulation of a parabolic trough solar collector. Solar Energy 86: 746-755.
International Energy Agency (IEA) World energy Outlook. 2012. http://www.worldenergyoutlook.org/
Kalagirou, S. 2004. Solar thermal collectors and applications. Progress in energy and Combustion Science 30: 231-295.
Kalagirou, S., Lloyd, S., Ward, J. 1997. Modelling, optimization and performance evaluation of a parabolic trough solar
collector steam generation system. Solar Energy 60: 49-59.
Kosmadakis, G., Manolakos, D., Kyritsis, S., Papadakis, G. 2009. Comparative thermodynamic study of refrigerants to select
the best use in the high-temperature stage of a two-stage organic Rankine cycle for RO desalination. Desalination 243: 74-
94.
Kotas, T.J. 1995. The exergy method of thermal plant analysis, London-Boston: Krieger Publishing Company.
Liu, B., Riviere, P., Coquelet, C., Gicquel, R., David, F. 2012. Investigation of a two stage rankine cycle for electric power
plants. Applied Energy 100: 285-294.
Manikandan, K.S., Kumaresan, G., Velraj, J., Inıyan, S. 2012. Parametric study of solar parabolic trough collector system.
Asian Journal of Applied Sciences 5(6): 384-393.
Marques, R.P., Hacon, D., Tessarollo, A., Parise, J.A.S. 2010. Thermodynamic analysis of tri-generation systems taking into
account refrigeration, heating and electricity load demands. Energy and Buildings 42: 2323-2330.
Ozbilen, A., Dincer, I., Naterer, G.F., Aydin, M. 2012. Role of hydrogen storage in renewable energy management for
Ontario. International Journal of Hydrgen Energy 37: 7343-7354.
Petela, R. 2005. Exergy analysis of the solar cylindrical-parabolic cooker. Solar Energy 79 (3), 221–233.
Sencan, A., Yakut, K.A., Kalagirou, S. 2005. Exergy analysis of lithium bromide/water absorption systems. Renewable
Energy 30: 645-657.
Tyagi, S.K., Wang, S., Singhal, M.K., Kaushik, S.C., Park, S.R. 2007. Exergy analysis and parametric study of concentrating
type solar collectors. International Journal of Thermal Sciences 46: 1304-1310.
Zhai, H., Dai, Y.J., Wu, J.Y., Wang, R.Z. 2009. Energy and exergy analyses on a novel hybrid solar heating, cooling and
power generation system for remote areas. Applied Energy 86: 1395-1404.

236
 THERMODYNAMIC ANALYSIS OF AN AUTOMOTIVE AIR CONDITIONING SYSTEM USING
R1234YF AND R134a
* ** **
A. Alkan , A. Kolip , S.Cosman
*
Department of Mechanical Education, Faculty of Technical Education, Sakarya University 54187, Esentepe, Sakarya
**
Department of Mechanical Engineering, Faculty of Technology, Sakarya University 54187, Esentepe, Sakarya
aalkan@sakarya.edu.tr, akolip@sakarya.edu.tr, sertaccosman@sakarya.edu.tr

ABSTRACT
In this study, thermodynamic simulation of an automotive air conditioning system using R1234yf and R134a
refrigerants was conducted to determine and compare the performance of these two refrigerants. For this aim,
thermodynamic properties of R1234yf and R134a were modelled as a function of pressure and temperature. Then,
a program was developed using MATLAB to obtain performance parameters of the automotive air conditioning. By
means of this program, the compressor capacity, discharge temperature, cooling capacity, COP and total exergy
destruction of the system were determined as functions of evaporating and condensing temperatures. For both
refrigerants, the results were presented by comparative diagrams. According to the thermodynamic analysis
results, it is seen that the compressor capacity, discharge temperature and total exergy destruction of the system
decrease while cooling capacity and COP increase with increasing evaporating temperature. On the other hand,
the compressor capacity, discharge temperature and total exergy destruction increase while cooling capacity and
COP decrease with increasing condensing temperature. R1234yf refrigerant causes less discharge temperature
(in the range of 9-15 oC), less compressor capacity (0.59-0.35kW) and less total exergy destruction (0.57-0.34 kW)
than R134a refrigerant does. However, as a result of thermodynamic analysis it has been determined that R1234yf
refrigerant has less COP (2.6-5%) and less cooling capacity (21-25%) than R134a refrigerant.

INTRODUCTION
Industrial and technological development has brought various environmental problems. Especially, it has been
understood that refrigerants used in air conditioning system cause ozone layer and global warming by scientific
researches. In 1987; because of damaging to the ozone layer, it was limited to use of refrigerants that containing
chlorine atoms. Then they were banned gradually. With the Kyoto Protocol in 1997; it has been decided to restrict
using of hydrofluorocarbon group refrigerants and banned gradually which have high global warming potential.In
addition; it is banned with the global warming values (GWP) of refrigerants used in vehicles which will be produced
in 2017 and later, GWP of in all available vehicles more than 150 by European Parliament. However, R1234yf
refrigerant has very low GWP and nonintoxicant.

The open literature on the theoretical and experimental performance of automotive air conditioning system (AAC)
is very restricted. Jung et al. (1999) presented experimental performance of supplementary/retrofit refrigerant
mixtures for R12 used in existing AAC systems. Al-Rabghi and Niyaz (2000) compared the coefficients of
performance (COP) of an AAC system, originally designed to use R12 and R134a. Hosoz (2005) performed
exergy analysis of an experimental AAC system and determined the rates of exergy destruction in the components
of the system. Hosoz et al. (2006), investigated the experimental performance of an AAC system capable of
operating as a heat pump for the cooling and heating modes. Esen et al. (2006), determined comparative
experimental performance of an AAC system for R12 and R134a. Alkan et al. (2010), investigated effect to the
system performance with using fixed and variable capacity compressor on experimental AAC system that consist
of original components and working with R134a refrigerant.

Some releated researches are available for computer simulation of AAC system. Lee et al. (2000) developed some
mathematical models of refrigeration circuit elements in an AAC system with using a computer program, and made
performance analysis of this AAC system. Jabardo et al. (2002), developed a computer simulation model for the
steady-state performance of an AAC system with a variable capacity compressor, and investigated the accuracy of
the model on an experimental system. Joudi et al. (2003), advanced a simulation model of an ideal AAC system
for R12 and various alternative refrigerants. Esen et al. (2005) presented performance analysis results of an AAC
system that using R12 and R134a refrigerants by means of computer program. Hosoz et al. (2006), modelled the
performance parameters of an AAC system using artificial neural networks and observed that their model
predicted the performance of the system accurately. Zilio et al. (2011), tested a "drop-in" R1234yf system together
with two modified R1234yf systems with the primary goal to document some laboratory results and their analyses.
Zhao et al. (2012), developed a simulation model of the mini-channel evaporator using R1234yf as a working fluid.
Col et al. (2012), reported that the local heat transfer coefficient measured during flow boiling of R1234yf in a 1mm
diameter circular microchannel and compared to R134a.

Although numerous studies have done on the use of various refrigerants in AAC systems, comparative
thermodynamic simulation of R134a and R1234yf showing the effect of these refrigerants on the system
performance have not been reported in the literature yet. In this study, thermodynamic simulation of an automotive

237
air conditioning system using R1234yf and R134a refrigerants was conducted to determine and compare the
performance of these two refrigerants. Considered from the viewpoint of thermodynamic properties of a new
generation refrigerant, it will be a guide for the future projects and experimental works.

MODELLING OF AAC SYSTEM

In this modelling study, thermodynamic properties for comparing R1234yf and R134a refrigerants were obtained
by using Refprop (2010) program. Then, performance results of refrigeration circuit are obtained as a function of
the parameters consisting of temperature or the temperature-pressure points. According to the function conditions
of temperature or temperature-pressure parameters, two or three dimensional curves of these functions were
obtained in Matlab. Then, tenth degree of polynomial equations with one uknown and fifth degree of polynomial
equations with two unknowns were created for these curves in Matlab. A program was developed and
thermodynamic properties and thermodynamic performance values of refrigeration circuit were obtained using this
program. Finally, the results of the program are shown in comparative diagrams. Also a flow diagram for
thermodynamic analysis and schematic view of refrigerating circuit are shown in Fig. 1 and Fig. 2 respectively.

Fig. 1. A flow diagram of developed program for thermodynamic analysis of automobile air conditioning system.

Fig. 2. Shematic view of automobile air conditioning system refrigerating circuit.

238
Modelling of Evaporator and Condenser saturation pressures of R1234yf and R134a refrigerants
Using Refprop Program; corresponding to the evaporator and condenser saturation temperatures (-25/+30 C and
+25/+90 C respectively) to saturation pressures of refrigerants were obtained. Depending on the saturation
temperature, saturation pressure curves were created using datas in Matlab. Polynomial constants were obtained
by fitting polynomial curves to the created curves. Acquired constant values for both refrigerants are given below.
The tenth degree equation for the saturation pressure at the evaporator for R1234yf and its constants are given
below:

(1)

G1= -7.6782e-007; G2= 1.3896e-006; G3= 1.527e-005; G4= 4.1862e-005; G5= 0.00053705; G6= 0.00021531;
G7= 0.080325; G8= 3.1537; G9= 36.325; G10= 181.25; G11= 343.47;

The tenth degree equation for the saturation pressure at the evaporator for R134a and its constants are given
below:

(2)
A1= 9.7476e-008; A2= 5.5216e-007; A3= 1.5895e-006; A4= 5.0607e-005; A5= 0.00050791; A6= -0.001861;
A7= 0.1081; A8= 3.7298; A9= 39.016; A10= 180.45; A11= 320.26;

The tenth degree equation for the saturation pressure at the condenser for R1234yf and its constants are given
below:

(3)

H1= 0.0035228; H2= 0.0070182; H3= -0.0092393; H4= -0.0037505; H5= 0.082019; H6= 0.1844; H7= 0.61168;
H8= 8.434; H9= 107.42; H10= 664.74; H11= 1551.4

The tenth degree equation for the saturation pressure at the condenser for R134a and its constants are given
below:

(4)

B1= 0.00060456; B2= 0.0015651; B3= 5.4166e-005; B4= 0.0034438; B5= 0.033085; B6=0.13337; B7= 0.65949;
B8= 9.7156; B9= 121.1; B10= 713.87; B11= 1584.5;

Modelling of the Thermodynamic Properties at the Refrigeration Circuit (third point) for R1234yf and R134a
refrigerants
In refrigerating circuit, condenser outlet is the same as expansion valve inlet and known as third (3) point. In this
section, refrigerant is in the compressed liquid state. Also the temperature is called in that point and it is given
Eq.(5). At third point, thermodynamic properties of refrigerant are as a function of temperature. Using Refprop
Program; corresponding to temperatures (+25/+90 oC) to thermodynamic property values of refrigerant ( , )
were obtained. With the help of these data by using Matlab; depending on the temperature, and curves
were created. Polynomial constants were obtained by fitting tenth degree of polynomial curves to the created
curves. Taken constant values from the model are given below for both refrigerants.

Temperature equation of R1234yf and R134a refrigerants at (3) point:

(5)

The tenth degree equation for the enthalpy ) at (3.) point of refrigerating circuit for R1234yf and its constants
are given below:

239
 (6)

K21= 0.0063011; K22= 0.010786; K23= -0.025745; K24= -0.033714; K25= 0.067614; K26= 0.091516;
K27= 0.054486; K28= 0.31753; K29= 1.618; K210= 29.806; K211= 130.46;

The tenth degree equation for the enthalpy ) at (3.) point of refrigerating circuit for R134a and its constants are
given below:

(7)

C21= 0.00061762; C22 = 0.0013804; C23 = -0.0012268; C24 = -0.0013278; C25 = 0.01116; C26= 0.027687;
C27= 0.068712; C28= 0.27009; C29= 1.435; C210= 30.152; C211= 135.32;

The tenth degree equation for the entropy ) at (3.) point of refrigerating circuit for R1234yf and its constants are
given below:

(8)

K31= 1.698e-005; K32= 2.9005e-005; K33= -6.9637e-005; K34= -9.0829e-005; K35= 0.00018292; K36= 0.0002424;
K37= 0.00012567; K38= 0.00083282; K39= 0.0019469; K310= 0.088037; K311= 0.45878;

The tenth degree equation for the entropy ) at (3.) point of refrigerating circuit for R134a and its constants are
given below:

(9)

C31= 1.642e-006; C32= 3.6285e-006; C33= -3.3908e-006; C34= -3.5813e-006; C35= 2.9446e-005;
C36= 7.1468e-005; C37= 0.00016864; C38= 0.00070538; C39= 0.0013755; C310= 0.089164; C311= 0.4773;

Modelling of Thermodynamic Properties at Refrigerating Circuit (first point) for R1234yf and R134a
refrigerants
In refrigerating circuit, evaporator outlet is the same as compressor inlet and known as first (1) point. At this point,
refrigerant is superheated vapour and at the same time it is a function of temperature and pressure. By
using Refprop; depending on different variation of and values, refrigerant thermodynamic property values
( , ) were obtained. Then, surface curves were achieved as a function of temperature and pressure by
using Matlab. Polynomial constants were obtained by fitting fifth degree of polynomial curves with two unknowns to
the present curves. Acquired constants are given below:

Temperature equation of R1234yf and R134a refrigerants at (1) point:

(10)

The fifth degree equation for the enthalpy ) at (1.) point of refrigerating circuit for R1234yf and its constants are
given below:

(11)

240
L200= 209.1; L210= 8.196; L201= -2.523; L220= 0.03795; L211= 0.3139; L202= -0.1112; L230= 0.002871;
L221= -0.03185; L212= 0.03374; L203= -0.008702; L240= -0.000209; L231= 0.002768; L222= -0.006045;
L213= 0.003781; L204= -0.0007606; L250= 6.218e-005; L241= -0.0003945; L232= 0.001014; L223= -0.0009944;
L214= 0.0004359;L205= -6.947e-005;

The fifth degree equation for the enthalpy ) at (1.) point of refrigerating circuit for R134a and its constants are
given below:

(12)

D200= 254.1; D210= 9.232; D201= -3.099; D220= -0.07913; D211= 0.6011; D202= -0.1982; D230= 0.03379;
D221= -0.1418; D212= 0.1239; D203= -0.02652; D240= -0.006477; D231= 0.03467; D222= -0.05128; D213= 0.02548;
D204= -0.003906; D250= 0.001148; D241= -0.006832; D232= 0.0136; D223= -0.01053; D214= 0.003326;
D205= -0.0003604;

The fifth degree equation for the entropy ) at (1.) point of refrigerating circuit for R1234yf and its constants are
given below:

(13)

L300= 0.7993; L310= 0.03061; L301= -0.03755; L320= -0.0003756; L311= 0.0009916; L302= 0.006378;
L330= 0.0002371; L321= -0.0008201; L312= 0.0008024; L303= -0.002087; L340= -0.0002195; L331= 0.00124;
L322= -0.002421; L313= 0.002127; L304= -0.0003256; L350= 5.019e-006; L341= -0.0001214; L332= 0.0003207;
L323= -0.0002766; L314= 1.974e-005; L305= 1.314e-005;

The fifth degree equation for the entropy ) at (1.) point of refrigerating circuit for R134a and its constants are
given below:

(14)

241
D300= 0.9382; D310= 0.03311; D301= -0.03698; D320= -0.0005901; D311= 0.001436; D302= 0.004623;
D330= 5.389e-005; D321= -0.0003281; D312= 0.0005817; D303= -0.001562; D340= -0.0003388; D331= 0.001707;
D322= -0.003288; D313= 0.002874; D304= -0.0005163; D350= 8.458e-005; D341= -0.0004204; D332= 0.0007168;
D323= -0.0004289; D314= -0.0001424; D305= 9.102e-005;

Modelling of Thermodynamic Properties at Refrigerating Circuit (second point) for R1234yf and R134a
refrigerants
In refrigerating circuit, compressor outlet is the same as condensor inlet and known as second (2) point. At this
point, the refrigerant is superheated vapour and temperature properties are as a function of temperature and
pressure. Acquired constants are given below:

Entropy equation of isentropic compressor for R1234yf and R134a;

, (15)

The fifth degree equation for the enthalpy ) at (2.) point of refrigerating circuit for R1234yf and its constants are
given below:

(16)

M100= 254.6; M110= 6.019; M101= 4.472; M120= -1.243; M111= 0.2084; M102= 0.02026; M130= 0.2785;
M121= -0.0191; M112= -0.002976; M103= 0.0003474; M140= -0.1115; M131= 0.005772; M122= -0.0001137;
M113= 0.0001323; M104= -1.183e-005; M150= 0.03445; M141= -0.001953; M132= 8.174e-005; M123= 2.705e-009;
M114= -6.392e-006; M105= 4.887e-007;

The fifth degree equation for the enthalpy ) at (2.) point of refrigerating circuit for R134a and its constants are
given below:

(17)

E100= 298.4; E110= 7.508; E101= 7.163; E120= -1.399; E111= 0.3382; E102= 0.05618; E130= 0.3104;
E121= -0.04032; E112= -0.004147; E103= 0.001073; E140= -0.1252; E131= 0.01191; E122= -1.997e-005;
E113= 0.0002006; E104= -4.253e-005; E150= 0.03853; E141= -0.003554; E132= 6.112e-005; E123= -3.63e-006;
E114= -7.661e-006; E105= 2.19e-006;

Depending on the isentropic efficiency of compressor, equation of the enthalpy value at (2) point of refrigerating
circuit for R1234yf and R134a refrigerants;

, (18)

242
The fifth degree equation for the temperature ) at (2.) point of refrigerating circuit for R1234yf and its constants
are given below:

(19)

M200= 70.09; M210= 12.18; M201= 5.593; M220= 0.3221; M211= -1.478; M202= 0.2321; M230= -0.07087;
M221= -0.1052; M212= 0.1796; M203= -0.02966; M240= -0.02738; M231= 0.03634; M222= 0.01021; M213= -0.01708;
M204= 0.002373; M250= -0.001824; M241= 0.01689; M232= -0.02016; M223= 0.009519; M214= -0.003368;
M205= 0.0006478;

The fifth degree equation for the temperature ) at (2.) point of refrigerating circuit for R134a and its constants
are given below:

(20)

E200= 79.13; E210= 11.19; E201= 8.426; E220= -0.108; E211= -1.443; E202= 0.2895; E230= -0.02414; E221= 0.02083;
E212= 0.1553; E203= -0.04372; E240= -0.00581; E231= 0.009931; E222= -0.00681; E213= -0.01533; E204= 0.004695;
E250= -0.0002447; E241= 0.003315; E232= -0.004342; E223= 0.003945; E214= -0.0002318; E205= 5.989e-005;

The fifth degree equation for the entropy ) at (2.) point of refrigerating circuit for R1234yf and its constants are
given below:

(21)

M300= 0.8176; M310= 0.07376; M301= -0.0701; M320= -0.01208; M311= 0.03507; M302= -0.01606; M330= 0.0106;
M321= -0.04143; M312= 0.05069; M303= -0.02008; M340= -0.01115; M331= 0.05162; M322= -0.08823;

243
M313= 0.06486; M304= -0.01671; M350= 0.001081; M341= -0.01018; M332= 0.02919; M323= -0.03606;
M314= 0.01955; M305= -0.003771;

The fifth degree equation for the entropy ) at (2.) point of refrigerating circuit for R134a and its constants are
given below:

(22)

E300= 0.9471; E310= 0.07013; E301= -0.06345; E320= -0.007538; E311= 0.01925; E302= -0.003731; E330= 0.003185;
E321= -0.0119; E312= 0.01264; E303= -0.005; E340= -0.001438; E331= 0.006864; E322= -0.01123; E313= 0.007381;
E304= -0.001173; E350= 7.051e-005; E341= -0.0009764; E332= 0.002884; E323= -0.003244; E314= 0.001412;
E305= -0.0003036;

Modelling of Thermodynamic Properties at Refrigerating Circuit (fourth point) for R1234yf and R134a
refrigerants
In refrigerating circuit, expansion valve outlet is the same as evaporator inlet and known as fourth (4) point. At this
point, refrigerant pressure is reduced by expansion valve and then, it becomes a mixture of saturated liquid and
vapour state at low pressure. İt is assumed that refrigerant enthalpy value (fourth point) equals to enthalpy value at
the evaporator inlet. This point is assumed to be equal to the evaporator pressure thermodynamically. At this point,
the entropy values of the refrigerant should be calculated using equations (31-32).

The tenth degree equation for the liquid enthalpy ) at (4.) point of refrigerating circuit for R1234yf and
its constants are given below:

(23)

N11= -1.2642e-007; N12= 1.3033e-006; N13= 8.0987e-006; N14= 8.7358e-006; N15= -1.7434e-005;
N16= 0.00011428; N17= 0.0016441; N18= 0.015424; N19= 0.46706; N110= 20.654; N111= 52.171;

The tenth degree equation for the liquid enthalpy ) at (4.) point of refrigerating circuit for R134a and its
constants are given below:

(24)

F11= 3.822e-008; F12= 9.2488e-008; F13= 1.8852e-007; F14= 3.3918e-006; F15= 2.8359e-005; F16= 0.00024585;
F17= 0.0022767; F18= 0.024506; F19= 0.35558; F110= 21.441; F111= 55.223;

The tenth degree equation for the vapour enthalpy ) at (4.) point of refrigerating circuit for R1234yf
and its constants are given below:

(25)

N21= 5.9063e-007; N22= -2.5853e-006; N23= -1.4866e-005; N24= 2.2592e-005; N25= 0.00012763;
N26= -0.00050034; N27= -0.0051633; N28= -0.044585; N29= -0.21548; N210= 10.259; N211= 213.84;

244
The tenth degree equation for the vapour enthalpy ) at (4.) point of refrigerating circuit for R134a and
its constants are given below:

(26)

F21= -1.8768e-008; F22= -1.3541e-007; F23= -8.7028e-007; F24= -5.0222e-006; F25= -4.5396e-005;
F26= -0.00044732; F27= -0.0025895; F28= -0.038175; F29= -0.29894; F210= 9.1743; F211= 251.91;

The tenth degree equation for the liquid entropy at (4.) point of refrigerating circuit for R1234yf and its
constants are given below:

(27)

N31= -7.6966e-009; N32= 1.2916e-008; N33= 9.0355e-008; N34= -3.4339e-008; N35= -2.7775e-007;
N36= 4.3107e-007; N37= 2.6062e-006; N38= 6.401e-005; N39= -0.00057241; N310= 0.074367; N311= 0.20433

The tenth degree equation for the liquid entropy at (4.) point of refrigerating circuit for R134a and its
constants are given below:

(28)

F31= -7.6233e-009; F32= 2.5742e-008; F33= 6.1495e-008; F34= -1.5835e-007; F35= -9.2099e-008;
F36= 1.0994e-006; F37= 2.5469e-006; F38= 0.00011508; F39= -0.001056; F310= 0.077273; F311= 0.21658;

The tenth degree equation for the vapour entropy at (4.) point of refrigerating circuit for R1234yf and
its constants are given below:

(29)

N41= -3.3376e-007; N42= 1.1336e-007; N43= 2.4622e-006; N44= -9.4077e-007; N45= -5.9507e-006;
N46= 1.1512e-006; N47= 2.1508e-006; N48= -0.00025084; N49= 0.0010678; N410= 0.0028724; N411= 0.79084;

The tenth degree equation for the vapour entropy at (4.) point of refrigerating circuit for R134a and its
constants are given below:

(30)

F41= 9.9529e-010; F42= -1.6686e-009; F43= -2.2658e-008; F44= -5.8367e-009; F45= 3.8943e-008;
F46= -2.0287e-006; F47= 6.2254e-006; F48= -0.00025924; F49= 0.0015001; F410= -0.0083261; F411= 0.93013;

For R1234yf refrigerant; equation of the entropy values at (4) point:

(31)

For R134a refrigerant; equation of the entropy values at (4) point:

(32)

245
THERMODYNAMIC ANALYSIS OF THE MODELLING SYSTEM
Thermodynamic model for the system that uses R1234yf and R134a is shown in Figure 1. Performance
parameters were obtained and presented in graphical forms for two refrigerants. For this aim, thermodynamic
analysis equations are given below;

Cooling capacity can be evaluated from:

(33)

The power absorbed by the refrigerant in the compressor can be calculated from:

(34)

The COP for the AAC system:

(35)

When heat transfer to the environment is neglected, rate of the exergy destruction in compressor:

(36)

Rate of exergy destruction in condenser and liquid line due to heat transfer to the environment can be calculated :

(37)

Assuming that the expansion process is adiabatic, rate of exergy destruction in expansion valve can be expressed
as:

(38)

Rate of exergy destruction in evaporator can be determined from:

(39)

Finally, the rate of total exergy destruction in the refrigeration circuit of the system can be evaluated from:

(40)

RESULT AND DISCUSSION

In automotive air conditioning system that using R1234yf and R134a refrigerants, variations in the performance
parameters with the evaporating, condensing and superheat temperatures are shown in Figs.3-7. In these figures,
superheat, evaporating and condensing temperatures and other system parameters are given in Table 1. Also
deviation ratios are given in Table 2.

Table 1. System inlet parameters and values for thermodynamic analysis

Inlet Parameters Values Unit
Evaporating temperature -10,-6,-2,2,6,10 (°C)
Condensing temperature 40,45,50,55,60 (°C)
Superheat temperature 2,6,10 (°C)
Subcooling temperature 8 (°C)
Mass flow rate of refrigerant 0.050 (kg/s)
İsentropic compressor efficiency 0.7
Ambient temperature 27 (°C)
Evaporator and condenser inlet temperature 35 (°C)

246
Table 2. Deviations (Difference between modelled results and real values)
R1234yf R134a
T2 %±0.095 %±0.062
Qevap %±0.065 %±0.060
Wkomp %±0.094 %±0.150
Qkond %±0.011 %±0.045
Cop %±0.150 %±0.210
Edtot %±0.294 %±0.350

Variations in the compressor discharge temperature of refrigerant with the condensing and evaporating
temperatures are shown in Fig. 3. It is seen that the compressor discharge temperature increases with increasing
superheat and condensing temperature while it decreases with increasing evaporating temperature. Also the
system using R1234yf refrigerant has less compressor discharge temperature (in the range of 9-15 oC) than the
other one. In the case of increasing condensing and evaporating temperatures of system using R1234yf
refrigerant, it becomes more steady than the other system.

Variations in the compressor power with the condensing and evaporating temperatures are shown in Fig. 4. It is
seen that the compressor power increases with increasing superheat and condensing temperature while it
decreases with increasing evaporating temperature for both systems. Also R1234yf refrigerant has less
compressor capacity (0.59-0.35 kW) than R134a refrigerant.

Fig.3. Variations in the discharge temperature with the condensing and evaporating temperature

Fig. 4. Variations in the compressor power with the condensing and evaporating temperature

Variations in the cooling capacity with the condensing and evaporating temperatures are shown in Fig. 5. It is
seen that the cooling capacity decreases with increasing superheat and condensing temperature while it increases
with increasing evaporating temperature for both systems. Also R1234yf refrigerant has less cooling capacity
(21-25%) than R134a refrigerant.

Variations in the coefficient of performance (COP) with the condensing and evaporating temperature are shown in
Fig. 6. It is seen that the COP decreases with increasing superheat and condensing temperature while it increases
with increasing evaporating temperature for both systems. Also R1234yf refrigerant has less COP (2.6-5%) than

247
R134a refrigerant. In addition to this, the COP of system using R1234yf refrigerant is increasing when the
superheat temperature increases. However, a significiant change was not observed for system using R134a.

Fig. 5. Variations in the cooling capacity with the condensing and evaporating temperature

Fig. 6. Variations in the coefficient of performance (COP) with the condensing and evaporating temperature

Fig.7. Variations in the total exergy destruction of circuit with the condensing and evaporating temperature

Variations in the total exergy destruction of circuit (circuit) with the condensing and evaporating temperature are
shown in Fig. 7. It is seen that the total exergy destruction of circuit increases with increasing superheat and
condensing temperature while it decreases with increasing evaporating temperature for both systems. Also
R1234yf has less total exergy destruction (0.57-0.34 kW) than R134a refrigerant.

248
CONCLUSIONS
In this study, performance parameters of an automotive air conditioning system using R1234yf and R134a
refrigerants have been presented depending on different evaporating, condensing and superheat temperatures.
Thermodynamic properties of R1234yf and R134a were modelled as a function of pressure and temperature.
Then, a program was developed using Matlab to obtain performance parameters of the automotive air
conditioning. By means of this program, the compressor capacity, discharge temperature, cooling capacity, COP
and total exergy destruction of the system were determined as functions of evaporating and condensing
temperatures. Reached main results are as follows:

 For both systems; It was seen that the total exergy destruction of circuit, compressor power and discharge
temperature increase with increasing superheat and condensing temperature while they decrease with increasing
evaporating temperature. Also it was determined that R1234yf refrigerant have less discharge temperature (in the
range of 9-15 oC), less compressor capacity (0.59-0.35kW) and less total exergy destruction (0.57-0.34 kW) than
R134a refrigerant.

 For both systems; It was seen that the COP and cooling capacity decrease with increasing superheat and
condensing temperature while they increase with increasing evaporating temperature. Also it was observed that
the R1234yf refrigerant has less COP (2.6-5%) and less cooling capacity (21-25%) than R134a refrigerant.

SYMBOLS
AAC : Automotive air conditioning
E d :Rate of exergy destruction (W)
Qevap :Cooling capacity (W)
W komp :Compressor power (W)
COP :Coefficent of performance
s :Specific entropy ((kJ/kg K)
h :Specific enthalpy (kJ/kg)
m :Mass flow rate (kg/s)
T :Temperature (°C)
T0 :Ambient temperature (°C)
:Efficiency

Subscripts
evap :Evaporatör
comp :Compressor
cond :Condenser
r :Refrigeration
tot :total
g :Vapour
l :Liquid
1....4 :Refrigerating circuit points
is :isentropic

REFERENCES
Al-Rabghi, O.M. and A.A. Niyaz. 2000. Retrofitting R-12 car air conditioner with R-134a refrigerant. International
Journal Energy Research. 24:467-474.

Alkan, A. and M Hosoz. 2010. Comparative performance of an automotive air conditioning system using fixed and
variable capacity compressors. International Journal of Refrigeration. 33:487−495.

Col, D.D., S. Bortolin, D. Torresin, A. Cavallini. 2013. Flow boiling of R1234yf in a 1 mm diameter channel. 36:353-
362.

Esen, D.Ö. and M. Hoşöz. 2006. R12 ve R134a Soğutucu Akışkanları Kullanan Otomobil İklimlendirme Sisteminin
Deneysel Performans Analizi. Gazi Üniversitesi Mühendislik Mimarlık Fakültesi Dergisi. 21:703-709.

Esen, D.Ö. and M. Hoşöz. 2005. Soğutucu akışkanları ve kompresör devrinin otomobil klimalarının performansına
etkisinin deneysel analizi. Tesisat Dergisi. TMMOB Ulusal İklimlendirme Sempozyumu ve Sergisi.

249
Hosoz, M. and M. Direk. 2006. Performance evaluation of an integrated automotive air conditioning and heat pump
system. Energy Conversion and Management. 47:545-559.

Hosoz, M. and H.M. Ertunc. 2006. Artificial Neural Network Analysis of an Automotive Air Conditioning System.
Energy Conversion and Management. 46:1574-1587.

Hosoz, M. 2005. Exergy Analysis of Automobile Air Conditioning System Using R134a. Proceedings of the Second
International Exergy, Energy and Environment Symposium. Kos. Greece. July 3-7.

Jabardo, J.M.S., W.G. Mamani and M.R. lanella. 2002. Modelling and experimental evaluation of an automotive air
conditioning system with a variable capacity compressor. International Journal Refrigeration 25:1157-1172.

Joudi, K.A., A.S. Mohammed, and M.K. Aljanabi. 2003. Experimental and computer performance study of an
automotive air conditioning system with alternative refrigerants. Energy Conversion and Management. 44: 2959-
2976.

Jung, D., B. Park. and H. Lee. 1999 Evaluation of supplementary/retrofit refrigerants for automobile air-conditioners
charged with CFC12. International Journal Refrigeration. 22:558-568.

Kaynakli, O. and I. Horuz. 2003. An experimental analysis of automotive air conditioning system. International
Communications Heat and Mass Transfer. 30:273-284.

Kiatsiriroat, T. and T. Euakit. 1997. Performance analyses of an automobile air-conditioning with

R22/R124/R152a refrigerant. Applied Thermal Engineering 17: 1085-1097.

Lee, G.H. and J. H. Yoo. 2000. Performance analysis and simulation of automobile air conditioning system.
International Journal Refrigeration. 23:243-254.

Lemmon, E.W., M.L. Huber, M.O. McLinden. NIST Standard Reference Database 23: Reference Fluid
Thermodynamic and Transport Properties—REFPROP, Version 9.0 (National Institute of Standards and
Technology, Standard Reference Data Program, Gaithersburg, MD, 2010)

Zhao, Y., Y. Liang, Y. Sun, J. Chen. 2012. Development of a mini-channel evaporator model using R1234yf as
working fluid. International Journal Refrigeration. 35:2166-2178.

Zilio, C., J. S. Brown, G. Schiochet, A. Cavallini. 2011. The refrigerant R1234yf in air conditioning systems.
Energy. 36:6110-6120.

250
 PARAMETRIC ANALYSIS OF PUMPED STORAGE HYDROPOWER-COUPLED
WIND TURBINE PLANTS

K. Deger, B. Kilkis, T. Yavuz

MSc in Energy Engineering, BSCE
Baskent University, Mechanical Engineering Dept. Ankara, Turkey
k_deger@yahoo.com.tr
bkilkis@baskent.edu.tr
tyavuz@baskent.edu.tr

ABSTRACT

The concept of a hybrid system consisting of a wind turbine plant and a pumped storage hydropower plant is a
promising idea with respect to a number of important factors. Wind turbine plant power output has a fluctuating
and interrupted nature and the penetration of the electrical energy to the grid usually poses problems to the steady
operation of the system. Pumped storage hydropower is a mature technology in the energy storage sector.
Pumped storage hydropower has advantages compared to other energy storage technologies such as CAES,
batteries etc. due to its short discharge time, high energy storage capacities, and high power output. The first
objective of this study is to time-shift the power output of the wind turbine plant to the peak-demand periods by
means of storage and thus realizing increase in the penetration level to the grid. Second objective is to obtain
more stable power output from the wind turbine plant that inevitably has a variable power output. A one-year
operation simulation of the hybrid system was carried out under peak tariff strategy within the scope of these two
objectives. An operation simulation was also made covering only the wind turbine plant with same power output
data and same demand strategy and the results were compared. Without storage, the wind turbine plant feeds
6,350.30 MWh of electrical energy to the grid, this value increases to 7,656.00 MWh for the hybrid system, this
means that there is 17.05% increase in penetration and the system is technically applicable for the first objective.
Annual stable power output ratio that shows the fulfillment of demand is approximately 72.49% for wind turbine
plant without storage and this value is approximately 87.40% for the hybrid system. This means that there is a
14.91% increase and the system is technically applicable for the second objective. This paper provides the details
of the year-long analysis, draws generalized conclusions, and provides recommendations for future studies.

NOMENCLATURE

𝑐 Scale parameter (m/s)

𝐸(𝑡) Energy storage level in reservoir, during 𝑡 interval (MWh)
𝐸(𝑡 + 1) Energy storage level in reservoir, during (𝑡 + 1) interval (MWh)
𝐸 max Maximum energy storage capacity of reservoir (MWh)
𝐸 min Minimum energy storage capacity of reservoir (MWh)
𝐸 a𝑑 Electrical energy demand of grid (annual) (MWh)
𝐸 a𝑑𝑒𝑓 Unmet (deficient) electrical energy demand of grid (annual) (MWh)
𝐸 a𝑑𝑔 Total electrical energy delivered to grid (annual) (MWh)
𝐸 a𝑒𝑥 Excess electrical energy (annual) (MWh)
𝐸 aℎ Electrical energy produced at PSHP (annual) (MWh)
𝐸 a𝑝 Electrical energy consumed by PSHP’s pumps (annual) (MWh)
𝐸 a𝑣 Electrical energy produced at WTP (annual) (MWh)
𝐸 a𝑤 Electrical energy delivered to grid by WTP (annual) (MWh)
𝐸 in
1 Initial energy storage level of reservoir (MWh)
𝑘 Shape parameter
𝑃𝑑 (𝑡) Power demand of grid, during 𝛥𝑡 interval (MW)
𝑃𝑑𝑒𝑓 (𝑡) Unmet (deficient) power demand of grid, during 𝛥𝑡 interval (MW)
𝑃𝑑𝑔 (𝑡) Total delivered power to grid, during 𝛥𝑡 interval (MW)
𝑃𝑒𝑥 (𝑡) Excess power output, during 𝛥𝑡 interval (MW)
𝑃ℎ (𝑡) Power output of PSHP’s turbines, during 𝛥𝑡 interval (MW)
𝑃max
ℎ Maximum power output of PSHP’s turbines (MW)
𝑃min
ℎ Minimum power output of PSHP’s turbines (MW)
𝑃𝑝 (𝑡) Power of PSHP’s pumps, during 𝛥𝑡 interval (MW)
𝑃max
𝑝 Maximum power of PSHP’s pumps (MW)
𝑃min
𝑝 Minimum power of PSHP’s pumps (MW)
𝑃𝑣 (𝑡) Power output of WTP’s turbines, during 𝛥𝑡 interval (MW)

251
𝑃max
𝑣 Maximum power output of WTP’s turbines (MW)
𝑃min
𝑣 Minimum power output of WTP’s turbines (MW)
𝑃𝑤 (𝑡) Power delivered to grid from WTP, during 𝛥𝑡 interval (MW)
𝑃𝑤𝑡 Power output of wind turbine (MW)
����
𝑃ℎ𝑠 Stable power output ratio of HS (annual) (%)
�����
𝑃𝑜𝑤 Stable power output ratio of WTP (annual) (%)
𝑡 Time index
𝑣 Wind speed (m/s)
𝑣𝑚 Average wind speed (m/s)
𝛼 Hellman exponent
𝛥𝑡 Duration of interval (hour)
𝜂𝑐 Isentropic charge efficiency of HS
𝜂𝑑 Isentropic discharge efficiency of HS
𝜂𝑡 Isentropic cycle efficiency of HS
𝜌𝑎 Air density (kg/m3)

ACRONYMS

ESS Energy Storage System

HS Hybrid System
PSHP Pumped Storage Hydropower Plant
TGO Transmission Grid Operator
WTP Wind Turbine Plant

1. INTRODUCTION

Energy production from renewable sources is affected by private factors such as climatic and topographical
conditions of project implementation area and, as a result; energy production levels have a variable structure in
time. Wind turbine plants (WTP) could not provide continuous power output due to variable wind speed and air
density thus energy production level varies by time, therefore the produced energy level that is not harmonized
with grid load level could not be recovered, if power output level is lower than the grid load, demand could not be
met in full, if the power output amount is greater than the latter, some amount of excess power occurs. Also the
variable power output of WTPs could cause problems on grids that have inappropriate design; in this respect a
hybrid system (HS) design which energy storage system (ESS) is the other component could be a solution for to
time-shift power output and for to obtain a more stable power output. While considering the superior factors it has,
pumped storage hydropower plant (PSHP) could be opted as the component of HS among various ESSs, and
PSHP is selected as the ESS component of the HS. An operation simulation numerical model is developed to
investigate technical benefits of a hypothetical HS model. The numerical model is based on power and energy
units and the results are compared with the results obtained from the simulation that was carried out for WTP
without storage.

The HSs are investigated commonly for isolated grids in literature. Results of HS model’s optimization for an
isolated grid show that there is an increase in wind energy penetration to the grid (Brown et. al. 2008). A general
model was also presented for the technical and economic sizing of HSs in another study (Bueno et al., 2005), and
installation of a HS for an isolated grid is proposed by same researchers with a following study, according to the
results; with the application of optimum sized economic model, the renewable sourced energy penetration to the
grid increases (Bueno et al., 2006). In particular, the aspects of control and operation of a variable speed PSHP for
the objective of wind power integration into an isolated grid is also studied (Suul et al., 2008). Stochastic based
approach for optimization of HSs is another fact also. WTP power output amount and electricity unit selling prices
vary in time; these parameters could be evaluated in scope of stochastic models (García-Gonzáles et al., 2008),
(Gao et al., 2009). In various studies economic benefit based models are developed for HSs (Castronuovo et al.,
2004), (Ngoc et al., 2009). In those studies objective function is derived to gain maximum profit and linear
algorithm approach is used for optimization, and the future values of variables such as power output or grid
electrical energy demand were obtained by forecasting techniques.

Installation objectives directly influence the design approach of the HS model. There are two objectives of the HS
model analyzed in this study. Climatic and meteorological conditions influence wind speed and air density and
consequently WTP power output fluctuates in time therefore the power output level of WTP could not be
harmonized with grid load or with the level of power amount committed to the transmission grid operator (TGO) in
some periods. If the energy produced at WTP is higher than the level committed to TGO, then excess energy
occurs, otherwise hence the produced energy level is lower than the committed level, then the commitment is not

252
fulfilled and economic loss may occur in producer side. The first objective of the HS model is to time-shift the
power output of the WTP to peak demand periods by means of storage and thus realizing increase in the
penetration level to the grid. Also the fluctuation occurs in WTP power output could affect the grid. If the grid is not
designed especially for such conditions, the power fed to grid can cause stability problems. With an appropriate
operation approach, the power output of WTP could be smoothed by power output of PSHP and a more stable
power output level could be obtained. Second objective is to obtain more stable power output from the WTP which
normally has a variable power output. The penetration restrictions are not included in HS model and also it is
assumed that when WTP itself feeds electrical energy to the grid, the stability of the grid is not influenced.
Renewable energy time-shift and wind generation grid integration are some benefits of ESSs. At the present day
there are various mature ESS technologies, besides this, a number of ESS technologies are in development
stage. ESSs could be evaluated according to factors such as their power output capacities, cycle life, energy
storage capacities etc., PSHP, as one of the mature ESS technology, could be opted as a component of HS in the
scope of mentioned objectives while considering the superior factors such as its long cycle life, high energy
storage capacity, high power output capacity, energy input and output in the form of electrical energy and low
environmental impact; they are also eligible for renewable concept. In this study an operation simulation numerical
model is developed for a hypothetical HS model. Power and energy units are used for numerical model and the
model is based on evaluation of technical benefits. In various studies economic benefit based models are
developed for HSs and objective function is derived to gain maximum profit, linear algorithm approach is used for
optimization and the future values of variables such as power output or grid electrical energy demand were
obtained by forecasting techniques (Castronuovo et al., 2004), (Ngoc et al., 2009). However in the present paper,
the simulation is not based on the economic benefit and electrical energy prices are not included in the model
structure. Forecasting techniques are not used for to obtain wind speed and grid demand values, instead
predetermined data is used. Approach of such studies (Castronuovo et al., 2004), (Ngoc et al., 2009) for power
limits of WTP turbines’, power limits of reversible PSHP pump-turbines’ and energy capacity limits of reservoir is
preferred in this paper. However, the general numerical model of the operation simulation is developed specifically
in the scope of this study (Değer, 2013).

2. HYBRID SYSTEM AND THE MODEL DEVELOPMENT

Generally PSHPs are designed considering either one of the three types of operation approach. In daily operation
strategy, energy is generated during peak periods; emptied reservoir is filled on the same day while the demand
level is low. In weekly operation, energy is generated on the weekdays and emptied reservoir is filled on the
weekends. Seasonal operation strategy is the last one. In the seasonal strategy the PSHP is designed for other
purposes besides energy production. In this study the daily operation strategy is opted.

There can be one or more than one penstock in the PSHP scheme. The configuration that has more than one
penstock enables pumping and energy production synchronously. Excess energy output of WTP could be stored
in PSHP’s upper reservoir while energy is generated at the turbines of PSHP, notwithstanding the mentioned
advantages; additional penstocks increase the initial cost. In this study the PSHP model has one penstock. PSHPs
could be designed in the way where the pump units and turbine units are separated units. Besides this, reversible
pump-turbine and variable speed motor-generator arrangement is another option and this option is preferred in the
study. This type of arrangement has benefits such as; adjustability, compensation of WTP power output fluctuation
and consequently improvement in the functionality of the grid.

Generally PSHPs consume electrical energy from the grid in low demand periods and produce electrical energy in
peak periods and feed the grid. Energy consumption from grid may be considered as an option for providing
electrical energy that is committed to TGO when amount of energy that could be provided by HS is insufficient and
for also preventing over sizing of the reservoirs. Here the technical benefits of the HS and WTP without storage
are compared under same conditions; for this type of approach HS just should consume electrical energy that is
produced by WTP and so electricity consumption from grid for pumping is not allowed for HS. The HS model
provides electrical energy for scheduled hours. Operation approach could be regarded as peak strategy while
considering the projected period is 6 hours a day. Power plants are generally classified as base-load plants, mid-
merit plants and peaking plants. WTP is not a peaking plant by itself. However, a design with storage may allow a
WTP to act as a peaking plant.

The HS is a hypothetical model. Specific topographical requirements for WTP and PSHP are out of the scope of
this study. HS consists of WTP, PSHP and the system operator. The previously electrical energy production
amount committed by HS is submitted to TGO at the pre-concerted time periods while the TGO is the part of the
system as the demand side component. It is assumed that WTP turbines and wind conditions are identical for
each turbine. A variable speed wind turbine is selected. The power output values of the wind turbine are given in
Table 1. Wind speeds were calculated according to the selected Weibull parameter values. The calculated wind
speeds are based on hourly intervals. Therefore there are 8760 wind speed values for a 1-year period and this

253
situation is harmonized with the duration of interval that is selected for the simulation. The synthesized wind speed
values were originally for 10 m height and the actual values for selected hub height were calculated with the power
law method (Nelson, 2009), (𝑣: Wind speed, 𝑃𝑤𝑡 : Power output of wind turbine)

Table 1. Wind turbine power output values for different wind speeds (pc_aaer.xls, 2012).
𝑣 (m/s) 𝑃𝑤𝑡 (MW)
0-3 0
4 0.029
5 0.071
6 0.138
7 0.238
8 0.361
9 0.507
10 0.673
11 0.860
12-21.99 1.001
22 0

The technical specifications of the WTP model are given in Table 2 (𝑘: Shape parameter, 𝑐: Scale parameter, 𝛼:
Hellman exponent, 𝜌𝑎 : Air density, 𝑣𝑚 : Average wind speed, 𝑃max
𝑣 : Maximum power output of WTP’s turbines, 𝑃min
𝑣 :
Minimum power output of WTP’s turbines).

Table 2. Technical specifications summary of the WTP model.

Model specifications Explanation/selection
𝑘 1.73
𝑐 (m/s) 8.11
𝛼 0.14
𝜌𝑎 (kg/m3) 1.225
𝑣𝑚 (m/s) 7.23
Wind speed measurement height (m) 10
Measurement interval (hour) 1
Wind turbine hub height (m) 70
Number of wind turbines 10
𝑃max
𝑣 (MW) 10.01
𝑃min
𝑣 (MW) 0.00

The technical specifications of the PSHP model are given in Table 3 ( 𝐸 in 1 : Initial energy storage level of
reservoir, 𝐸 max : Maximum energy storage capacity of reservoir, 𝐸 min : Minimum energy storage capacity of
reservoir, 𝑃max
ℎ : Maximum power output of PSHP’s turbines, 𝑃min ℎ : Minimum power output of PSHP’s
turbines, 𝑃𝑝 : Maximum power of PSHP’s pumps, 𝑃min
max
𝑝 : Minimum power of PSHP’s pumps)

Table 3. Technical specifications summary of the PSHP model.

Model specifications Explanation/selection
Type Pure PSHP
Operation period Daily basis
Number of the penstocks 1
Source of electrical energy for pumping WTP
𝐸 in
1 (MWh) 1.00
𝐸 max (MWh) 7.00
𝐸 min (MWh) 1.00
𝑃max
ℎ (MW) 7.00
𝑃min
ℎ (MW) 1.00
𝑃max
𝑝 (MW) 7.00
𝑃min
𝑝 (MW) 1.00

254
Operation simulation strategy of the HS is explained below.

• Operation of PSHP’s turbines: The potential energy stored in the upper reservoir of PSHP is converted to
electrical energy via reversible pump-turbines and variable speed motor-generators. With the variable
speed motor-generators the power output is adjusted between available minimum and maximum values.
This operation step is applicable for the intervals that wind energy is not available.

• Operation of PSHP’s pumps: Energy output of the WTP is stored as potential energy in the upper
reservoir by pumping. Variable speed motor-generators are operated in motor-pump mode between the
minimum and maximum power output values. As mentioned before; electrical energy from grid is not used
for pumping. This operation step is applicable as if there is sufficient wind energy, there is no grid demand
and also there is enough space for storage in the upper reservoir.

• Operation of WTP’s turbines: As the power output values depends on the wind speed; with the selection
of variable speed wind turbines the electrical power output could be adjusted between the minimum and
maximum values. Energy produced at WTP is fed to grid at first.

Also there are other situations for the operation simulation strategy of the HS.

• Operation of PSHP’s turbines and WTP’s turbines together: Grid electrical power demand is met with the
WTP and PSHP together. This operation step is applicable as if the electrical energy production level of
WTP is lower than the grid demand level and there is adequate stored potential energy in the upper
reservoir.

• Operation of PSHP’s pumps and WTP’s turbines together: If the electrical energy produced at WTP is
adequate and there is enough space for storage in the upper reservoir of PSHP; power output of the WTP
is fed to the grid and also it is consumed by PSHP’s pumps and energy is stored in the upper reservoir.

Other details of the operation simulation are listed below.

• Grid demand: Grid demand is selected as constant 4 MW for 6 hours a day. In the model, grid load is as
well as equal to the electrical energy production amount that is submitted to TGO by HS.

• Duration of the interval: While taking into account the start-up and shut-down periods of the components
of the HS, the selected 𝛥𝑡 value is one hour (𝛥𝑡: Duration of interval).

• Simulation period: Simulation period is determined as 1 year. Values of power output and energy
production variables of the HS model are calculated on hourly basis for one year. This approach is opted
for also WTP without storage.

A representation of HS model is given in Figure 1. The summary of the technical specifications of HS model is
given in Table 4 (𝑃𝑑 (𝑡): Power demand of grid, during 𝛥𝑡 interval, 𝑡: Time index, 𝜂𝑐 : Isentropic charge efficiency of
HS, 𝜂𝑑 : Isentropic discharge efficiency of HS, 𝜂𝑡 : Isentropic cycle efficiency of HS).

Figure 1. Representation of the HS model.

255
Table 4. Technical specifications summary of the HS model.
Model specifications Explanation/selection
Type of power plant Peaking plant
Daily power output period of HS (hour) 6 (Maximum)
𝑃𝑑 (𝑡) (MW) 4
𝛥𝑡 (hour) 1
𝜂𝑐 0.86
𝜂𝑑 0.88
𝜂𝑡 0.76

Selections and constraints for system components are given by following equations.

WTP’s turbines have minimum and maximum power output limits (1), (𝑃𝑣 (𝑡): Power output of WTP’s turbines,
during 𝛥𝑡 interval).

𝑃min
𝑣 ≤ 𝑃𝑣 (𝑡) ≤ 𝑃max
𝑣 (1)

PSHP’s turbines have minimum and maximum power limits, besides them the available energy level of the upper
reservoir is the other constraint (2), (𝑃ℎ (𝑡): Power output of PSHP’s turbines, during 𝛥𝑡 interval, 𝐸(𝑡): Energy
storage level in reservoir, during 𝑡 interval).

𝐸(𝑡) − 𝐸 min
𝑃min
ℎ ≤ 𝑃ℎ (𝑡) ≤ min �� � . 𝜂𝑑 , 𝑃max
ℎ � (2)
𝛥𝑡
PSHP’s pumps have minimum and maximum power limits, besides them the available space for of the upper
reservoir to store energy is the other constraint (3), (𝑃𝑝 (𝑡): Power of PSHP’s pumps, during 𝛥𝑡 interval).

𝐸 max − 𝐸(𝑡) 1
𝑃min
𝑝 ≤ 𝑃𝑝 (𝑡) ≤ min �� � . , 𝑃max � (3)
𝛥𝑡 𝜂𝑐 𝑝
Maximum power limits of PSHP’s turbines and pumps are equal and also minimum power limits of PSHP’s
turbines and pumps are equal similarly (4), (5).

𝑃max
ℎ = 𝑃max
𝑝 (4)
𝑃min
ℎ = 𝑃min
𝑝 (5)

PSHP’s upper reservoir has an initial storage value; this is equal to minimum energy level of the reservoir (6).

𝐸 in
1 = 𝐸
min
(6)

PSHP’s upper reservoir has minimum and maximum energy storage capacity limits (7), (8), (9).

𝑃ℎ (𝑡)
𝐸(𝑡) − 𝛥𝑡 ∙ ≥ 𝐸 min (7)
𝜂𝑑
𝐸(𝑡) + 𝛥𝑡. 𝜂𝑐 . 𝑃𝑝 (𝑡) ≤ 𝐸 max (8)
𝐸 min ≤ 𝐸(𝑡) ≤ 𝐸 max (9)

There are two main simulation steps according to the check of actual 𝑃𝑣 (𝑡) and 𝑃𝑑 (𝑡) values that is made by
system operator.

1th situation: Actual 𝑃𝑣 (𝑡) and 𝑃𝑑 (𝑡) values are controlled by system operator. If the power output of the WTP is
greater than or equal to grid demand during 𝛥𝑡 interval then grid demand is primarily met by WTP output (10), (11).
In this case WTP output is also equal to the total delivered power to the grid, 𝑃𝑑𝑔 (𝑡), (12), (𝑃𝑤 (𝑡): Power delivered
to grid from WTP, during 𝛥𝑡 interval, 𝑃𝑑𝑔 (𝑡): Total delivered power to grid, during 𝛥𝑡 interval).

𝑃𝑣 (𝑡) ≥ 𝑃𝑑 (𝑡) (10)

𝑃𝑤 (𝑡) = 𝑃𝑑 (𝑡) (11)

256
𝑃𝑤 (𝑡) = 𝑃𝑑𝑔 (𝑡) (12)
If the power output of the WTP is greater than the grid demand during 𝛥𝑡 interval then for the difference value of
the delivered grid and power output of WTP (𝑃𝑣 (𝑡) − 𝑃𝑑 (𝑡)) three probable sub-situations could occur according to
minimum and maximum power limits of the pumps and available storable space of the reservoir. 𝑃𝑒𝑥 (𝑡) term is
defined (13), this term indicates the excess power occurred due to mentioned restrictions, (𝑃𝑒𝑥 (𝑡): Excess power
output, during 𝛥𝑡 interval). The equation shows energy storage level regarding the operation of PSHP’s pumps
during the 𝛥𝑡 interval is given by (14), (𝐸(𝑡 + 1): Energy storage level in reservoir, during (𝑡 + 1) interval).

𝑃𝑒𝑥 (𝑡) = 𝑃𝑣 (𝑡) − 𝑃𝑑 (𝑡) − 𝑃𝑝 (𝑡) (13)

𝐸(𝑡 + 1) = 𝐸(𝑡) + 𝛥𝑡. 𝜂𝑐 . 𝑃𝑝 (𝑡) (14)

2nd situation: Actual 𝑃𝑣 (𝑡) and 𝑃𝑑 (𝑡) values are controlled by the system operator. If the power output of the WTP is
less than the grid demand during 𝛥𝑡 interval then (15), the stored energy amount in the upper reservoir is
controlled and three probable sub-situations could occur according to minimum and maximum power limits of the
turbines and available energy capacity of the reservoir. Here 𝑃𝑑𝑔 (𝑡) is equal to the sum of WTP power output and
PSHP’s turbines power output (16). 𝑃𝑑𝑒𝑓 (𝑡) term is defined (17), this term indicates the unmet energy demand level
even if the WTP and the PSHP contribute to energy production together (𝑃𝑑𝑒𝑓 (𝑡): Unmet (deficient) power demand
of grid, during 𝛥𝑡 interval). The equation shows energy consumption from upper reservoir regarding operation of
PSHP’s turbines during the 𝛥𝑡 interval is given by (18).

𝑃𝑣 (𝑡) < 𝑃𝑑 (𝑡) (15)

𝑃𝑑𝑔 (𝑡) = 𝑃𝑤 (𝑡) + 𝑃ℎ (𝑡) (16)
𝑃𝑑𝑒𝑓 (𝑡) = 𝑃𝑑 (𝑡) − 𝑃𝑑𝑔 (𝑡) (17)
𝑃ℎ (𝑡) (18)
𝐸(𝑡 + 1) = 𝐸(𝑡) − 𝛥𝑡 ∙
𝜂𝑑
In order to compare with HS model and consequently to evaluate the technical benefits of the HS, an operation
simulation for WTP without storage is also made under same conditions.

1th situation: Actual 𝑃𝑣 (𝑡) and 𝑃𝑑 (𝑡) values are controlled by the system operator. If the power output of the WTP is
greater than or equal to grid power demand during 𝛥𝑡 interval then it is met by WTP output, in case of the excess
energy, 𝑃𝑒𝑥 (𝑡) occurs (19), (20), (21).

𝑃𝑣 (𝑡) ≥ 𝑃𝑑 (𝑡) (19)

𝑃𝑤 (𝑡) = 𝑃𝑑 (𝑡) (20)
𝑃𝑒𝑥 (𝑡) = 𝑃𝑣 (𝑡) − 𝑃𝑑 (𝑡) (21)

2nd situation: Actual 𝑃𝑣 (𝑡) and 𝑃𝑑 (𝑡) values are controlled by system operator. If the power output of the WTP is
less than the grid power demand during 𝛥𝑡 interval then, the grid demand is met at the level of WTP output
and 𝑃𝑑𝑒𝑓 (𝑡) occurs (22), (23), (24).

𝑃𝑣 (𝑡) < 𝑃𝑑 (𝑡) (22)

𝑃𝑤 (𝑡) = 𝑃𝑣 (𝑡) (23)
𝑃𝑑𝑒𝑓 (𝑡) = 𝑃𝑑 (𝑡) − 𝑃𝑣 (𝑡) (24)

3. RESULTS AND DISCUSSION

While results are evaluated with respect to first objective of the HS; annual electrical energy demand of the grid is
8,760.00 MWh and annual electrical energy produced at WTP is 41,800.09 MWh. These values are same for HS
and WTP without storage as the comparison is made under same conditions.

Without storage, the WTP feeds 6,350.30 MWh of electrical energy to the grid, this value increases to 7,656.00
MWh for the HS; this means that there is 17.05% increase in penetration level. Without storage, annual excess

257
electrical energy value is 35,449.79 MWh, this value decreases to 34,103.47 MWh for the HS; this means that
there is 3.80% decrease. Without storage, annual unmet (deficient) electrical energy demand of the grid is
2,409.70 MWh, this value decreases to 1,104.00 MWh for the HS; this means that there is 54.19% decrease.

The annual energy values according to the results (25), (26), (27), (28), (29), (30), (31), (32) are given in Table 5
(𝐸 a𝑑 : Electrical energy demand of grid (annual), 𝐸 a𝑣 : Electrical energy produced at WTP (annual), 𝐸 a𝑤 : Electrical
energy delivered to grid by WTP (annual), 𝐸 aℎ : Electrical energy produced at PSHP (annual), 𝐸 a𝑑𝑔 : Total electrical
energy delivered to grid (annual), 𝐸 a𝑝 : Electrical energy consumed by PSHP’s pumps (annual), 𝐸 a𝑒𝑥 : Excess
electrical energy (annual), 𝐸 a𝑑𝑒𝑓 : Unmet (deficient) electrical energy demand of grid (annual)).

8760

� 𝑃𝑑 (𝑡) . 𝛥𝑡 = 𝐸 a𝑑 (25)
𝑡=1
8760

� 𝑃𝑣 (𝑡) . 𝛥𝑡 = 𝐸 a𝑣 (26)
𝑡=1
8760

� 𝑃𝑤 (𝑡) . 𝛥𝑡 = 𝐸 a𝑤 (27)
𝑡=1
8760

� 𝑃ℎ (𝑡) . 𝛥𝑡 = 𝐸 aℎ (28)
𝑡=1
8760

� 𝑃𝑑𝑔 (𝑡) . 𝛥𝑡 = 𝐸 a𝑑𝑔 (29)

𝑡=1
8760

� 𝑃𝑝 (𝑡) . 𝛥𝑡 = 𝐸 a𝑝 (30)
𝑡=1
8760

� 𝑃𝑒𝑥 (𝑡) . 𝛥𝑡 = 𝐸 a𝑒𝑥 (31)

𝑡=1
8760

� 𝑃𝑑𝑒𝑓 (𝑡) . 𝛥𝑡 = 𝐸 a𝑑𝑒𝑓 (32)

𝑡=1

Table 5. Annual energy values comparison of HS and WTP.

Annual values HS WTP
𝐸 a𝑑 (MWh) 8,760.00
𝐸 a𝑣 (MWh) 41,800.09
𝐸 𝑎𝑤 (MWh) 6,350.30
𝐸 aℎ (MWh) 1,305.70 _
𝐸 a𝑑𝑔 (MWh) 7,656.00 6,350.30
𝐸 a𝑝 (MWh) 1,346.32 _
𝐸 a𝑒𝑥 (MWh) 34,103.47 35,449.79
𝐸 a𝑑𝑒𝑓 (MWh) 1,104.00 2,409.70

The 𝑃���� �����

ℎ𝑠 term is defined, this term indicates the ratio of the fulfillment of demand for HS and 𝑃𝑜𝑤 term is defined,
this term indicates the ratio of the fulfillment of demand for WTP without storage (33), (34), (𝑃 ����
ℎ𝑠 : Stable power
�����
output ratio of HS (annual), 𝑃 𝑜𝑤 : Stable power output ratio of WTP (annual)).

����
𝑃ℎ𝑠 = . 100 (33)

258
�����
𝑃𝑜𝑤 = . 100 (34)

Annual stable power output ratio is approximately 72.49% for WTP without storage and this value is approximately
87.40% for the HS. This means that there is 14.91% increase, Table 6.

Table 6. Annual stable power output ratio values comparison of HS and WTP.
Annual values ����
𝑃ℎ𝑠 (%) �����
𝑃𝑜𝑤 (%)
Stable power output ratio 87.40 72.49

The diagrams showing annual energy values for 1 year are plotted. In this paper the diagrams for one month
(January) are presented; Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, Figure 10.
The simplified flow-chart of the operation simulation algorithm is given in Figure 11.

4. CONCLUSIONS

In this paper a hypothetical HS model that consists of a combination of a PSHP and a WTP is parametrically
analyzed. As the installation objectives of the HS directly influence the design approach, two objectives were
determined in the first place. Factors such as climatic and topographical conditions of project implementation area
affect the wind speed and air density and consequently the energy production levels of WTPs have a variable
structure in time. If power output level is lower than the grid load, demand could not be met properly and deficiency
occurs, if the power output value is greater than the demand value, then an amount of excess power occurs. The
first objective of the HS is to time-shift the power output of the WTP to peak demand periods by means of storage
and thus realizing increase in the penetration level to the grid. Due to fluctuating and interrupted nature of the
power output of the WTP, penetration of the electrical energy to the grid may poses problems to the steady
operation of the grids that are not reliable for such conditions. The second objective is to obtain more stable power
output from the WTP which inevitably has a variable power output. The technical specifications of WTP and PSHP
models were determined. A general numerical model of the operation simulation is developed specifically in the
scope of the mentioned two objectives. A one-year operation simulation of the HS was carried out under peak tariff
strategy, one-year power output values of the WTP were calculated and this data was employed in the simulation.
Interval time was selected to be one hour. Also another simulation was made, results are obtained from the
operation simulation that was made covering only the WTP with same power output data; annual electrical energy
produced at WTP is 41,800.09 MWh, and with same demand strategy; constant 4 MW for 6 hours a day, in total,
annual electrical energy demand of grid is 8,760.00 MWh.

The following conclusions are drawn from the study. Annual total electrical energy amount delivered to grid for the
WTP without storage is 6,350.30 MWh and this value increases to 7,656.00 MWh for the HS; there is 17.05%
increase in the penetration level. Annual excess electrical energy value is 35,449.79 MWh for WTP without
storage and 34,103.47 MWh for HS; there is 3.80% decrease and annual unmet (deficient) electrical energy
demand of the grid decreases from 2,409.70 MWh to 1,104.00 MWh; there is 54.19% decrease. When the results
relevant to these three variables were compared, the penetration levels increase for the HS and HS is technically
applicable for the first objective.

For the evaluation of the second objective of the system, stable power output ratios are considered. Annual stable
power output ratio of WTP without storage is approximately 72.49% and annual stable power output ratio of HS is
approximately 87.40%; there is 14.91% increase. The increase in the variable indicates that there is an
improvement in stable power output in comparison with WTP without storage. According to these values HS is
technically applicable for the second objective.

In further studies the HSs that are coupled with other types of ESSs may be investigated. Also with the selection of
a solar power system instead of WTP, the HSs may be investigated. Designing a model with stochastic approach
and simulation of this stochastic model is also another subject.

259
Figure 2. 𝑃𝑑 (𝑡)-time diagram in the month of January. Figure 3. 𝑃𝑣 (𝑡)-time diagram in the month of January.

Figure 4. 𝐸(𝑡)-time diagram in the month of January. Figure 5. 𝑃𝑑𝑔 (𝑡)-time diagram in the month of January.

Figure 6. 𝑃𝑤 (𝑡)-time diagram in the month of January. Figure 7. 𝑃ℎ (𝑡)-time diagram in the month of January.

Figure 8. 𝑃𝑝 (𝑡)-time diagram in the month of January. Figure 9. 𝑃𝑒𝑥 (𝑡)-time diagram in the month of January.

Figure 10. 𝑃𝑑𝑒𝑓 (𝑡)-time diagram in the month of January.

260
Figure 11. Simplified flow-chart that represents the operation simulation algorithm of HS.

5. REFERENCES

Paul D. Brown, J.A. Peças Lopes, Manuel A. Matos, 2008. Optimization of Pumped Storage Capacity in an
Isolated Power System with Large Renewable Penetration, IEEE Transactions on Power Systems, vol.23, no.2,
523-531.

C. Bueno, J.A. Carta, 2005. Technical-economic analysis of wind-powered pumped hydrostorage systems. Part I:
model development, Solar Energy, vol.78, iss.3, 382-395.

C. Bueno, J.A. Carta, 2006. Wind powered pumped hydro storage systems, a means of increasing the penetration
of renewable energy in the Canary Islands, Renewable and Sustainable Energy Reviews vol.10, iss.4, 312-340.

Jon Are Suul, Kjetil Uhlen, Tore Undeland, 2008. Wind Power Integration in Isolated Grids enabled by Variable
Speed Pumped Storage Hydropower Plant, IEEE ICSET 2008, 309-404.

Javier García-Gonzáles, Rocío Moraga Ruiz de la Muela, Luz Matres Santos, Alicia Mateo González, 2008.
Stochastic Joint Optimization of Wind Generation and Pumped-Storage Units in an Electricity Market, IEEE
Transactions on Power Systems, vol.23, no.2,460-468.

Feng Gao, Arne Hallam, Chien-Ning Yu, 2009. Wind Generation Scheduling with Pump Storage Unit By
Collocation Method, IEEE PES General Meeting 2009, 1-8.

261
Edgardo.D. Castronuovo, João A. Peças Lopes, 2004 Optimal operation and hydro storage sizing of a wind-hydro
power plant, Electrical Power and Energy Systems, vol.26, iss.10, 771-778.

Phuc Diem NGUYEN NGOC, Thi Thu Ha PHAM, Seddik BACHA, Daniel ROYE, 2009. Optimal operation for a
wind-hydro power plant to participate to ancillary services, IEEE ICIT 2009, 1-5.

Kurtuluş DEĞER, Hybrid Systems of Pumped Storage Hydropower Plants and Wind Turbine Plants, 2013. MSc.
Thesis, Başkent University, Turkey.

Vaughn Nelson, Wind Energy: Renewable Energy and the Environment, 2009. NW: CRC Press.

http://www.inl.gov/wind/software/powercurves/pc_aaer.xls, access date: 2012.

262
 DESIGN OF A MICRO HYBRID TRIGENERATION SYSTEM FOR MEETING ENERGY
DEMANDS OF AN ENERGY EFFICIENT HOUSE USING RENEWABLE ENERGY SOURCES

Marko Mančić, Dragoljub Živković, Milena Todorović

Faculty of Mechanical Engineering, University of Nis in Serbia
Street: Aleksandra Medvedeva 14, 18000 Nis, Serbia
markomancic@yahoo.com

ABSTRACT
A trigeneration plant simultaneously produces heating, cooling energy electricity. The share of electricity
consumption in buildings in Serbia is 48%, 65% of which is used for residential needs. A mini trigeneration
plant is designed based on availability of renewable energy sources, for energy demands of a single
residential building. In this paper modelling and simulation of the annual electricity, heating and cooling
energy loads are simulated for an energy efficient household. The building is modelled as multi-zone, and
the envelope is modelled as well insulated object with mass. The object is connected to the electricity grid.
Based on the analysis of energy loads of the simulated object, the synthesis of micro hybrid trigeneration
plant is done with focus on using renewable energy sources. The hybrid trigeneration plant uses PV/T
collectors, heat storage and the water-water heat pump. Simulation of the system demonstrates that the
formatted system can provide heating and cooling energy demands of the building using only renewable
energy sources, and approximately one half of annual electricity demand.

Key words: trigeneration, Photo Voltaic, heat pump, building simulation, energy system simulation

INTRODUCTION
Over the last thirty years, countries of Western Europe have reduced the consumption of heating energy to
2
the values of less than 100kWh/m per year, which is on average two times less than energy consumption in
housing industry in Serbia. With the "liberalized" electricity market in Serbia, and rising energy prices,
distributed energy production on-site based on renewable energy is attracting more attention (Schneider et
al., 2007). Evaluation of advantages and disadvantages of mitigation to decentralized electricity production
against the general sustainability criteria indicates that decentralized energy can contribute to CO 2 emission
reduction and lower consumption of primary energy carriers (Karger et al., 2009). Models and approaches for
decentralized energy planning can differ worldwide, since development strategies are area-based, shaped to
meet energy needs and develop renewable energy potentials with least cost to the economy and
environment (Hiremath et al., 2007). Labels promoting higher energy performance standards in housing
industry have been promoted in many European countries as market mechanisms, while for less advanced
countries this should stand as an opportunity for lessons to be learned towards developing such labels to suit
market demand and implementation of European Energy Performance of Buildings Directive recast
(Ramanathan et al., 1994). Application for membership in European Union imposes the issue of compliance
with European objectives in this area by reducing energy consumption for 20%, increasing energy production
from renewable sources (RES) to 20%, and reducing CO 2 emission for 20%.
Ministry of Natural Resources, Mining and Spatial Planning of the Republic of Serbia has adopted an Energy
Development Strategy for the period until 2015, which defines a program for the use of new renewable
energy resources as third priority, whereas the first three priority programs represent the basic precondition
for stable economic, energy efficient and environmentally-friendly energy development of Serbia in the future
(Ministry of Mining and Energy et al., 2004). The strategy presents analysis of steady dynamic economic
development plan of the country, with special attention paid to the issue of using renewable energy
resources [Fig 1]. As shown in Fig 1. the renewable energy market in Serbia is developing, and the use of
renewable was negligible prior to 2005, if large scale hydro power plants are excluded from the balance. In
this document, two scenarios are analyzed [Fig 1]: (1) dynamic economic development, i.e. optimistic
scenario where economic capacity in the country will have a sufficient growing trend to support and enable
reaching the preset goals of using renewable; and (2) disturbed economic development, which envisages
minimal economic growth and a pessimistic renewable energy market growth scenario.

263
 2015.

2012.

2009.

2006

2003

0 0,05 0,1 0,15 0,2 0,25

Fig. 1. Estimated use of RES in Serbia: blue - dynamical economic development; red - disturbed economic development

In this paper, a mini hybrid trigeneration system was designed. Photovoltaic (PV) technologies are
recognized as one of the most suitable for decentralized electricity production in urban areas (Marszal et al.,
2011), (Bojic et al., 2006). Adequate technical approach implies economical PV production for maximum
electricity generation by mounting the panels only on the south roof surface of the house (Bojic et al., 2006).
Nevertheless, in this paper potential of electricity production using photovoltaic/thermal (PV/T) collectors is
estimated for panels mounted in south and east vertical facades in addition to the south facing roof. Solar
and low temperature geothermal energy are chosen based on their availability and estimated potential.
Possible electrical and heating energy production using RES is estimated with respect to climate conditions
data for the City of Nis (http://apps1.eere.energy.gov). Houses equipped with heat pumps as main source for
heating are considered to have the lowest energy consumption and the lowest energy intensity (Elsawaf et
al., 2013). Availability of acquired heating energy is increased by a low temperature thermal storage tank,
which is heated by a water-water heat pump. System consisting of PV/T collectors, heat storage tank and
represents a hybrid trigeneration system. The system is integrated to be capable of maintaining the set-point
temperature in the residential area. The aim of the paper is to investigate the behaviour of the designed
trigeneration system, which takes advantage of available RES, and produces heating, cooling and electricity.
Finally, an annual electricity balance is conducted, determining the ratio of on-site electricity production to the
simulated energy demand of the residential object. Since the effectiveness of electricity production of the
investigated trigeneration system depends on integration of PV/T collectors to the building, some remarks
are made for building design pathways for more effective electricity production (Bojic et al., 2006), (Bojic et
al., 2011), (Rosta et al., 2008), (Saman et al., 2013).

SIMULATION OF ENERGY LOADS OF A RESIDENTIAL BUILDING

The investigated trigeneration model designed for the energy demands of a single residential building, with
the net heated area of 160 square meters. The investigated model is a simple two floor house model, with a
basement, and compact shape. The building thermal loads are simulated using TRNSYS 17 software multi-
zone building model. The building is modelled as energy efficient, with an U-value of the building envelope
2
non transparent elements of less than 0.2 w/m k. Building ventilation is natural, and the building takes
advantage of radiant floor heating and radiant ceiling cooling. Aluminium-wood windows with a "thermal
brake" are considered with glazing of low emissivity, three-layer glass with inter-sheet space filled with argon.
The building was divided into three zones: non-heated attic space, non-heated basement and residential
area. Energy balance of heating and cooling loads was calculated for the residential area, with defined
o
interior temperature of 20 C.
Simulation used Energy Plus weather data (http://apps1.eere.energy.gov) for Nis. Since these were
statistical values of hourly temperature changes, design conditions defined more strict criteria for equipment
sizing. Annual Energy load simulation for the building was performed with time step of 1h. Thermal load
change on the model in both heating and cooling regime is presented at Fig. 2.

264
 5
4,5
4
3,5
3
Q_HEAT [kW] 2,5
Q_COOL [kW] 2
1,5
1
0,5
0
1
267
533
799
1065
1331
1597
1863
2129
2395
2661
2927
3193
3459
3725
3991
4257
4523
4789
5055
5321
5587
5853
6119
6385
6651
6917
7183
7449
7715
7981
8247
8513
Fig. 2. Annual hourly thermal load change of the Energy Efficient House model, heating and cooling [kW]

Profile of electricity demand profile of a building depends on characteristics of its envelope, quality of joinery,
behaviour and habits of its residents, but it also depends on weather conditions throughout a year
(http://apps1.eere.energy.gov). The above said, along with the fact that there are neither statistical nor
measured data on electricity demand of a household in Serbia, complicates designing annual household
electrical load model (Jovanovic et al., 2012).
By using energy efficient household electric appliances and applying “Demand side management”, annual
electricity consumption in a household can be reduced by 50% (Flores, 2006). Larsen et al, developed an
econometric model for estimating electricity consumption and found it better than the engineering ERÅD
model (Laresen et al., 2004), but the model is survey based. A model of household electricity consumption
based on statistical data for a Serbian household has been developed by (Vojinovic et al., 2007) for use with
multi-agent based simulator, but the model integration with the chosen software was found impracticable.
Interpretation of electricity use for a typical day of a typical household using electrical consumption
monitoring charts, can be summed for all appliances determining daily consumption (Flores, 2006), however
the trend of change of consumption during the year is neglected in this model.
In this paper, Jukka and Lund mathematical model (Paatero et al., 2006) was used to create hourly electricity
household load profile. The model results with a time dependant electricity load profile, as a sum of the
consumption cycle values of electricity loads of the given household appliance in stand-by and in operating
regime over a given time step. According to this model, parameters of active and stand-by consumption of
an appliance are linked to the annual consumption model according to the following equation Eq.(1):

(1)

Where: is the mean consumption (kWh); is the electric load of an appliance in stand-by,
electric load during a mean consumption cycle (W); time step n duration of a mean
consumption cycle (s); number of time steps of the mean consumption cycle.

265
 450
Electricity
demand [W] 400
350
300
250
200
150
100
50
0
1
294
587
880

1466
1173

1759
2052
2345
2638
2931
3224
3517
3810
4103
4396
4689
4982
5275
5568
5861
6154
6447
6740
7033
7326
7619
7912
8205
8498
Fig. 3. Hourly electricity load profile change: Daily load and annual hourly load

The bottom-up electrify load model results with a time dependant electricity load profile, as a sum of the
consumption cycle values of electricity loads of the given household appliance in stand-by and operating
regime over a given time step. According to this model, parameters of active and stand-by consumption of
an appliance are linked to the annual consumption model (Paatero et al., 2006). The difference in behaviour
profile of electrical load on typical working day and at weekend was considered negligible. Only basic,
energy efficient rated, appliances were included in by the model: TV, electrical stove, CD player, PC,
refrigerator with freezer, microwave oven, washing machine, and dishwasher. The assumed energy efficient
appliances significantly reduces electric load of the household compared to use of standard appliances.
Thermal gain from use of these appliances was included in thermal balance of the building. Trend of change
in daily electricity consumption due to climate changes throughout the year was synchronized with data from
(Jovanovic et al., 2012) by multiplying the time step value of electrical load by a monthly weight factor of the
corresponding time interval. The results of simulation of electric load are lower than the results found in
literature (Jovanovic et al., 2012), (Flores, 2006), (Larsen et al., 2004), (Vojinovic et al., 2007), which is
expected, since the used model does not account for using electricity for space and sanitary hot water
heating. Obtained annual profile of electrical load with time step of 1h, is presented at Fig. 3 and daily profile
of demand. Monthly balance of electricity consumption is represented on Fig. 4.

266
 electricity demand 180
[kWh] 160
140
120
100
80
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12

Fig. 4. Monthly balance of electricity consumption

DESIGN AND SIMULATION OF PERFORMANCES OF TRIGENERATION PLANT

Synthesis, simulation and behaviour analysis of the system for the model energy demands was performed
using TRNSYS software. Trigeneration system produces heating, electrical and cooling energy for which
nowadays various technologies are available. Synthesis of the trigeneration system for the house model
was done with respect to available RES, possibilities to exploit them, performance of specific components
and possible energy flows in the system. Solar and geothermal energy are selected as most adequate for the
model.
Solar energy can be utilized by use of solar collectors for production of heating and electricity. Obtained
thermal energy can be used directly for heating water or cooling by heat powered cooling systems, or for
accumulating the energy by thermal storage. Based on characteristics of plate, stationary, parabolic,
concentric and vacuum solar collectors, conditions of building location, requirements of consumers,
photovoltaic/thermal (PV/T) solar collectors were selected.

267
Fig. 5. Simplified scheme of the solar trigeneration plant designed to meet energy demands of the Energy Efficient House model for Decentralized
Electricity Production

In this paper, the effect of trigeneration system is achieved as follows: electrical and thermal energy are
obtained by application of hybrid solar PV/T collectors, the “waste” heat of which is used for heating water,
while heat pump water-water is used for additional heating and cooling. Presented PV/T collectors are water
cooled PV panels mounted on the absorber of a flat plate solar collector. The controller system is set to
maintain the temperature of PV panel of hybrid collector at a nominal value of the given collector in order to
improve the effect of electricity production. For utilization of low temperature energy sources, heat pumps
which change the temperature level of available heating energy by heat source and adsorption of vapour or
gas sorption can be used (Ziegler, 1999). Application of trigeneration systems based on use of PV/T
collectors has already been analyzed (Calise et al., 2012) in combination with absorption heat pump and
supplementary water heater and concentric PV/T collectors (Matuska et al., 2006). The performance of PV/T
solar collectors is further analyzed in (Calise et al., 2012), (Matuska et al., 2006), (Calise et al., 2012),
(Fraisse et al., 2007). Water temperature at hybrid PV/T collector outlet is usually not sufficient to directly
power the sorption chillers, thus imposing the need for additional heating. In majority of the cases, this can
be achieved by combustion of biomass, natural or generator gas or some other fuel, which increases local
production of CO2, etc. Higher price of sorption cooling devices and the analysis results of necessary cooling
energy of the model for the climate in Serbia, which is by far less than the energy necessary for heating, give
advantage to the use of water-water heat pump. Heating and cooling coefficients of modern water-water heat
pumps, with the application of low temperature heating systems, enable comfort with relatively low electricity
consumption. The proposed system concept reduces the local CO 2 and other green house gas (GHG)
emissions, enables decentralized electricity production, possible surplus electricity export to grid, quiet
operation while maintaining thermal conditions in the residential area of the building.
Based on available temperature levels in the system, it is necessary to use low temperature heating system
0
and heat pump. Underground water temperature used as heat source for heat pump is assumed as 14 C. As
the thermal load analysis showed a small demand for cooling energy, the system is not equipped with
“cooling” energy storage; instead the heat pump directly supplies ceiling panel cooling system. Thermal

268
 3
storage is 2,5m high with the volume of 1m , and it is modelled as a stratified storage tank with 10 nodal
points and fixed inlets and outlets. System scheme is presented at Fig. 5.
Solar PV/T collectors are integrated with the building thermal envelope and grouped into three panels.
2
Panels at southern and eastern vertical plane are 13.4m in surface, while the total dimensions of roof panels
2 o
are 28.3m with the slope of 45 . Simulated electricity production is presented at Fig. 6. Power of heat pump
for heating and cooling is 8.3kW, and nominal electrical power 1.67kW.
A monthly balance of electrical energy produced by application of selected surfaces of PV/T collectors is
presented at Fig. 6. It should be noted that the simulation assumes that the solar subsystem operates and
produces electricity throughout the year.

3,5
S_0_45 [kW]
S_0_90 [kW] 3

E_270_90 [kW] 2,5

1,5

0,5

0
1

3837

4659

5481

6303

7125
275
549
823
1097
1371
1645
1919
2193
2467
2741
3015
3289
3563

4111
4385

4933
5207

5755
6029

6577
6851

7399
7673
7947
8221
8495
S_O_90 [kWh] 800
S_0_45 [kW] 700
E_270_90 [kWh]
600

500

400

300

200

100

0
1 2 3 4 5 6 7 8 9 10 11 12
Fig. 6. Annual electricity production: Hourly production in [kW] and monthly production in [kWh]: S_0_45 -Solar collectors facing south, slope 45o;
S_0_90- Solar collectors turned facing south, slope 90o; E_270_90 - Solar collectors facing east, slope 90o

Basic function of designed system is to enable adequate thermal conditions in residential area. The system
is set to ensure the interior temperature of 20˚C in heating season, and maximum temperature in cooling
o
season is 26 C. Engaged power of heat pump is presented at Fig. 7, annual temperature change in the
residential area of the building hose model is shown on Fig. 8, and monthly balance of electricity used to fire
the heat pump is on Fig. 9. The heat pump controller is set to monitor the temperature in the residential area,
forcing heat pump to provide the given set-point temperature in the residential building zone constantly. The
consequence of this control strategy is frequent change in cooling and heating regimes in spring and autumn
transition periods as seen in Fig 7 and 8. Significant energy savings in this period can be made if with
equipment turned off with no strict control of indoor temperature conditions, which is a standard practice.

269
 1,4
Heat pump power 1,2
cooling [kW]
1

Heat pump power 0,8

heating [kW] 0,6

0,4

0,2

0
1

3613

5161
259
517
775
1033
1291
1549
1807
2065
2323
2581
2839
3097
3355

3871
4129
4387
4645
4903

5419
5677
5935
6193
6451
6709
6967
7225
7483
7741
7999
8257
8515
Fig. 7. Electrical power engaged by the heat pump for heating and cooling

30
25
20
15
10
5
0
1

1387
1618

7624
7855
232
463
694
925
1156

1849
2080
2311
2542
2773
3004
3235
3466
3697
3928
4159
4390
4621
4852
5083
5314
5545
5776
6007
6238
6469
6700
6931
7162
7393

8086
8317
8548
Fig. 8. Temperature in the residential object area
1000

POWER HEAT [kWh] 900

800
700
POWER COOL[kWh]
600
500
400
300
200
100
0
1 2 3 4 5 6 7 8 9 10 11 12
Fig. 9. Monthly balance of electricity consumed by the heat pump

Although the outlet temperature of the PV/T collectors is not sufficient for direct heating, and has a significant
temperature drop in the winter, implying the use of ethylene glycol as a het transport medium, the obtained
simulation temperatures in the summertime indicate a possibility of using the PV/T outlet as a heat source for
the heat pump cooling. Annual balance was performed based on simulation data, which shows if the total
produced electrical energy is sufficient to power the designed HVAC system and electricity consumption in
household. Final annual balance of electrical energy is presented at Fig. 11.

270
 40

P3 [kWh] 35

P2 [kWh] 30

P1 [kWh] 25

20

15

10

0
1 2 3 4 5 6 7 8 9 10 11 12

Fig. 10. Monthly electricity consumption of the supply pump groups

Fig. 11 shows that the largest electrical consumer of the analyzed model is the heat pump. Total annual
electricity production of the proposed model is 6143kWh, 66,55% of which is produced by roof panels, 22,
26% is produced by the southern facade panels and 11,15% by the eastern facade panels. Average annual
electricity production per square meter of PV/T collector is 146,023kWh for the roof panels, 102,166 of the
south facade panels, and 51,14 for the east facade panels. Total annual electricity consume of the model is
estimated to 12077,05 kWh, where the heat pump consumes 6617,99 kWh for heating, and 3837,9 kWh for
cooling. Hence, the house model produces around 51% of the energy it consumes.

Houhold electricity [kWh]

Total Heat pump [kWh]

Total Pumps [kWh]

Total S_0_90 [kWh]

Total S_0_45 [kWh]

Total E_270_90 [kWh]

0 2000 4000 6000 8000 10000 12000

Fig. 11. Annual balance of electricity generated by solar collector planes and electricity consumed by consumer groups: Total E_270_90–
Electricity generated by the solar collectors turned to the east; Total S_0_45 – Electricity generated by the solar collectors placed on roof; Total
S_0_90 - Electricity generated by the solar collectors turned to the south; Total Pumps – Total electricity consumed by the trigeneration plant
supply pump groups ; Total Heat Pump – Electricity consumed by heat pump; Household Electricity consumption – Total electricity
consumption of the household appliances.

271
 Annual Electricity consumption [kWh]
Annual Electricity Production [kWh]

0 2000 4000 6000 8000 10000 12000 14000

Fig. 12. Annual balance of production and consumption of electricity for the Energy Efficient House model

As presented in Fig.12 electrical energy produced by solar collectors is not sufficient for meeting the
estimated total electricity demand of the household. The deficit of electricity consumed over the electricity
produced by PV/T collectors implies that the model should be modified, to increase electricity production.
Increased electricity production using solar collectors can be achieved in two ways: by increasing the surface
of installed collectors and by changing the slope of vertical collectors. Based on literature (Bojic et al., 2006)
and simulation data, change of the slope angle of vertical collectors on the southern side is more
economical, while the installation of collectors on the eastern side is questioned due to their low performance
with given orientation and high investment costs.

DISSCUSSION
The simulation of heating and cooling demands of the given residential building model was relatively low as
anticipated. According to the simulation, the chosen heat pump was able to maintain the temperature in the
residential area of the building within the pre determined range. The heat pump operated at the regime which
was below its maximum power throughout the year. This is expected, since the energy system was sized
according to the sizing temperature, while the simulation used average hourly weather data. The heat pump
was set to operate continuously and keep the temperature in the residential area within the pre set range.
Hence, as shown by the electricity annual balance, the heat pump is the greatest electricity consumer in the
proposed model.
The significant difference in electricity consumed and electricity produced by PV/T collectors implies that the
model should be modified to research feasibility of increased on-site electricity production. In this case, this
means increasing electrical energy production by application of solar collectors. This can be achieved in two
ways: by increasing the surface of installed collectors and by changing their orientation towards the sun.
Changing the slope of vertically installed collectors is the first step towards this goal. Based on simulation
data, change of the slope angle of vertical collectors on the southern side is more economical, while the
installation of collectors on the eastern side is questioned due to their low performance with given orientation
and high investment costs.
Change of slope angle of PV/T collectors on the southern facade can be performed in two ways: by changing
the slope of the façade wall or by installing special structure to enable more favourable slope of collectors on
the southern side. Since the construction of wall with slope which follows the roof plane is optimal, from the
aspect of utilization of solar energy, but regarding aesthetics and architecture this is neither functional nor
acceptable, a combination of the two mentioned solutions is suggested. Facade wall can be constructed at
o
angle of 15 , and it is possible to install bearing structure for solar modules onto it with the slope of 30% from
the facade. Thus, desired slope angle of collector is achieved, investment costs are reduced and
performance of electricity production by only two fields of solar collectors is obtained. In this case, the
electricity production on southern facade wall would be equal to specific electricity production per square of
roof field of PV/T solar collectors, and it would amount around 144 kWh per year, according to the results of
the simulation. Instead of instalment of non-efficient solar collector field on the eastern facade, instalment of
PV collectors above parking space, garage is suggested. Furthermore, change of base geometry should also
be considered, i.e. designing larger facade surface orientated towards south and installing solar collectors by
defined principle. Application of solar tracking system was not considered in this study. The results show, that
by orienting the same total collector area towards south, and by slope adjustment, the vertical facade
collectors would be able to produce 3912.8 kWh instead of 2054.83kWh, whereas total electricity production
could raise to 8450.56 kWh. This means an increase of 90% of the effectiveness of the same area of solar
collectors, and increased total production 1.37 times. However, even with these modifications, the proposed
simple house model is not able to produce all the energy it demands.

272
CONCLUSION
The renewable energy market is still not developed in Serbia. Rulebook for Building Energy Efficiency and
procedures and conditions for determining Building Energy codes in Serbia was adopted in September 2012,
however it does not promote on-site energy production. The problem of saving and utilization of renewable
energy resources is very serious, primarily due to the lack of funds for investments (Bojic et al., 2006) into
systems that would help solve this problem.
Heating and cooling demand, were met by the designed trigeneration system, but the simulation resulted
with a deficit of electricity produced on-site suggesting design modifications for better solar energy
harvesting. A square meter of a south oriented roof collector is three times higher than that of an east facade
collector and 50% higher than that of south facade collector. The model modification for more effective solar
energy harvesting represent an architectural design challenge and should be aimed for increasing south
oriented roof, increasing south facade area under and slope adjustment of the south facade solar collectors.
If we take into consideration the purchase price of electrical energy produced by application of solar system
in Serbia which is €0.23 per delivered kWh, and very low purchase price of electrical energy per kWh from
the supply network of 0.045-0.055 €/kWh on average, financial evaluation of the proposed model may prove
positive. However, detailed analysis of investment costs (Bojic et al., 2006), (Bojic et al., 2011) and
profitability should be performed before coming to a conclusion.
.

NOMENCLATURE

REFERENCES
Schneider, D., N. Duic, I. Raguzin, Z. Bogdan, M. Ban, B. Grubor, P. Stefanovic, D. Dakic, B. Repic, Z.
Stevanović, A. Zbogar, M. Studovic, S. Nemoda, N. Oka, D. Djurovic, N. Kadic, V. Bakic, S. Belosevic, A.
Eric, R. Mladenovic, M. Paprika, N. Delalic, A. Lekic, R. Bajramovic, A. Teskeredzic, I. Smajevic, E.
Dzaferovic, F. Begic, H. Lulic, S. Metovic, S. Petrovic, A. Djugum, Dz. Kadric, N. Hodzic, F. Kulic, A. Kazagic
and A. Gafic. 2007. Mapping the potential for decentralized energy Generation based on res in Western
Balkans. Thermal Science 11(3):7-26.
Karger, C. and W. Hennings. 2009. Sustainability evaluation of decentralized electricity generation.
Renewable and Sustainable Energy Reviews 13:583-593.
Hiremath, R., S. Shikha and N. Ravindranath. 2007. Decentralized energy planning; modeling and
application-a review. Renewable and Sustainable Reviews 11:pp 729-752
Ramanathan, R. and L. Ganesh. 1994. A multi-objective evaluation of decentralized electricity generation
options available to urban households. Energy Conversion Management 38(8):661-660.
Ministry of Minings and Energy, Government of the Republic of Serbia. 2004. The Development Strategy of
Serbia until 2015. Belgrade.
Marszal, A., P. Heiselberg, J. Bourrelle, E. Musall, K. Voss, I. Sartori and A. Napolitano. 2011. Zero Energy
Building – A review of definitions and calculation methodologies. Energy and Buildings 43:971–979.
Bojic, M. and M. Blagojevic. 2006. Photovoltaic electricity production of a grid connected urban house in
Serbia. Energy Policy 34:2941–2948.
Website: http://apps1.eere.energy.gov/buildings/energyplus/weatherdata_about.cfm 28.01.2012
Elsawaf, N., T. Abdel-Salam and H. Abaza. 2013. Economic evaluation and calculations of energy savings by
upgrading the heating systems in pre manufactured homes. Energy and Buildings 59:187–193.
Bojic, M., N. Nikolic, D. Nikolic, J. Skerlic and I. Miletic. 2011. Toward a positive-net-energy residential
building in Serbian conditions. Applied Energy 88:2407–2419.
Rosta, S., R. Hurt and R. Boehm. 2008. Performance of a zero-energy house. Journal of solar energy
Engineering 130:021006-1.
Saman, W. 2013. Towards zero energy homes down under. Renewable Energy 49:211-215.
Jovanovic, S., Z. Djordjevic, M. Bojic, S. Savic and B. Stepanovic. 2012. Weather conditions impact on
electricity consumption. Proceedings Conference on Mechanical Engineering Technologies and Applications,
COMETA 2012.
Flores, A. 2006. Study of an ordinary home electrical consumption in order to make it more energy-efficient.
Renewable Energy & Power Quality Journal 4:426-434.
Larsen, B. and R. Nesbakken. 2004. Household Electricity end-use consumption: results from econometric
and engineering models. Energy Economics 26:179-200.
Vojinovic, O., I. Milentijevic, V. Ciric and A. Vukašinovic. 2007. Household Model for multi-agent electricity
consumption simulator. Proceedings of 13th YuInfo Conference, Kopaonik, Serbia.

273
Paatero, J. and P. Lund. 2006. A model for generating houshold electricity load profiles. International Journal
of Energy Research 30:273-290.
Ziegler, F. 1999. Recent developments and future prospects of sorption heat pump systems. International
Journal of Thermal Science 38:191-208.
Calise, F., M. Dentice d’Accadia and L. Vanoli. 2012. Design and dynamic simulation of a novel solar
trigeneration system based on hybrid photovoltaic/thermal collectors (PVT). Energy Conversion and
Management 60:214–225.
Matuska, T. and B. Sourek B. 2006. Facade solar collectors. Solar Energy 80:1443–1452.
Calise, F., M. Dentice d’Accadia, A. Palombo and L. Vanoli. 2012. Dynamic simulation of a novel high-
temperature solar trigeneration system based on concentrating photovoltaic/thermal collectors. Energy (in
press), available online 01.11.2012
Fraisse, G., C. Menezo and K. Johannes. 2007. Energy performance of water hybrid PV/T collectors applied
to combisystems of Direct Solar Floor type. Solar Energy 81:1426–1438.

Acknowledgement
This paper was done within the research framework of project III 42006 - “Research and development of
energy and environmentally highly effective polygeneration systems based on renewable energy sources”,
supported and financed by the Republic of Serbia Ministry of Education, Science and Technological
Development.

274
 ASPHALTENE AS LIGHT HARVESTING MATERIAL IN DYE-SENSITIZED SOLAR CELL
* **
R.Abujnaj and R.R. Chianelli
*Earth and Environmental Science Department, College of Science Elmergib University Alkhoms Libya
Email. reabujnah@mail.elmergib.edu.ly
**Materials Research and Technology Institute University of Texas at El Paso USA
chianell@utep.edu

ABSTRACT
In this study asphaltene, waste hydrocarbons and problematic constituent present in heavy oil have been
investigated for its use in dye-sensitized solar cells (DSSCs). It presents the first asphaltene (DSSCs), in
which light to electricity conversion efficiency has improved from 0 % to 1.8 %. Four natural asphaltene
portions were extracted and used as light absorbers in TiO2 based dye-sensitized solar cell. The photovoltaic
performance of the cells were analyzed by determining the various I-V cell parameters such as open-circuit
voltage, short-circuit current, fill factor, and series resistance. The overall energy conversion efficiency was
also measured to correlate the effect of the different asphaltene portions as well as concentration on the
significant improvement of the solar cell parameters. The sensitization of TiO 2 electrode with H-E Asph.Fr3
asphaltene portion has resulted in a maximum energy conversion efficiency of 1.8 % of the solar cell. In
2
particular, a maximum photocurrent density and solar conversion efficiency of 16.9 mA/cm and 1.8 % have
been obtained, respectively. A strategy to minimize, series resistance, and improve photo current as well as
open circuit voltage is also studied. Purifying asphaltene, using RTV mask, and employing the TiO2 blocking
layer as well as UV-O3 treatment have shown to improve the performance of asphaltene DSSCs.

INTRODUCTION
A photovoltaic solar cell is a proven technology to capture the solar energy and provides clean and
renewable electrical energy that can reduce the world’s dependency on fossil fuels, and reduce Green
House Gas (GHG) emission, global warming and other related environmental issues. In spite of the
substantial growth over the past decades the high cost of photovoltaic solar cells has remained a limiting
factor for the implementation of the solar electricity in a large scale. Currently, the dominant photovoltaic
technology is based on solid-state pn junction devices, in which semiconductor absorbers produce electron-
hole pairs and the electron-hole pairs are separated by a build-in electrical field in the pn junction to generate
electricity. These types of solar cells have high power conversion efficiency; however, suffer from high
manufacturing and material cost.
Among the material systems currently of great interest for this reason are the group of molecular-based
materials combinations often referred to as ‘organic’ or ‘molecular’ photovoltaic materials. These refer to
conjugated molecular species, such as polymers, molecules and dyes, which are capable of absorbing light
and conducting charge and thereby acting as organic semiconductors (Troshin, 2008,). Their attraction lies
primarily in the possibility of processing such materials directly from solution, and so with bulk synthesis of
the chemical materials and bulk solution processing of photovoltaic modules, the cost of the photoactive
material could fall by an order of magnitude compared even with thin film inorganic semiconductors
(Shaheen, 2005), (Kalowekamo, 2009). Additionally, the less challenging manufacturing environment,
compared, for example, with crystalline silicon wafer production, promises to reduce the capital cost for
production facilities and to make the technology more widely accessible, especially in developing countries
(Brabec., 2004). Moreover it is also believed that if such materials are naturally synthesized and stable the
cost could be further reduced.
Dye sensitized solar cell (DSSC) is one of the low cost alternatives for the conventional pn junction
based solar cells, and commercially promising because it can be made from low-cost materials and do not
require elaborated manufacturing facilities. At the moment there are three types of dyes (Gratzel M., 2003)
that can produce DSSCs with AM1.5 conversion efficiencies over 10%. However, these types of dye suffer
from the drawback that they are based on the rare ruthenium transition metal.
Asphaltene is one of these materials that for sure fulfill the requirements of organic semiconductors; i.e.
absorbing a broad range of visible and near infrared light (Ruiz-Morales Y., 2007) (Salmanova, 2007)
capable of conducting charge (Jiqian Wang, 2009), stable and naturally abundant. Asphaltene is an organic
moiety that contains large number of structures, in particular high molecular weight fused aromatic
hydrocarbons components with hetero-atoms (Mullins, 2007). It also contains some organic molecules that

275
already used in DSSC and show a promising performance, such as Porphyrin. Most of the researchers have
found that the sign of asphaltene charge is positive in organic solvent such as heptane, toluene, ethanol, and
nitro methane as studied by zetametry (Siffert, Kuczinski, & Papirer, 1990), and electrophoresis (Neves,
Lucas, & Gonz´alez, 2001). The origin of the electric charges is a consequence of an electron transfer
between the organic solid particles and the liquid organic phase (Jiqian Wang, 2009).
Asphaltene therefore is an excellent candidate for use in DSSC. Despite of the long term research on
this material, there was no any study directed to its potential use in such purpose. This article introduces the
first real data on using asphaltene portions as sensitizer in Dye sensitized solar cell and describes ways of
improvements.
MATERIALS AND METHODS
Materials
Commercial chemical agents and materials were used to prepare typical DSCs, including TiO2 pastes,
electrolytes and the transparent conductive glass. The FTO are (F-doped SnO2, FTO, (Sloarnoix TCO30-8
SA Switzerland 8 Ω/cm2/&,3.3 mm thickness and 500- 1000 nm transmittance). The TiO2 paste was a highly
transparent Ti-Nanoxide HT/SP and Ti-Nanoxide R/SP obtained from Solaronix Co. Switzerland. The paste
contains approximately 18 % wt. of nano-crystalline titanium dioxide with 8-10 nm particle size, with terpineol
and other organic binders. A layer of highly porous anatase nanocrystals is obtained after firing at 450 °C for
30 minutes. The obtained sintered layer is highly transparent. Three types of commercial electrolytes
obtained from Solaronix company have been used in this study. They are Standard Iodolyte MPN-100. It
consists of 100 mM of tri-iodide in methoxypropionitrile solution. The second one is Iodolyte TG-50. It is
Iodide Based Redox Electrolyte consists of 50 mM of tri-iodide in tetraglyme.
Asphaltene extraction and purification
The crude oil sample was separated into maltenes (n-heptane soluble) and asphaltene (n-heptane
insoluble) by adding n-heptane in a volume-volume ratio of n-heptane to crude oil of 40: 1 followed by
filtration. A 50 ml of crude was mixed with 2000 ml of n-heptane in a beaker, the mixture then stirred by
magnetic tip overnight at room temperature. The used solvent-to-oil ratio of 40:1 has been demonstrated to
be suitable for avoiding errors in the determination of the amount of asphaltene fraction and in its
characterization (Ancheyta, 2002). The mixture was filtered using a filter paper No 40 with a pore diameter of
8 µm. The obtained asphaltene contains parts of oil that does not dissolve in heptanes such as clay, sand
etc. for this reason it was purified by dissolving it completely in toluene at room temperature and then filtered
again to remove debris. The obtained solution of asphaltene and toluene was covered and kept under
vacuum at room temperature for 24 hours to dry out. The precipitated asphaltene then weighted for
determining asphaltene content, marked as crude asphaltene and will be referred to as un- fractionated
asphaltene.
Fractionation of asphaltene
A sample of 3 g asphaltene was dissolved in 100 mL toluene, and 300 mL n-pentane was added into the
solution and mixing by stirring with a magnetic bar at ambient temperature. After adding n-pentane, the ratio
of the total volume of the n-pentane to toluene is kept at 75/25 in volume. The solution was covered and
stirred for half an hour for precipitation of insoluble. After filtration, the insoluble fraction was collected and
dried at room temperature and left under vacuum overnight to obtain the first fraction (First precipitate. Fr.1).
another amount of 270 ml pentane then added to the filtrate n-pentane and toluene to keep the ratio of the
total volume of the n-pentane to toluene at 85/15 in volume.
The filtration and drying process then repeated to obtain the other fractions. Fr.2 was precipitated in a
solution with an n-pentane to toluene ratio of 85/15 in volume. Similarly, samples (Third precipitate. Fr.3) (n-
pentane to toluene = 90/10), were precipitated in binary solutions with the increasing proportion of n-
pentane. The last fraction, (fourth precipitate Fr. 4), was obtained by evaporating the solvents with the
remaining asphaltene.
Preparation of the photo-electrode
The blocking layer was prepared in the following manner. The FTO glass was first cleaned in a detergent
solution using an ultrasonic bath for 15 min, and then rinsed with water and ethanol, and treated with UV-O3
cleaner for 20 minutes after that the FTO glass plates were immersed into a 50 mM aqueous solution of TiCl 4
at 70 C for 30 min and washed with water and ethanol. The TiCl4 solution was formed by mixing 0.54 ml
TiCl4 in 100 ml of ultra-water in an ice path.
After applying the blocking layer, the photo electrodes then rinsed with d-water and ethanol and are
ready for meso-layer application. The nanocrystalline TiO2 layer was prepared by applying two layers of Ti-
Nanoxide HT/SP paste by spreading (doctor blading). In the processes the FTO-coated glass was covered

276
with two layers of parallel adhesive Scotch tape 1 cm apart to control the area and the thickness of the TiO2
film (2,5 x 1 cm) see figure 1. Tape is used to hold the slides as well as to create the spacing required to
produce the nanoparticle film. A solution of Ti-Nanoxide HT/SP paste is then applied to the top slide. A glass
road is then used to draw the solution down to the slide of interest in a smooth and quick motion to coat the
working electrode with nanoparticles. The film is then left to dry and the tape is removed. Films are treated at
500 ˚C for half an hour before soaking in a dye solution. This treatment process is used to remove any
organic additives used as well as to sinter the nanoparticles together and remove any water from the active
region of the cell. It is important to note that the samples should be hot (70˚C) when immersed in a dye
solution to prevent H2O adsorption. The samples were left overnight in the dark for dye adsorption. The film
is then cut into the desired cell area using a glass cutter and the titania film is scratched into by 5 mm
squares using a razorblade figure 1.
Preparation of the counter electrode
Counter electrodes were fabricated by first drilling two holes using a diamond tip drill through an FTO
piece of dimensions 2.5 cm by 5 cm. Drops of water were applied during the drilling process to prevent
fracture of the glass. Once the electrodes have holes in them, they are washed with a d-water and detergent
solution using an ultrasonic bath for 20 minutes to remove any organic residues, then washed in Ethanol
solution for 20 minutes and Acetone for 20 minutes and then cleaned with UV-O3 cleaner for 20 minutes just
before Pt application. Before being used in a cell, counter electrodes are coated with platinum using a 5 mM
solution of Hexachloroplatinic acid in Ethanol. A few drops are placed on the counter electrode and left to dry
for half an hour. The counter electrode is then heated at 450˚C in a furnace for half an hour. Upon cooling,
the counter electrode is ready to be used in a working cell.
Assembling the cell
Surlyn spacers are cut from a large sheet of Surlyn sandwiched between two grid paper sheets, into the
desired cell active area of 10 mm by 15 mm. After soaking the working electrode in dye solution overnight,
the cells were assembled. To begin, the working electrodes are removed from the dye solution and washed
with ethanol for 15 seconds waving mildly with tweezers. The electrode is then left to dry in a vertical
position. The Surlyn gasket is placed around the titania square and the counter electrode is then placed
above the working electrode sandwiching the Surlyn gasket and placed under the hot press for 30 seconds.
Once the Surlyn is melted (easily seen visually), the cell is left to cool in the dark. The next phase of cell
assembly is to fill the solar cell with electrolyte and seal the cell. To do this, a Solaronix' Vac'n'Fill Syringe
was used to fill the electrolyte and Polyimide Film Tape with Acrylic Adhesive was used to cover the holes of
the counter electrode see figure 1.
RESULTS AND DISCUSSION
Introducing TiO2 blocking Layer
The importance of a blocking layer in organic dye-sensitized solar cells has been demonstrated in many
studies (Ito. S., 2005), (Gan, 2007), (Thelakkat, 2002), (Okuya. M., 2002). It was shown that upon the
addition of a compact layer, the light-harvesting efficiencies were more than tripled. Such a compact layer
improves the adhesion of the TiO2 to the TCO and provides a larger TiO2/TCO contact area and more
effective electron transfer from the TiO2 to the TCO by preventing the electron recombination process, which
occurs in the interface between the redox electrolyte and the TCO surface. The compact layer can be
fabricated by various methods such as sputter deposition (Ito. S., 2005), dip-coating (Gan, 2007), chemical
vapor deposition (Thelakkat, 2002), and spray pyrolysis (Okuya. M., 2002). Especially, the fabrication of the
TiO2 compact layer using a TiCl4 aqueous solution has been widely adopted (Ito. S., 2005) and is the
method used in this research.
To test the effect of the compact TiO2 layer on the performance of asphaltene solar cells we made two
asphaltene solar cells both with three TiO2 layers (one compact and the other is Nano-crystalline layer). The
used paste in these cells was commercial highly transparent Ti-Nanoxide HT/SP paste stated above, and the
TiO2 compact layer was prepared using aTiCl4 aqueous solution. The asphaltene was un-fractionated
unpurified Hunt asphaltene obtained from Altimera crude oil with utilized concentration of 0.5 g/l. The cells
were tested under solar simulator at standard condition of 1.5 sun and the results are shown in table.1. It is
clear that, the introduction of TiO2 compact layer prepared by the above mentioned method has increased
the short circuit current, the close circuit voltage as well as overall cell efficiency.

277
 Preparation of the working electrode

Sintering at 500 Dye absorption Cell dimension

Preparation of the Counter electrode

Holes drilling Pt application Air drying Heat at 450

Cell building

Spacer application Electrodes sandwiching Electrolyte submission Contact application

Figure: 1 processes of manufacturing DSSC

278
Table: 1 The effect of TiO2 compact layer on asphaltene cell performance

Cell id Voc (v) Jsc (mA/cm2) FF % IPCE % Voc Slope (ohms)

Without compact 0.4784 0.158 35.42 0.03 1830

With compact 0.420 0.388 46.46 0.07 398

Purifying asphaltene and using RTV silicone rubber mask layer on the photoelectrodes

A typical DSSC usually contains three interfaces formed by FTO/TiO 2, TiO2/dye/electrolyte, and
electrolyte/Pt-FTO. Simultaneously, electrons are transferred to tri-iodide at the TiO2/dye/electrolyte interface
and iodide is reduced to triiodide at the counter electrode. Series resistance in DSSC could arise from one of
the above interfaces and have a great impact on cell performance. The values of series resistance can be
raised by varying the electrical properties of TCO, increasing the thickness of the electrolyte layer and the
catalytic activity of the counter electrode. Many researchers (Hoshikawa, 2005), (Anneke. H, 2001) (Weiqing.
L., 2010) have found that, under short circuit conditions, electron transport was predominately affected by the
series resistance in TCO-dye interface, and the electrolyte/Pt-TCO interface.
The extraction of asphaltene from crude oil involves dissolving the crude oil in n-heptanes followed by
filtration. Asphaltene is the part of crude oil that does not dissolve and stay on top of the filter paper.
Asphaltene extracted by this method could contain inorganic impurities which may increase resistivity of the
asphaltene films when applied on top of TiO 2. Hence, it may have an effect on charge transfer and increases
the series resistance at TiO2/asphaltene/electrolyte interface. For this purpose we employed the purification
method described above to take off these impurities, before using it in DSSC.
In addition to that, when applying the compact layer in our previous experiments, the FTO glass was
immersed completely in a TiCl4 aqueous solution. Subsequently, TiO2 compact layer was formed on all the
TCO coating of the FTO glass used to prepare the photoelctrode. This has shown to increase the resistivity
of the TCO layer, for that reason we used RTV silicone rubber paste as mask and cover the backside of
glass completely and leave less than half side of conducting surface uncovered, which will be used to
deposit the paste. Following these two strategies the efficiency of our cells has increased significantly when
compared to previous cells from 0.07 see table 1 above to 0.25 see table 2 below.

Table: 2 Purified Un-fractionated Hunt asphaltene cell parameters with 0.5 gm/ l
2 2
Iv parameter 80 mW/cm2 100 mW/cm 120 mW/cm

Voc (V) 0.54 0.55 0.54

Jsc (mA/cm2) 0.59 0.72 0.91

Fill factor % 65.4 63.2 60.6

Voc slope (ohms) 43.6 42.6 38.7

Efficiency (%) 0.26 0.25 0.25

It is clear that the enhancement of cell efficiency in this cell was a consequence of decreasing series
resistance and increasing close circuit current which leads to increase fill factor. This is in agreement of other
people findings, who found an increase of FF with decrease in series resistance (Naoki K., 2006) . To study
the effect of varying light intensity on cell performance, the cell was tested under three different intensity
levels 80, 100, and 120 mW/cm2. As was expected higher light intensity produces higher current density but
less fill factor and therefore lower cell efficiency. This observation is a well-known phenomenon in DSS cells
(Kun-Mu Leea, 2009 ). This is probably due to increase recommendation processes. The existence of the
recombination process can be verified in the dark current measurement column five in the above table.
UV-Ozone treatment of the photo-electrode
The main components of the TiO2 pastes utilized in this study as well as in most DSSC research labs
more than TiO2 particles are organic compounds used to disperse and link the semiconductor particles to
form a network film. At high temperature sintering processes these compounds decompose and permit the
formation of porous TiO2 films with high surface area. However, some carbon atoms will probably deposit on

279
top of TiO2 particles and hinder the absorption of the dye, and the electron injection. UV–O3 treatment is well
known method to remove organics on transparent conducting oxides and has been widely used in organic
electronic devices to clean the surface and modify the work function of ITO. Recently UV-O3 treatment was
applied to synthesize porous nano-particulateTiO2 films at room temperature and remarkable improvement of
photo-conversion efficiency was obtained using the treatment. It is believed that the main effect of the UV–O3
treatment was reported to be the removal of residual organics and positive shift in the conduction band of the
nano-crystalline titanium dioxide, which promotes electron injection from the dye (Byoung-Kuk Lee, 2009)
(Dongshe Zhang, 2006).
To see if UV–O3 treatment has an effect on the performance of asphaltene solar cells, UV-O3 treatment
was first applied on the last TiO2 layer. Two asphaltene cells were made this time one with the commercial
paste and the other is homemade paste described above. Both cells have given much better Voc and
photocurrent when compared to the previous cells without UV-O3 treatments see table 3

Table: 3 the effect of UV-O treatment on top TiO2 layer

2
Cell id Voc (V) Jsc m A/cm FF % IPCE %
Ti HT/SP + R 0.806 8.51 15 1.03
H P25 + R 0.809 6.77 14.7 0.8

After, our UV–O3 treatment on top layer of the photo-electrode showed improvement in cell operation,
we apply UV-O3 in three processing steps; on bare FTO before the compact layer, before and after the
compact layer and after the mesolayer. The used asphaltene here was fr.2 asphaltene with applying
concentration of 1mg/ml. Due to difficulty of applying the previous sealing material, the sealant has changed
to Polypropylene w/ Rubber Medium Adhesive tape with 228.6 micrometers thick obtained from Grainger
company item no 6JT54, the adhesive was not strong enough so some drops of crazy glue was put in on top
of the tape between the electrodes. So the thickness was further increased a few micrometers. The results
are shown in table 4.
The efficiency of these cells has increased four times when compared with the ones without UV–O3
treatment. The improvement in efficiency was a result of a big boost in both short circuit current and open
circuit voltage. This enhancement was a result of increase in asphaltene absorption, eliminating organic
contamination and therefore, higher photon capture and higher electron injection. However these cells shows
very low fill factor. The decrease in the fill factor is due to increasing the electrolyte thickness as a result of
increasing the sealant layer. The electrolyte ions will take longer time to reach the counter electrode before it
reduced back. Similar results were observed with increasing the electrolyte thickness.

Table: 4 The effect of ozone treatment on the asphaltene DSSC

IPCE (%) Voc (V) Jsc (mA/cm^2) FF (%) FTO UVO CL UVO ML UVO
1.44 0.8252 13.19 13.1 X X
1.34 0.7818 13.55 12.9 X X
0.76 0.7790 7.74 12.2 X X X

The low photo current in asphaltene solar cell when UV-O3 treatment was done on all the three TiO2
layers was probably due to electrolyte touching the counter electrode because of improper sealing. So
changing the sealant is vital at this point.
Varying asphaltene fraction
The open circuit voltage and short circuit current are both depend on the extent of light absorption. The
most straight forward way to increase Jsc is to absorb a greater fraction of the incident light. The optical gap
of the Ru dye in the most efficient DSSC to date is 1.8 eV, allowing it to absorb essentially all the light out to
700 nm. Increasing the photocurrent density requires decreasing the optical gap to extend the dye’s
absorption into the near-infrared. Our first thought of increasing short circuit current is to try different
asphaltene fraction. Fractionation of asphaltene produces slightly different absorption spectra (see figure 1)

280
 0,18

0,16

0,14
Unfrac.
0,12 Fr. 3
Absorbance

Fr. 1
0,1 Fr.2
0,08 Fr. 4

0,06

0,04

0,02

0
350 400 450 500 550 600 650 700 750 800
Wavelength nm

Figure: 1 Absorption spectra of asphaltene thin films

Neglecting the fill factor at the moment, the cells made of Fraction four which is the most toluene soluble
fraction produces the less photocurrent and photo voltage when compared to the other asphaltene fractions.
The less soluble asphaltene part (fraction one) generates higher photo voltage but lower photocurrent then
fr.2, (see table 5) fr.3, and un-fractionated asphaltene.

Table:5 Asphaltene fraction cell parameters

IPCE (%) Voc (V) Jsc (mA/cm^2) FF (%) Frac #

1.80 0.7793 13.69 16.6 3
1.38 0.7098 4.76 40.1 4
1.02 0.7724 10.52 13.0 1
1.76 0.7744 13.45 13.0 Unfrac.

These results are in agreement with the absorption spectra of asphaltene thin films presented in figure
(1). Both Fr.1 and Fr.4 have lower absorption in the visible part of light and therefore, will have lower
photocurrent. The high fill factor in Fr.4 cell is a result of changing the sealant to lower thickness and easy to
process; fastelfilm 230110 with 130 micrometer thick film, obtained from Fastel Adhesive.
CONCLUSION
The essence of this research work is well encapsulated by the title of this work, namely, “Asphaltene as
light harvesting material in dye sensitized solar cell”. Despite the fact that there were considerable doubts
about the suitability of these materials as sanitizer in DSSC at the start of this work, the research presented
in this article has comes a long way in removing many of these reservations. It has been proved that for at
least one part of the asphaltene or asphaltene in whole can be effectively utilized. Asphaltene indeed is an
excellent sanitizer in DSSC. Asphaltene the most abundant naturally forming material and undesirable part of
crude oil that causes a lot of trouble to most oil refining companies has been shown in this study to be good
sensitizer in DSSC. The results obtained show that every step in manufacturing the asphaltene cell has an
effect on different cell parameters. According to results shown above, asphaltene solar cells have achieved a
fill factor of 65% -70 %, when series resistance was minimized, Voc of 0.7 – 0.8 volt, and photocurrent of 10
- 13.69 mA/ cm2 when UV-O3 treatment were employed and contaminates were eliminated. Therefore, a
total overall efficiency of 4.55 % – 7.0 % could be easily attained with good engineering design and material
process. Fraction three (Fr.3) obtained at 90/10 pentane to toluene ratio has produced the best asphaltene
solar cell so far with 1.8 total efficiency. Un-fractionated asphaltene produced comparable result with fraction
three.

281
REFERENCES
Troshin, P. A Lyubovskaya, R. N. and Razumov, V. F. 2008 Organic Solar Cells: Structure,
Materials,Critical Characteristics, and Outlook. . Nanotechnologies in Russia, . Vol. 3, 5-6.

Shaheen, S. E. Ginley, D. S. Jabbour, G. E. 2005 Organic-based photovoltaics:Toward low-cost

power generation . Materials Research Society bulletin . Vol. 30 , 1.

Kalowekamo, J. and Baker, Erin. 2009. Estimating the manufacturing cost of purely organic solar
cells. Solar Energy., Vol. 83 .

Brabec., Christoph J. 2004. Organic photovoltaics: technology and market . Solar Energy Materials &

M. Gratzel. Dye-sensitized solar cells. 2003. journal of Photochemistry Photobiology, C Vol. 4.

Ruiz-Morales Y., and. Mullins, O. C. 2007. Electronic Absorption Edge of Crude Oils and
Asphaltenes Analyzed by Molecular Orbital Calculations with Optical Spectroscopy. Energy &
Fuels., Vol. 21, pp. 944-952.

Salmanova, Ch. K. Akhmedbekova, S. F. Mamedov, A. P. Kyazimov, S. M. and Abdulova, Sh.

2007. Transformation of resins and asphaltene in photoirradiation. Chemistry and Technology of
Fuels and Oils., Vol. 43, 5, pp. 415-421.

Jiqian Wang, Chuan Li, Longli Zhang, Guohe Que, and Zhaomin Li. 2009.The Properties of
Asphaltenes and Their Interaction with Amphiphiles. Energy Fuels, Vol. 23, 7, pp. 3625–3631.

Mullins, O.C., Sheu, E.Y., Hammami, A., and Marshall, A. Asphaltenes. Heavy oils, and
petroleomics. s.l. : Springer, 2007.

Siffert, B., Kuczinski, J. and Papirer, E. 1990. Relationship between electrical charge and
flocculation of heavy oil distillation residues in organic medium. s.l. : Journal of Colloid Interface
Science,. pp. 107–117. Vol. 135.

Neves, G. B. M. de Sousa, M. A. Travalloni-Louvisse, A. M., Lucas, E. F. and Gonz´alez, G. 2001

Characterization of asphaltene particles by light scattering and electrophoresis. Petroleum Science
and Technology., Vol. 19, pp. 35–43.

Ancheyta, J. Centeno, G. Trejo,F. Marroquín,G. García, J. A. Tenorio,E. and A. 2002.Torres.

Extraction and Characterization of Asphaltenes from Different Crude Oils and Solvents. Energy &
fuel., Vol. 16, 5, pp. 1121–1127.

Seigo Ito, Peter Chen, Pascal Comte, Mohammad Khaja Nazeeruddin, Paul Liska,2007. Fabrication
of screen-printing pastes from TiO2 powders for dye-sensitised solar cells. Progress in
Photovoltaics : Research and applications., Vol. 15, 7, pp. 603-612.

S.Ito, P.Liska, P.Comte, R.Charvet, P.Pechy, U.Bach, L.Schmidt-Mende, S.M. Zakeeruddin,2005.

-
A.Kay, M.K. Nazeeruddin, M.Graetzel,. Control of dark current in photoelectrochemical(TiO 2/I I3)
and dye-sensitized solar cells. Chemical Communication , pp. 4351–4353.

W.Y.Gan, S.W.Lam,K.Chiang,R.Amal,H.Zhao,M.P.Brungs,. 2007.Novel TiO2 thin film with non-UV

activated super wetting and anti fogging behaviours,. Journal of Material Chemistry ., Vol. 17, pp.
952–954.

M.Thelakkat, C.Schmitz,H.-W.Schmidt, 2002. Fully Vapor-Deposited Thin Layer Titanium Dioxide

SolarCells,. Advanced Materials., Vol. 14, pp. 577–581. .

M.Okuya, K.Nakade,S.Kaneko, 2002. PorousTiO2 thin films synthesized by aspray pyrolysis

deposition(SPD) technique and their application to dye-sensitized solar cells,. Solar Energy Material
Solar Cells., Vol. 70, pp. 425–435.

282
 DETERMINATION OF A GEOTHERMAL ENERGY FIELD WITH AUDIO-MAGNETOTELLURIC
(AMT) DATA AT THE SOUTH OF MANISA, TURKEY

Hatice Karakilcik
Department of Geology Engineering, Faculty of Engineering and Architecture, University of Cukurova,
Adana 01330, Turkey

ABSTRACT
In this study, we present an investigation of the geothermal energy field and its energy potential at the south
of Manisa in Turkey. The resistivities of the geothermal energy field have key role to determine goethermal
energy reservoir and potential in the field. This work focuses on low resistivity zones in order to determine
the geometry and extension of the geothermal system by using the Audio-Magnetotelluric (AMT) method that
is to image rock units below the shallow subsurface and determine deeper structure that might represent
permeability in a geothermal system. Therefore, AMT stations were arranged with an interval of about 1 km
array and its measurements were carried out at 40 sites, which were located on approximately 9.5 km long
profiles at the south of Manisa. The resistivity model obtained from 2D inversion of AMT data to show
subsurface specific conductive and resistive features in the field. As a result, two geothermal anomalies were
defined at the study area by using AMT data and the resistivity of soft rock is found generally quite low (<200
m), however hard rock is displayed higher resistivity. But, low resistivities can also evaluate in the fault
zones, because of water content and rock softening.

Keywords: AMT method, geothermal energy, low resistivity, magnetotellurics

INTRODUCTION
Applied electromagnetic research in recent years has been influenced by the growing importance of
geothermal energy. Geothermal energy is a major resource and potential source of low emissions renewable
energy suitable for base-load electricity generation and direct-use applications. It is the natural heat of the
earth stored in rocks and in the fluids within them. Significant geothermal resources can also be associated
with basement rocks heated by natural radioactive decay of elements (e.g., uranium, thorium and potassium)
and in naturally-circulating waters deep in sedimentary basins.
Electromagnetic methods have become powerful geophysical tools in mapping subsurface conductivity
variations. The method is widely used in the exploration of geothermal resources (Árnason et al., 2010). It
mainly involves the propagation of continuous wave or transient electromagnetic fields in the earth. The
source may be natural or artificial. In a natural source method like Magnetotellurics (MT), the fluctuation in
the earth’s natural magnetic field induces an electric field. By measuring the electrical and magnetic field at
the surface of the earth, inferences are made about the conductivity distribution in the subsurface.
MT/AMT survey, targetted at defining the main characteristics of the geothermal reservoir at study area, can
also be supplemented with other types of geophysical surveys that assist us in understanding the regional
geology and the local geological structure in a geothermal prospect.
Audio-Magnetotelluric (AMT) method has found an application in geothermal exploration primarily because
of its ability to detect the depth at which rocks become conductive because of thermal excitation.
Turkey has a unique geographic position at the cross roads between Europe and Asia. It is located on an
active tectonic, orogenetic belt, the Alpine‐Himalaya Orogen, with young faults and active volcanism. Turkey
is rich in geothermal energy resources. Geothermal exploration in Turkey started in the early 1960s. The first
geothermal well for power generation was drilled in 1968 on high enthalpy fields for potential power
production. The well has a depth of 540 meters and a temperature of 198 °C (Kaya, 2012). Turkey has
significant potential geothermal resources associated with buried high heat-producing granites and lower
temperature geothermal resources associated with naturally-circulating waters in aquifers deep in
sedimentary basins. Total geothermal potential of Turkey is estimated around 30000 MW. In geothermal,
Turkey has a proven 4000 MW of resources over 20 °C of which 600 to 1000 MW are high temperature
fields able to produce electricity.

Geology
The study area is located within the northern Menderes Massif in western Turkey. Western Turkey, a region
characterized by the widespread exposures of metamorphic rocks (known as the Menderes Massif−one of
the largest metamorphic core complexes in the world) dissected by approximately E-W trending grabens,
forms a very good example for three modes of extension in the world (Fig.1).

283
Within the study area, four groups of rocks are recognized: (a) the metamorphic rocks–orthogneisses and
metasediments; (b) pegmatoids; (c) Neogene sedimentary rocks; and (d) Quaternary alluvial sediments. The
structural framework of the study area is characterized by gently to moderately tilted (5-40°) Neogene
sedimentary rocks cut by ~E-striking gently to steeply dipping normal faults and northerly (NNW to NNE)
striking oblique-slip cross (or transverse) faults. The predominant E-W strike of normal faults and tilting
reflects regional north-south extension. The strata are also deformed into multiple, discontinuous WNW - to
E-trending folds (Faulds et al., 2009). Similar to other studies in the region resulting primarily from local
reversals in the predominant dip direction of normal fault systems.
In geothermal prospect the required information includes low resistivity bodies, dikes, faults, irregular shaped
bodies and volcanic plugs. The two largest faults in the area are the moderately N-dipping Alaşehir frontal
fault zone and gently (~10-30°) N-dipping Gediz detachment fault. Menderes metamorphic complex, is
typically marked by a thick zone of cataclasis that overprints an earlier mylonitic fabric, but is currently
inactive, with no evidence for Quaternary displacement.

Figure 1. A) Study area. B) Geology and tectonics map of the study area (Faulds et al., 2009).

284
Methods
The Magnetotelluric (MT) method (Cagniard, 1953; Chave and Jones, 2012), which operates in the
frequency range 0.001 to several Hertz (Hz) is a well established technique for imaging the conductivity
structure of the Earths crust. Its depth of investigation is typically from several hundred meters below ground
surface to several kilometers or even deeper. The depth of investigation in MT is a function of subsurface
resistivity and frequency (or the inverse of the period) of the electromagnetic signals. The penetration depth
can be roughly related to the period by the use of the skin depth (δ meter).
An Audio-Magnetotelluric (AMT) sounding system appears to be an effective technique for reconnaissance
exploration to detect shallow resistivity anomalies associated with geothermal reservoirs. Data acquisition for
a single AMT station (Fig.2) is done by measuring the input fields, two horizontal magnetic components Hx
and Hy and the resulting earth response, two horizontal electrical fields, Ex and Ey and the vertical magnetic
field, Hz. The recorded time-series data were converted to the frequency domain and processed to determine
the impedance tensors, which is used to derive apparent resistivities and phases at each site (Fantaye,
2010).
AMT method has proved very useful for subsurface mapping purpose by determining of the low resistivity
area. 40 AMT soundings were collected in study area (Fig.2). Seven profiles AMT measurements were
carried out at this part of the study area using the short-period automatic magnetotelluric system with a
frequency range of 128–0.16 Hz and a maximum depth penetration of about 5 km.

Figure 2. Location map (the profiles consist of AMT data stations at the study area).

RESULTS AND DISCUSSION

Geothermal research in the vicinity of Manisa that make up the study area dates back to beginning in 1976
up to the present by working geological and geophysical studies at Salihli, Alaşehir, Kula, Turgutlu and
Sarıgöl geothermal fields, appropriate drilling locations have been identified. To date, the license area of 18
at a depth of 16,648 m has been 22 geothermal drilling activity and 1978 MWt heat energy was obtained.
The results of the drilling operations are given in Table 1.

Table 1. The results of the geothermal drilling operations in around Manisa from 1976 until 2012
(Republic of Turkey Ministry of Energy and Natural Resources, 2013).
City Town Village Code Year Depth Temp. l/s Potential
(m) (°C) Compressor (MWt)
Manisa Salihli Kurşunlu K-1 1976 42.50 90 20 4.60
Manisa Salihli Caferbeyli SC-1 1989 1189 150 2 0.96
Manisa Salihli Kurşunlu K-2 1992 70 94 45 11.11
Manisa Salihli Kurşunlu K-3 1992 114.3 94 80 19.762
Manisa Köprübaşı Saraycık S-1 1992 386 64.5 22 2.72
Manisa Köprübaşı Demirci S-3 1992 135.2 74 45 7.35
Manisa Köprübaşı Demirci AS-1 1994 122 43 23 0.77
Manisa Köprübaşı Demirci AS-2 1994 270 29 15 0
Manisa Kavaklıdere AK-1 1996 750 63 3 0.35
Manisa Kurşunlu K-4 1996 262 58.6 45 4.45
Manisa Alaşehir Demirci 1997 300 36 25 0.1
Manisa Salihli Kurşunlu K-5 1997 120 83 40 8.04
Manisa Kula Emir E-1 1998 33 65 40 5.02

285
 Manisa Kula Emir E-2 1999 162.5 63 100 11.72
Manisa Turgutlu Urganlı U-1 2001 440.6 61 22 2.39
Manisa Alaşehir Kavaklıdere KG-1 2002 1447.1 182 15 9.23
Manisa Köprübaşı Demirci 2003 116 42 5 0.15
Manisa Alaşehir Kurudere AK-2 2004 1507 213.43 6.74 5.03
Manisa Alaşehir Kavaklıdere 2010/ 2011 2750 287.5 35 36.99
14
Manisa Alaşehir Kavaklıdere 2010/ 2011 1750 159.53 2 2.61
15
Manisa Alaşehir Kavaklıdere 2011/ 2011 2250 88 90 57.64
3
Manisa Sarıgöl Alemşahlı 2011/ 2012 2672 125.21 18 6.8
17

In general, the data quality was very good for all AMT measurement points. AMT data were analyzed and
modeled using a 1D inversion scheme. 1D and 2D inversions are conducted to resolve the conductive
structures. Then corresponding data on seven profiles (A1, A2, A3, B1, B2, B3 and B4) were inverted using
2D inversion schemes (Fig.2).
AMT data results reveal the presence of three major conducting zones beneath the surface in the basement
(Fig.3, 4 and 5). Since these resources may be found in fractured, tectonically stressed areas, their presence
is often marked by microseismic events that also serve as a quide to drilling into the fractured rocks once
other favorable geothermal conditions are esteblished. Since low resistivity usually indicates low permeability
conductive clays, AMT surveys may be used to locate the base of a geothermal caprock and, indirectly, its
high thermal gradient. The dimensions of the reservoir can then be mapped and used to identify drilling
targets and prospective locations of production and injection wells. As it seen in Fig.3, at A1 profile, there are
two low resistivity zones. Resistivity value is lower than 20 Ωm.

Figure 3. AMT measuring points at A1 profile (between 101-102 and 111-113 AMT stations are
observed low resistivity value).

As it seen in Fig.3, depth in the areas with low resistivity values vary between 900 and 2500 m. As the rock
electric conductivity depends on the rock type, the minerals present, the fluid temperature and the fluid
salinity, these surveys can provide excellent imaging of subsurface by delineating rock types, locating deep
fault zones, and mapping subsurface hydrothermal zones which are characterized by extremely low electric
conductivity. The AMT method has found an application in geothermal exploration primarily because of its
ability to detect the depth at which rocks become conductive because of thermal excitation.

286
Figure 4. B2 profile consist of 5 AMT stations and low resistivity values.

As shown in Fig.4, at B2 profile, between 301-302, and 311 AMT stations are also zones with low resistivity
values. Possible geothermal zone detected by the AMT method and the zone of low resistivity values.
Conductive units are clear specifications of areas which are affected by geothermal fluid circulations and
resolving of conductors at deep levels by AMT directly depends their scale. Most of the geothermal fields in
the study area first, most of the fields occur along ~E-W-striking normal faults, but generally not near the
maximum displacement on these fault zones (Faulds et al., 2009). In cases analyzed to date, the structural
settings favoring geothermal activity generate conduits of highly fractured rock along ~E-W-striking fault
zones oriented approximately perpendicular to the least principal stress.

287
Figure 5. It is seen in the AMT stations at B1 profile (red color corresponds to a value lower resistivity).

The geothermal fields in the study area occur along the active southern margin of the Gediz graben.
Fractured rocks of the Menderes Massif, such as mica-schist, gneiss, and especially marbles, are the
reservoir rocks. Cap rocks for the geothermal fluids include clay-rich intervals within the Neogene
sedimentary units (Tarcan et al., 2000). Most hot springs and hot wells with good flow rates lie near the
gently N-dipping Gediz detachment fault (Fig.1), where it intersects and is cut by ~N-S-striking transverse
faults. Karstic marble and breccia along and near the detachment fault provide good channel ways for flow,
possibly somewhat distal to the main upwelling zone. The reservoir for the Kurşunlu field varies between 10-
200 m in depth and resides in highly fractured, commonly karstified zones along the detachment fault.
Temperatures at the Kurşunlu field range from 57 to 120 °C. Empirical chemical geothermometers applied to
the thermo-mineral waters tentatively suggest that reservoir temperatures at Kurşunlu vary between 150 and
230 °C (Tarcan et al., 2000).
Thermal-mineral waters from both systems are used for bathing and medicinal purposes, but those from
Kurşunlu are also utilized in a district heating system for 5,000 homes in the city of Salihli. The total
discharge rate for the main producing wells at Kurşunlu is 145 l/s. The discharge rates are highest where the
reservoir is predominantly formed by marbles (versus schist). Average production temperatures are ~ 90°C
for the district heating.

Key Ingredients of Geothermal Energy Production

Geothermal exploration aims at identifying the geothermal resource, in terms of surface extent, volume, rock
and fluid properties, and collect all necessary information for taking decisions on investing towards a
geothermal power plant. Geothermal energy can simply be described as heat extracted from the earth. The
extraction of this energy relies on the transport of heat from a geothermal reservoir to the surface via a fluid.
This fluid is in direct contact with reservoir rocks and may naturally occur in the reservoir or may have to be
introduced into the system. To understand the key ingredients for a viable geothermal system, it is useful to
look at the net amount of energy produced by a single well. So, the key ingredients for geothermal energy
production can be given as;

MW  c pFT  η  P (1)

288
where cp is the specific heat of the working fluid, F is the flow rate from the production well. η is the efficiency
with which the heat energy can be used, P is the parasitic losses. ΔT is the sensible heat that can be
extracted from the fluid produced by the production hole. The ΔT (Trs-Trj) is dependent on the reservoir
temperature (Trs) and the rejection temperature (Trj) of the energy conversion system. Maximising
Trs requires locating anomalously high thermal gradients, and drilling as deep as possible and Trj
requires cooling and minimisation of scaling in plant and pipelines due to mineral precipitation as
the fluid cools. Parasitic losses are due to pumping of the working fluid out of production wells and
then pumping the fluid back in to the reservoir via injection wells. These losses are likely to be high
due to pumping and cooling requirements. Using the current global experience in power generation from
convective hydrothermal resources, the minimum amount of net energy produced by a well in a geothermal
power system is around 4 MW (Huddlestone and Hayward, 2011)

CONCLUSIONS
Resistivity modeling of the Audio-Magnetotelluric (high frequency MT) data can be used to help resolve the
size and resistivity of the geothermal field. Analysis of the AMT data-set suggests likely signatures of hot
water reservoirs in the area which are distinguished positive to contain low resistivity values. The low
resistivity zone is called conductive zone a deep geothermal reservoir in the study area. The conductive
zones identified in the resistive basement at the AMT measurement sites can clearly be interpreted as the
flow of the fluids in the faults and fractures of the metamorphic rocks. Geothermal activity may be at these
zones at the study area. Numerous geothermal systems probably still remain to be discovered, since many
systems have no surface activity.

ACKNOWLEDGEMENT
The author are thankful to MSc. Geological Engineer Utku Egemen that supports the achievement of AMT
data at study area.

NOMENCLATURE
MT magnetotelluric
AMT audio-magnetotelluric
E East
W West
N North
S South
cp specific heat, J/kg°C
F flow rate, l/s
P parasitic losses, J
ΔT temperature difference, °C
T temperature, °C

Greek letters
η thermal energy efficiency

Subscripts
rs reservoir
rj rejection

References
Arnason, K., Eysteinsson, H., and Hersir, G. P., 2010. Joint 1D inversion of TEM and MT data and 3D
inversion of MT data in the Hengill area, SW Iceland. Geothermics 39, 13-34.

Cagniard, L., 1953. Basic theory of the magneto-telluric method of geophysical prospecting, Geophysics
18:605–635.

Chave, A.D., and Jones, A.G.(eds)., 2012. The Magnetotelluric Method:Theory and Practice, Cambridge
University Press, pp.552

289
Fantaye, T., 2010. Magnetotelluric and transient electromagnetic methods in geothermal exploration with
examples from the Krysuvik area, SW-Iceland. United Nations University ‘Geothermal Training Programme’
Reports 2010, Number 12, pp.151-182 Orkustofnun, Grensásvegur 9, IS-108, Reykjavík, Iceland.

Faulds, J.E., Bouchot, V., Moeck, I., and Oguz, K., 2009. Structural Controls on Geothermal Systems in
Western Turkey: A Preliminary Report.

Huddlestone H.C and Hayward, J., 2011. The potential of geothermal energy, www.csiro.au, Australia.

Kaya, T., 2012. An overview on Drilling and Geothermal Energy Projects in Turkey. GRC Annual Meeting
International Luncheon Reno, NV, USA.

Tarcan, G., Filiz, S., and Gemici, U. 2000, Geology and geochemistry of the Salihli geothermal fields, Turkey:
Proceedings of the World Geothermal Congress, Kyushu-Tohoku, Japan, May 28 – June 10, 2000, pp.1829-
1834.

Republic of Turkey Ministry of Energy and Natural Resources, 2013, Ankara,Turkey.

290
 ENERGY PRODUCTION FROM MUNICIPAL SOLID WASTE USING PLASMA GASIFICATION

C. Dereli and H. Topal

MSc Student, Lecturer in Gazi University
Gazi Üniversitesi Mühendislik Fakültesi Makina Mühendisliği Bölümü
caner_dereli@hotmail.com
htopal@gazi.edu.tr

ABSTRACT
The increasing quantity of Municipal Solid Waste (MSW) is one of the most serious environmental problems in
Turkey. In order to solve this problem the the government has failed to introduce any sustainable and environmental-
friendly remedies. However, among several MSW treatment technologies, plasma gasification may be an alternative
and reliable solution. Plasma gasification is a novel technology which does not fire the MSW as Conventional
combustors do. It converts organic ingredients of the waste into a combustible gas (syngas). This paper deals with
the feasibility study of utilizing plasma gasification and also co-gasification (i.e. gasification of waste with coal) for
energy production from MSW. For this purpose, two methods were used; first determining contents of the released
syngas by mathematical modeling of the gasification process, and second, simulating the same process with the help
of a simulation software (VMGSim). Finally, the results obtained by both methods are compared. Comparing the
results revealed that calculation results from mathematical model were consistent with those outlined from simulation
software.

Keywords: plasma gasification, municipal solid waste, energy, disposal

INTRODUCTION
Wastes can be defined as: “The unused or disposed part of the materials that we use to meet our needs”. In
industries, transportation, agriculture, tourism and construction an enormous amount of materials change form. As we
provide and consume energy for these activities, we produce gas, liquid and solid wastes while a portion of these
wastes is disposed the remaining can be recycled to be reused (www.aso.gov.tr).

The plasma reactor is an enclosed chamber into which the waste is fed. Plasma torches provide the heat, 3000 oC or
higher, in the chamber which converts organic material to a gas and inorganic material into a glassy slag. The plasma
facility may generate electric power, using the fuel gases produced in the reactor. These fuel gases may be
combusted in a waste-heat boiler, or cleaned and fed into a combustion turbine or other combustion device. However,
the plasma facility must be large enough, in terms of waste throughput, to justify the cost of a power generation unit.
The environmental controls on a plasma facility will be located downstream of the reactor and may include scrubbers,
a carbon injection system, or a baghouse, whether or not the facility is generating electricity (R.W. Beck Inc, 2003).

Thermal plasmas have the potential to play an important role in a variety of chemical processes. They are
characterized by high electron density and low electron energy. Compared to most gases even at elevated
temperatures and pressures, the chemical reactivity and quenching rates that are characteristic of these plasmas is
far greater. Plasma technology is very drastic due to the presence of highly reactive atomic and ionic species and
the achievement of higher temperatures in comparison with other thermal methods. In fact, the extremely high
temperatures (several thousands degrees in Celsius scale) occur only in the core of the plasma, while the
temperature decreases substantially in the marginal zones (Moustakas et al., 2005)

Overview of Combustion Technology

We can list the incineration methods used for waste disposal as follows: (R&R Bilimsel ve Teknik Hizmetler Ltd. Şti,
2007)
Grate firing systems,
1. Fluidized bed furnaces,
2. Rotary kilns,
3. Alternative systems,
 Wet oxidation,
 Pyrolysis,
 Gasification and combined systems,
4. Plasma Technologies,

291
Thermal plasma waste disposal

The application of plasma torches for environmental purposes is a relatively new process. Plasma torches operate
simultaneously as a plasma-chemical and a thermal apparatus. The electrical energy of the torches goes into the
plasma which transfers its energy to the substances to be treated, thereby triggering a dual simultaneous reaction
process in the plasma-chemical reactor: the organic compounds are thermally decomposed into their constituent
elements (syngas with more complete conversion of C into gas than in incinerators), and the inorganic materials are
melted and converted into a dense, inert, nonleachable vitrified slag, that does not require controlled disposal.
Therefore, it can be viewed as a totally closed treatment system (Oost, et al, 2006)

The biggest disadvantage of thermal plasma process is its making use of electricity which is a costly energy type.
However when it is considered as a longterm investment to provide a sustainable waste management system it is
also promising from economical point of view. Despite the high costs of electricity, by using transfered arch
equipment the energy will be used efficiently and there will be no extra load to heat the nitrogen-containing air
(Gomeza, 2008). Volatile metal substances will be gasified and along with the off-gas, halogenes and other sour
gases will be carried out of apparatus. Also for degassing purposes several equipments including dust precipitator
filters and absorbers have been installed to the system.

The advantages of plasma compared to the other disposal methods

The advantages of plasma technologies over traditional combustion technologies are given in the table below: (The
Northspan Group Inc., 2008)

Table1.The advantages of plasma technologies in comparison to combustion methods.

Plasma gasification Combustion
Does not release air during the production of Emissions include high level of greenhouse gas emissions,
synthetic gas. other air pollutants, dioxins and furans.
Fumes stack does not occur High-dose fumes stack reveals
In order to void the commercially valuable slags, 30% of solid wastes and potentially hazardous solid waste
solids are reduced to the level of 1/150 which is remained as ash,
Takes place in a deoxygenated channel changer. More air is added to the incinerator.
Plasma generator provides all the energy
Additional fuel is required to sustain the process.
needed for operation
Energy-rich waste are seperated as a fuel. All the energy is converted into heat.

Plasma Gasification Unit

The process that wastes are treated in a plasma solid waste disposal unit is shown in Figure 1. After the excess
humidity is taken away in the dryer, wastes are converted to syngas by means of plasma torch in the gasification
reactor. In order to purify the dust that contains syngas from particles like heavy metals, it is directed to dust
collectors and absorber/scrubbers. Next it is led to the second combustion chamber to be transformed to thermal
energy (heat) and combustion gases leave the system through stack.

Figure 1. Plasma gasification plant flowsheet

292
1.Kinetic Parameters

Gasification process is the result of carbon based kinetic reactions inside the gasification boiler. Reaction balance
is associated with A, E, Alpha (a) and Beta (b) parameters (Virtual Material Group Inc.,2010)

Basic reactions which are formed in the boiler:

Carbon monoxide from, carbon (1); 2C + O2 = 2CO (1.1)
Carbon monoxide from, carbon (2); C + CO2 = 2CO (1.2)
Carbon monoxide from, carbon (3); C + H2O = CO + H2 (1.3)
Carbon monoxide from, carbon; C + O2 = CO2 (1.4)
Carbon dioxide,from carbon monoxide; CO + H2O = CO2 + H2 (1.5)
From carbon to, hydrocarbons; C + 2H2 = CH4 (1.6)
From carbon monoxide to hydrocarbons;CO + 3H2 = CH4 + H2O (1.7)

0.5
(1.8)

(1.9)

(1.10)

(1.11)

(1.12)

(1.13)

3 exp
(1.14)

Figure 2 log N-1 / T graph

293
Reactions occurring in the boiler is drawn on the log K-1/T diagram given above. As it can be seen from the
reaction kinetics temperature is an influential parameter (Dereli, 2012).

2.Analytical and Simulation Methods

Modeling different waste-coal mixtures with analytical methods and a simulation software

Analytical Modeling

Although several different reactions take place in the boiler, five main reaction have been observed:
C + H2O = CO + H2 (Endothermic gasification reactions are oxidation.) (2.1)
C + CO2 = 2CO (Endothermic Boudouard equivalence) (2.2)
C + 2H2= CH4 (Exothermic formation of methane equivalent) (2.3)
CH4 + H2O = CO + 3H2 (Endothermic decomposition of methane equivalent) (2.4)
CO + H2O = CO2 + H2 (Exothermic conversion of gas-water balance) (2.5)
Mountouris et al. have expressed the above reactions as the following general equation : (Mountouris,et al, 2006)
C(x)H(y)O(z) + wH2O + mO2 + 3,76mN2 = n1H2 + n2CO + n3CO2 + n4H2O + n5CH4 + n6N2 + n7C (2.6)

Using the above equations and assuming n5 and n7 values as zero, product content of gasified municipal waste,
with known mass contents, and coal mixture is calculated analytically and the results are presented in the Table 6.

Simulation Methods

Simulation of gasification process of municipal waste and imported coal mixture is carried out by means of a
package software (educational version) and the contents of the produced syngas is calculated for each case.

Table 2. Content of municipal solid waste in Turkey (www.cevreorman.gov.tr)

Component Percent of content Component Percent of content
paper 50 C 47.9
sawdust 3 H 6
wood 11 N 1.2
vegetable waste 3.7 Cl 0.1
fabric 5.5 S 0.3
plastic 10 O 32.9
rubber 3.5 SiO2 11.7
resins 1 Lower heating value:
electronic waste 3 12400 kJ/kg

Construction waste 4.5 Humidity:% 30

glass 4
other 0.8

Similar to what was presented for municipal waste, the mass content of the imported coal used in the research is
presented in Table 3.

294
Table 3. Mass content of the imported coal (Akçura, et al, 1982)

Component C O H N S SiO2 Al2O3 Fe2O3 CaO MgO Na2O

% Content 61.70 13.20 4.10 1.20 0.39 11.17 4.42 1.97 1.18 0.37 0.30

In this study, mass content and calorific values of the syngas produced from different waste and coal mixtures are
analyzed. Different combinations of coal and waste that were studied in this research are presented in Table 4.
Calculations for each of these combinations are carried out analytically and via a simulation software. Obtained
results from each of these methods are discussed and compared.

Municipal waste and coal are analyzed individually and also with waste to coal respective ratios of 25% to 75%,
50% to 50% to 50%, 75% to 25%. Atomic composition for all of the combinations is given in Table 5.

Table 4. Models in the percentages of the waste and coal.

Model 1 100% waste 0% coal

Model 2 75% waste 25% coal

Model 3 50% waste 50% coal

Model 4 25% waste 75% coal

Model 5 0% waste 100% coal

Table 5. Composition of the fuel used in the modeling

Model 2 Model 3 Model 4
Model 1 Model 5
Component (75% MSW (50% MSW (25% MSW
(100% MSW) (100% Coal)
25% Coal) 50% Coal) 75% Coal)

C 47.9 51.35 54.8 58.25 61.7

O 32,9 27.935 23.05 18.125 13.2

H 6 5.525 5.05 4.575 4.1

N 1.2 1.2 1.2 1.2 1.2

S 0.3 0.3225 0.345 0.3675 0.39

SiO2 11.6 11.4925 11.385 11.2775 11.17

Al2O3 - 1.105 2.21 3.315 4.42

Fe2O3 - 0.4925 0.985 1.4775 1.97

CaO - 0.295 0.59 0.885 1.18

MgO - 0.0925 0.185 0.2775 0.37

Na2O - 0.075 0.15 0.225 0.3

humidity 20 18,75 17 15,5 14

Heating Value 12400 15058 17715 20373 23030

(kJ / kg)

295
 Table 6. Synthesis of gas released as a result of modelling compositions (Dereli, 2012)

100% MSW (Software)

25% MSW-%75% Coal

25% MSW-75%Komur

100% Coal (Software)

100% MSW (Analytic)

100% Coal (Analytic)

75% MSW- 25% Coal

50% MSW-50% Coal

50% MSW-50% Coal

75%MSW-25% Coal
Synthesis Gas
Components

Difference%
Difference%

Difference%

Difference%

Difference%
(Software)
(Software)

(Software)
(Analytic)

(Analytic)

(Analytic)
nitrogen 14.339 16.03 -10.5 27.042 27.78 -2.65 36.028 35.56 1.3 42.87 43.47 -1.38 48,585 49.101 -1.12

hydrogen 43.078 42.178 2.07 32.626 33.32 -2.08 25.834 27.95 -7.57 20.27 21.3 -4.84 15.856 16.95 -6.45

hydrogen dyanates 0.088 - - 0.075 - - 0.068 - - 0.06 - - 0.055 - -

hydrogen Sulphide 0.086 - - 0.08 - - 0.076 - - 0.073 - - 0.07 - -

Water 1.255 1.8 -30.27 2.36 0.44 81.35 1.89 1.02 46.03 2.515 0.71 +71.77 2.577 0.95 63.13

carbon monoxide 40.29 38.88 3.5 36.72 37.67 -2.52 35.023 35.18 -0.44 32.81 33.01 -0.6 31.224 30.68 1.74

carbon dioxide 0.635 1.1 -42.27 0.89 0.77 13.48 0.906 0.27 70.2 1.24 1.52 -18.42 1.479 2.31 -35.97

ammonia 0.22 - - 0.19 - - 0.169 - - 0.151 - - 0.137 - -

carbonyl sulfide 0.01 - - 0.009 - - 0.008 - - 0.008 - - 0.008 - -

296
CONCLUSION
In this study five simulations have been carried out for different mixtures of municipal waste, knowing its thermal
capacity and mass content, and imported coal by means of a package software. The purpose of this simulation is to
investigate the contents of the syngas produced by gasification of fuel feeds obtained from each of these various
mixtures. It has been observed that the content of the syngas depends on the mass content of the fuel feed. If the
humidity level is high enough during the gasification process it can be observed that the amount of produced
hydrogen and carbon monoxide increases. As a result, enough humidity must be provided inside the fuel feed.
Similarly in order to increase the amount of carbon dioxide and carbon monoxide in the contents of the syngas
produced via gasification reaction, the amount oxygen being delivered to the system must be raised.

As it can be seen from Table 6 municipal waste, imported coal and various combinations of these two fuels have
been simulated both analytically and via a package software. The point to be considered here is that the amount of
hydrogen calculated for model 1, which includes 100% municipal waste, is considerably more than other cases in
which only coal or different combinations of two fuels were used. The main reason for this is the excessive amount of
humidity in the municipal waste.

Moreover, regarding equation 1.3 presented in the kinetic parameters section, if enough humidity is available in the
setting the rate of produced carbon monoxide will raise. As shown in the Table 6 for model 1, which includes 100%
municipal waste, since it embodies more humidity than all other models, the rate of carbon monoxide produced
during gasification is higher.

In addition, the amount of nitrogen caluclated for municipal waste is far less than the cases where 100% of coal and
various combinations of two fuels were used as model. The reason for this can be explained as follows; Municipal
waste contains a higher rate (32.9% of mass content) of necessary O2 for the production of CO and CO2 during
gasification and as a result less oxidizing air is required. While for the other models containing 100% coal and
different coal-waste mixtures, since they contain a low rate (respectively 13.2% and 23.05% of mass content) of O2,
more oxidizing air is needed for CO and CO2 to be formed during gasification. As a result air/waste combinations are
assumed as, 0.416 mol air/mol waste for municipal waste (model 1), 1.207 mol air/mol coal for coal (model 5), 2.023
mol air/mol mixture for mixtures of municipal waste and coal. Considering that the highest rate of oxidizing air was
used for gasification of coal and as a result of high nitrogen amount in the air, the syngas produced via gasification in
this model will contain the highest amount of nitrogen.

In waste disposal facilities although the amount of produced heat and electric energy is an important factor but the
main purpose of these facilities is to dispose hazardous wastes properly in accordance with rules and regulations. If
we observe the subject from this point of view, compared to other incineration technologies plasma technology is a
cleaner and more environmental-friendly technology. Since in this method the gasification process takes place in high
temperatures, formation of dioxins and furans, which is one of the major problems of incineration facilities, is avoided.
Also stack effluents like heavy metals are accumulated at the bottom of the boiler as vitrified slime and can be
collected safely and reused as packing material. Consequently compared to other incineration methods, disposal via
plasma technology exposes less waste and plays a more environment-friendly role.

As a result of this study plasma gasification technology which is a brand new technology is simulated analytically and
via computer software, amounts of obtained products are compared and mathematical data are presented as a
guidance for the industrialists who want to employ this technology in our country.

297
REFERANCES
Website : http://www.aso.org.tr/kurumsal/index.php?sayfa_no=55 (2011)

Book: R. W. BECK INC. 2003. City of Honolulu Review of Plasma Arc Gasification and Vitrification Technology for
waste disposal final report, 4

Journal: K. Moustakas et al. / Journal of Hazardous Materials B123 (2005) 120–126

Book: R&R Bilimsel ve Teknik Hizmetler Ltd. Şti. 2007. Katı Atık Yakma Tesisleri İçin Teknolojiler ve Yer Seçimi

Journal: Oost, V., Hrabovsky, M. 2006. Pyrolysis of waste using a hybrid argon–water stabilized torch. Vacuum,
80(11-12), 1132-1137

Journal: Gomeza, E., Amutha Rania, D. 2008. Thermal plasma technology for the treatment of wastes: A critical
review Journal of Hazardous Materials

Journal: The Northspan Group Inc., 2008, Does it make sense to move forward with a full feasibility analysis?,
Aitkin County Plasma Gasification Study, Duluth, 3

Book: Virtual Material Group Inc., 2010, VMGSIM ver 6.0 Program Tutorial, 1715-1736

Journal: Mountouris, A., Voutsas, E. 2006 Solid waste plasma gasification: Equilibrium model development and
exergy analysis, 47, 1723-1737

Website :www.cevreorman.gov.tr (2008)

Book: Akçura, F. ve Gerger M., 1982, Başlıca Türk Kömürlerinin özellikleri MTA Enstitüsü, Ankara

Thesis: Dereli C. 2012. Solid waste disposal with plasma gasification technology. MSc Thesis, Gazi University,
Ankara, Turkey (in Turkish)

298
 THE IMPLEMENTATION AND ANALYSIS OF A THERMOELECTRIC GENERATOR FOR
GEOTHERMAL AND WASTE HEAT AREAS
1 2
R. Ahiska , H. Mamur
1
Gazi University, Faculty of Science, Department of Physics, Teknikokullar, Ankara 06500, Turkey, ahiska@gazi.edu.tr
2
Cankiri Karatekin University, Vocational School, Cankiri 18200, Turkey, hmamur@karatekin.edu.tr

ABSTRACT
In the world, Turkey is the fifth and is the first in European in terms of geothermal resources. As a way of
evaluating the rich–resource, this study has proposed a new designed portable thermoelectric generator
employing waste heat in industry and geothermal areas. The implementation and analysis of a new designed
portable thermoelectric generator has been carried out for different conditions. Effects of the hot–cold water
flows, the temperature differences between the surfaces and the load resistance affecting on output power
and efficiency of the thermoelectric generator have been investigated. In the established system, the hot–
cold water flows passed throughout the surfaces of the thermoelectric generator were increased up to 3.7
L/min and 12.8 L/min, respectively. Then, the temperature difference between the surfaces of the
thermoelectric generator was measured as 67 °C. When the load resistance of the thermoelectric generator
was about 10–15 Ω, the maximum power of the thermoelectric generator was obtained 41.6 W.

INTRODUCTION
Geothermal energy is one of the renewable energy sources and it has great potential as a renewable energy
source. In the world, Turkey is the fifth and is the first in European in terms of geothermal resources. One of
the methods converting geothermal energy into electrical energy is the usage of the thermoelectric module
(TEM) as a thermoelectric generator (TEG) (Gou et al., 2013). TEGs convert geothermal energy into
electrical energy directly (O’Shaughnessy et al., 2013). Recently, TEGs attract the attention among the
energy resources because of efficient use of energy, environmental concerns, and global warming. The
advantages of the TEGs in obtaining electrical energy are that they have no moving parts, have simple
construction, silent, environmentally friendly, very sensitive to temperature changes (± 0.1 °C), extremely
reliable, long lasting (100 thousand hours) and their maintenance costs are very lower than other renewable
technologies (Zhang et al., 2013). Besides these advantages, however, they have some disadvantages such
as low efficiency (<10%), and setup costs (Lu et al., 2013). Here, one disadvantage is low efficiency.
However, waste geothermal heats are considered no–cost resources to us. Eventually, low efficiency is not a
serious problem taking into account (Ahiska et al., 2011). The basic structure of a TEM in the TEG mode is
given in the Fig. 1.

heat
TH
Electrical insulator
p/n semiconductors P N P N P N ceramic layer

TC heat
Electrical conductor
copper IL

PL, RL

Fig. 1. A TEM in the TEG mode.

,
The maximum power generated from the TEG is obtained nearly half of the open circuit voltage V OC values.
This value is used in the maximum power point tracker (MPPT) circuits in order to track maximum power
point (MPP) (Lesage et al., 2013). The maximum power, used widely in practice, at any temperature
difference is also expressed by the short circuit current I SC and the open circuit voltage VOC of the TEG as
follows (Kim et al., 2009):

VOC I SC
PL max  (1)
4

299
While TEGs are producing electrical energy from geothermal energy, on the other hand they contribute to
energy efficiency (Yazawa et al., 2013). The usage areas of TEGs can vary. These are such as replacing
battery in self-powered wireless sensor nodes for areas where no electricity network (Leonov et al., 2007),
converting solar energy into electrical energy directly in the grid on–off systems (Xi et al., 2007), the
biomedical systems using the temperature difference between human body temperature and ambient
temperature (Leonov et al., 2009), the telemetry systems requiring less energy demands (Dalola et al.,
2009), heat recovery systems using engine coolant in the internal combustion engines (Karri et al., 2011),
electrification of isolated rural homes (Rinalde et al., 2010).

In this study, the implementation an analysis of a new designed portable TEG has been fulfilled for
converting waste heat energy into electrical energy in waste and geothermal areas and its experimental
tests have been executed. Effects of the flows of hot–cold flow, the temperature difference and the load
resistor of the TEG, which affect the generated power and the efficiency of the TEG, have been investigated.

Application and Experimental Setup of the TEG

Fig. 2. The experimental setup.

The experimental setup of a new designed portable TEG is given in the Fig. 2. The experimental setup
consists of the TEG, (1) the TEG, (2) the TEG data acquisition and test system (Ahiska et al., 2012), and (3)
the hold–cold water circulator, (4) light emitting diode (LED) armature, (5) the hot–cold water flow meters and
(6) the running water. During the experiments and tests of the TEG, a variable load resistance of 500 W
making a nickel–chromium wire has been used as a load resistance of across the output terminals of the
TEG. Twenty TEMs, supplied TES company and have the sizes 5  5  0.4 cm, have been used. The internal
resistance of each one of these is approximately 0.56 Ω at the ambiance temperature of 21.7 °C. These
TEMs have been placed between plates, designed 5.5  5.5  2 cm hollow profile brass, in order to increase
thermal conduction into two groups of teens. All TEGs have been connected electrically in series to increase
the operating voltage and thermally in parallel to decrease the thermal resistance and sandwiched between
two hollow profile brasses. Silicon grease has been applied to their surfaces in order to minimize the thermal
conduction resistance between the cold and hot surfaces of the TEG. Then, they have been inserted
between iron clamps to not move. TEMs and other thermal and electrical components have been put in a
stainless box sizing 20  20  40 cm. Connectors sizing 1/2" on both sides of the TEG have been used to
connect easily the input–output of the TEG to hoses of the hot–cold water, which provide temperatures of the
hot–cold surfaces of the TEG. Similarly, to ease of electrical connection of the TEG, two electrical jacks have
put for the electrical output of the TEG on the box. When the ambient temperature is at 21.7 °C, the total
internal resistance of the TEG has been measured approximately ΣR in = 11.25 Ω. This value is important
because it could be used where the load resistance RL gives the maximum power to us and it changes
depending on the temperature difference of the TEG.

EXPERIMENTAL RESULTS AND DISCUSSION

The hot–cold circulator system and the running water have been used to ensure the heat of the hot–cold
surfaces of the TEG, respectively. Three different hot water flows have been obtained from the TEG, since

300
the hot water circulation has been fulfilled a one phase three speed stage circulation motor branding Alarko.
The temperature of the hot water in the system has been reached approximately 97 °C. The temperature of
the running water has been measured minimum 12 °C. The cold water flow has been varied from about 1.8
L/min to 12.8 L/min. As depending on the hot–cold water flows changes, when the temperature of the hot
water in the circulator has been kept constant value of 3.0 L/min and the temperature of the running water
has been varied, the values of the temperature difference ΔT between the hot–cold surfaces of the TEG and
produced power of the TEG P were given in Fig. 3.

Fig. 3. When the hot water flow and load resistance are constant and the cold water flow is changed, the temperature difference and power of the
TEG changes.

According to the Fig. 3, provided that the temperatures of the hot–cold water are kept constant temperatures,
as the cold water flows increased, the temperature difference between the surfaces of the TEG and the
produced power increased. Also, the internal resistance R in of the TEG varies depending on the temperature.
Therefore, matched load resistance of the TEG changes. This change was from 11.25 Ω at ambient
temperature of 21.7 °C to 15.85 Ω at the maximum hot side temperature of 81.8 °C.

To measure performance values of the TEG, the temperature of the hot water has been begun with at about
30 °C, and it has been increased until at 97 °C gradually. The temperature difference between the surfaces
of the new designed portable TEG has been measured by T-type thermocouples and the temperature
difference has been calculated. By the reason of interacting the hot–cold water mass and their flows, the
temperature difference has been increased up to 67 °C. The hot–cold water flows have been increased up to
maximum 12.8 L/min and 3.7 L/min, respectively. These flows can be higher in the geothermal areas. Thus,
more the temperature difference can be obtained and consequently more the power can be generated.

Experimental data obtained from the new designed portable TEG at the various temperatures have been
given as in the Fig.4. During at the temperature difference between the surfaces of the TEG ΔT = 67 °C, the
open circuit voltage and the short circuit current have been obtained V OC = 51.5 V and ISC = 3.1 A,
respectively. In experiments of the new designed portable TEG, the experimental maximum power has been
obtained PLmax = 41.6 W, when the internal resistances of the TEG and the load have matched R L = Rin = 15
Ω. When the load resistance value of across the output terminal of the TEG generated the maximum power
has been kept constant value RL = 15 Ω and the temperature difference of the TEG is ΔT = 20 °C, the
current, voltage and power varying with the temperature difference between the surfaces of the TEG have
been obtained I = 0.6 A, V = 9.1 V, and P = 5.5 W, respectively. Also, when the load resistance is kept
constant value RL = 15 Ω and the temperature difference is ΔT = 40 °C, the current, voltage and power have
been obtained I = 1.1 A, V = 17 V, and P = 18.7 W, respectively. Then, when the maximum temperature
difference is reached ΔT = 67 °C in the experiment, in this case the hot–cold temperatures of the surfaces of
the TEG are TH = 81.8 °C and TC = 15 °C, the current, voltage and power have been obtained I = 1.6 A, V =
26 V, and P = 41.6 W, respectively. If the temperature difference is reached ΔT = 100 °C, the generated

301
power of the TEG could be increased PLmax = 100 W, since the generated power changes nonlinear as
depending on the temperature difference. In here, the hot–cold water flows used as sources of the TEG will
also important to increase the temperature difference.

Fig. 4. When the load resistance value of across the output terminal of the TEG has been kept constant value RL = 15 Ω, curves of current, voltage
and power varying with the temperature difference between the surfaces of the TEG.

The values of current and voltage obtained from the TEG varies wide ranges due to the connected load of
across the output terminal of the TEG. The efficiency of the TEG increased, as the hot–cold water flows have
been increased. It was measured maximum 3.9%. The conversion efficiency of the TEG as a function of
temperature difference ∆T was given in Fig. 5.

Fig. 5. Conversion efficiency of the TEG as a function of temperature difference.

CONCLUSION
In this study, a new designed portable TEG has been designed to convert waste heat recovery from
geothermal areas into electrical energy and its experiments and tests have been carried out for the various
temperature differences, the hot–cold water flows and the values of the load resistance. The results have
showed that 3.9% efficiency and 41.6 W maximum power output at average temperature difference between
the surfaces of the TEG of 67 °C have been obtained, when the hot–cold water temperatures and their flows
are kept TH = 81.8 °C, TC = 15 °C, dH = 3.7 L/min, and dL = 12.8 L/min, respectively. Effects of the hot–cold

302
water flows providing the temperatures of hot–cold surfaces of the TEG, their temperatures and the load
resistance have been investigated. The curves of power, voltage and current depending on the load
resistance, the temperature difference and the hot–cold water flows have been obtained for this TEG. In the
application areas, if the hot–cold water flows can be increased, the output power of the TEG will be
increased higher values.

In the world, Turkey is the fifth and is the first in European in terms of geothermal resources. As a way of
evaluating the rich–resource, this study has proposed a new designed portable thermoelectric generator
employing waste heat in industry and geothermal areas.

Acknowledge
This study was supported by the Ministry of Science, Industry and Technology, Republic of Turkey as No.
00769–STZ.2011–1 SANTEZ project. We thank to the head of Ankara Chamber of Industry 1. Organized
Industrial Zone and the members of its board of directors for valuable contributions.

NOMENCLATURE
PLmax maximum power, W
VOC open circuit voltage, V
ISC short circuit current, A
Rin internal resistance, Ω
RL load resistance, Ω
∆T temperature difference of the TEG surfaces, °C
η conversion efficiency
dH hot water flow, L/min
dL cold water flow, L/min

REFERENCES
Ahiska, R. and Mamur, H. 2012. A test system and supervisory control and data acquisition application with
programmable logic controller for thermoelectric generators. Energy Conversion and Management 64:15–22.

Ahiska, R., Mamur, H. and Ulis, M. 2011. Modelling and experimental study of thermoelectric module as
generator. Journal of the Faculty of Engineering and Architecture of Gazi University 26:889–896.

Dalola, S., Ferrari, V., Guizzetti, M., Marioli, D., Sardini, E., Serpelloni, M. and Taroni, A. 2009. Autonomous
sensor system with power harvesting for telemetric temperature measurements of pipes. IEEE Transactions
Instrumentation and Measurement 58:1471–1478.

Gou, X., Yang, S., Xiao, H. and Ou, Q. 2013. A dynamic model for thermoelectric generator applied in waste
heat recovery. Energy 52: 201–209.

Karri, M. A., Thacher, E. F. and Helenbrook, B. T. 2011. Exhaust energy conversion by thermoelectric
generator: Two case studies. Energy Conversion and Management 52:1596–1611.

Kim, R. Y., Lai, J. S., York, B. and Koran, A. 2009. Analysis and design of maximum power point tracking
scheme for thermoelectric battery energy storage system. IEEE Transactions on Industrial Electronics
56:3709–3716.

Leonov, V. and Vullers, R. J. M. 2009. Wearable electronics self-powered by using human body heat: The
state of the art and the perspective. Journal Renewable Sustainable Energy 1:701–714.

Leonov, V., Torfs, T., Fiorini, P. and Van Hoof, C. 2007. Thermoelectric converters of human warmth for
self–powered wireless sensor nodes. IEEE Sensors Journal 7:650–657.

Lesage, F. J. and Potvin, N. P. 2013. Experimental analysis of peak power output of a thermoelectric liquid–
to–liquid generator under an increasing electrical load resistance. Energy Conversion and Management
66:98–105.

303
Lu, H., Wu, T., Bai, S., Xu, K., Huang, Y., Gao, W., Yin, X. and Chen, L. 2013. Experiment on thermal
uniformity and pressure drop of exhaust heat exchanger for automotive thermoelectric generator. Energy
54:372–377.

O’Shaughnessy, S. M., Deasy, M. J., Kinsella, C. E., Doyle, J. V. and Robinson A. J. 2013. Small scale
electricity generation from a portable biomass cookstove: Prototype design and preliminary results. Applied
Energy 102:374–385.

Rinalde, G. F., Juanicó, L. E., Taglialavore, E., Gortari, S. and Molina, M. G. 2010. Development of
thermoelectric generators for electrification of isolated rural homes. International Journal of Hydrogen Energy
35:5818–5822.

Xi, H., Luo, L. and Fraisse, G. 2007. Development and applications of solar–based thermoelectric
technologies. Renewable and Sustainable Energy Reviews 11:923–936.

Yazawa, K., Koh, Y. R. and Shakouri, A. 2013. Optimization of thermoelectric topping combined steam
turbine cycles for energy economy. Applied Energy 109:1–9.

Zhang, M., Miao, L., Kang, Y. P., Tanemura, S., Fisher, C. A. J., Xu, G., Li, C. X. and Fan, G. Z. 2013.
Efficient, low–cost solar thermoelectric cogenerators comprising evacuated tubular solar collectors and
thermoelectric modules. Applied Energy 109:51–59.

304
 ELECTROSYNTHESIS, CHARACTERIZATION AND ELECTROCATALYTIC PROPERTIES OF
Pt-Fe NANOPARTICLES TOWARDS THE OXIDATION OF METHANOL AND FORMIC ACID
1* 1 2
Biuck Habibi , Zahra Gorbani and Elnaz Razmian
1
Electroanalytical Chemistry Laboratory, Faculty of Sciences, Department of Chemistry, Azarbaijan Shahid Madani
University, Tabriz, Iran, E-mail: B.Habibi@azaruniv.edu
2
Faculty of Sciences, Department of Chemistry, Payam Nour University, Ardabil, Iran.
ABSTRACT

Bimetallic Platinum-Iron alloy nanoparticles modified carbon-ceramic electrode as an electrocatalyst

was prepared by a two-step procedure: (I) the carbon-ceramic electrode (CCE) was produced by using sol-
gel technique and (II) Platinum-Iron nanoparticles alloy was deposited electrochemically on the surface of
CCE and the resulting modified electrode was referred as Pt-Fe/CCE. The obtained electrocatalyst, Pt-
Fe/CCE, was characterized by scanning electron microscopy, energy-dispersive X-ray spectroscopy, X-ray
diffraction and cyclic voltammetry. The electrocatalytic activity of the Pt-Fe/CCE towards the electrooxidation
of the methanol and formic acid was evaluated by cyclic voltammetric technique in 0.1 M of H 2SO4 solution. It
was found that the Pt-Fe/CCE was catalytically more active than Pt nanoparticles (alone) supported on the
same substrate. The obtained results also showed that the presence of Fe atoms greatly enhance the
activity of the Pt towards the electrooxidation of methanol and formic acid. Moreover, it contributes to reduce
the amount of the noble metal in the anode of direct liquid fuel cells (DLFCs), which remains one of the
challenges to make the technology of DLFCs economically viable.

Keywords: Methanol, Acid formic, nanoparticles, Alloy, Carbon-ceramic, Electrooxidation

1. INTRODUCTION
Due to the increasing demand of energy and the continuously rising oil price the use of renewables as a
source for the production of fuel and chemicals has become more and more important during the last few
years. Therefore, studies on the fuel cells currently increase because they have great potentials for portable,
transport and stationary power supply applications [1]. Methanol oxidation is of great importance in fuel cell
technology, namely in the direct methanol fuel cell (DMFC) aimed to power electric vehicles [2]. On the other
hand, formic acid has recently attracted attention as an alternative fuel for direct liquid fuel cells (DLFCs) due
to its high theoretical open circuit voltage (1.45 V) and reduced fuel cross-over compared to methanol. In
addition, formic acid is a strong electrolyte that facilitates proton transport in catalyst layers [3].
One of the main problems in DLFCs is the development of anodic materials with high electroactivity
towards the oxidation of methanol and formic acid. It is well known that the platinum (Pt) is an effective
electrocatalyst for these oxidations. However, Pt is also highly sensitive to CO poisoning: the catalyst surface
is progressively poisoned by the adsorbed CO (CO ads), which is formed as a result of the stepwise
dehydrogenation of these oxidants in oxidation reaction [4, 5]. Poisoning CO ads species can be oxidatively
removed from the Pt surface through a Langmuir-Hinshelwood-type surface reaction with neighboring OHad
species electrosorbed from water at more positive potentials.
Alloying of Pt with oxophilic metals enables electrochemical dissociation of water on oxophilic metal
sites at more negative potentials compared to pure Pt and, therefore, allows electrocatalytic oxidation of
COads at lower anodic over potentials. Most of literature reports concerning bimetallic electrocatalysts are
related to Pt-Ru [6–8], while combinations of Pt with other secondary metals such as Sn and Co have also
been reported [9, 11]. On the other hand, one of the 3d transition metals, iron (Fe), cobalt (Co) and nickel
(Ni), can be used to modify the nature of Pt catalysts for increased activity [12–15]. These transition metals
are expected to be cheaper than Ru. Recently, Pt-Fe catalysts have been intensively studied for use in fuel
cells. The studies conducted on Pt-Fe alloy as cathodic catalyst suggest that it contains a higher proportion
of Pt active sites with a distinctly different nearest neighbour environment in relation to nonalloyed Pt sample,
however, the preparation and electrocatalysis of Pt–Fe alloy as an anodic one are lacking in the literature.
Recently, Pt-Fe alloy catalysts were found to have good electrocatalytic performances for methanol [16, 17]
and formic acid electrooxidation [18].
In this paper, bimetallic Pt-Fe alloy nanoparticles modified electrode as an electrocatalyst was
prepared by a two-step procedure and characterized by scanning electron microscopy, energy-dispersive X-
ray spectroscopy, X-ray diffraction and cyclic voltammetry. The electrocatalytic activity of the Pt-Fe/CCE
towards the electrooxidation of the methanol and formic acid was evaluated by cyclic voltammetric technique
in 0.1 M of H2SO4 solution. It was found that the Pt-Fe/CCE was catalytically more active than Pt
nanoparticles (alone) (Pt/CCE) supported on the same substrate. The obtained results also showed that the

305
presence of Fe atoms greatly enhance the activity of the Pt towards the electrooxidation of methanol and
formic acid. Finally, the effect of some experimental factors was studied and optimum conditions were
suggested. The long-term stability of nanoparticles electrocatalyst has also been studied.

2. EXPERIMENTAL
2.1. Chemicals
Methyltrimethoxysilane (MTMOS), methanol, formic acid, H2PtCl6.5H2O, FeSO4, HCl, H2SO4 and graphite
powder of high purity were obtained from Merck or Fluka. All solutions were prepared with double distilled
water.

2.2. Procedure of CCE preparation

The sol-gel processing method was used for fabricating CCE according to the following procedure: The
amount of 0.9 ml MTMOS was mixed with 0.6 ml methanol. After addition of 0.6 ml HCl 0.1M as catalyst, the
mixture was magnetically stirred (for about 15 min) until it produced a clear and homogeneous solution.
Then, 0.3 g graphite powder was added and the mixture was shaken for another 5 min. Subsequently, the
homogenized mixture was firmly packed into a Teflon tube (with 3 mm inner diameter and 10 mm length) and
dried for at least 24 h at room temperature. A copper wire was inserted through the other end to set up
electric contact. The electrode surface was polished with 1500 emery paper and rinsed with water.

2.3. Procedure of Pt-Fe/CCE preparation

The Pt-Fe/CCE was prepared as follow: After polishing of the CCE, the Pt-Fe nanoparticles were
electrodeposited on the CCE surface from an aqueous solution of 0.1 M H 2SO4 containing H2PtCl6.5H2O and
FeSO4 via cyclic voltammertic method (potentiodynamic method). The electrode potential is swept between 0.5
-1
and -1.1 versus the saturated calomel electrode (SCE) (at a scan rate of 20 mV s ) for the certain number of
cycles. The Pt-Fe (50:50) was used as the electrocatalyst in methanol and formic acid electrooxidation. The
actual surface area of the Pt-Fe/CCE was calculated based on hydrogen adsorption/desorption peaks which
2
is over 43.8 times larger than that of polycrystalline Pt [actual surface area (Ar=0.1 cm )].

2.4. Instrumentation
The electrochemical experiments were carried out using an AUTOLAB PGSTAT-100
(potentiostat/galvanostat) equipped with a USB electrochemical interface and a driven GEPS software was
used for electrochemical experiments. A conventional three electrode cell was used at room temperature.
The smooth Pt or the nanoparticles modified electrode (Pt-Fe/CCE or Pt/CCE) was used as a working
electrode. A SCE and a platinum wire were used as the reference and auxiliary electrodes, respectively.
JULABO thermostat was used to control cell temperature at a constant value. Scanning electron microscopy
(SEM) and energy-dispersive X-ray spectroscopy (EDX) were performed on a LEO 440i Oxford instrument.
X-ray diffraction (XRD) of the Pt-Fe nanoparticles was studied using a Brucker AXF (D8 Advance) X-ray
power diffractometer with a Cu Kα radiation source (λ = 0.154056 nm) generated at 40 kV and 35 mA.

3. RESULTS AND DISCUSSION

3.1. Non-electrochemical characteristics of Pt-Fe/CCE
In order to surface characterization of the electrocatalyst, the surface micrograph of bare substrate and Pt-Fe
nanoparticles alloy on it has investigated by SEM and the corresponding results were shown in Fig. 1. Fig.
1A shows the surface structure of the bare CCE immediately after polishing with emery paper grade 1500.
As seen in this image the surface is dense, scaly and has a high porosity. The average pore size of CC
substrate is about 0.2 µm. Fig. 1B and 1C show the SEM image of the CCE surface after Pt-Fe deposition.
As seen, the wall and flower like Pt-Fe nanoparticles were obtained on the electrode (image 1B). High
magnification shows (Fig.1 C) that, each wall or flower like structures of Pt-Fe in this image is not the
individual Pt-Fe crystallites. They are clews consisting of crystallite aggregates.
EDX measurement was used to analysis the surface composition of the catalyst. The EDX spectra of
Pt-Fe/CCE is shown in Fig. 2A. The bare CCE data (Fig. 2B) is also shown for comparison. It is evident that
Pt and Fe are present on the CCE support, indicating that the H2PtCl6 and FeSO4 precursors can be co-
reduced to their respective metal phases by electroco-deposition method. The calculated chemical
composition of the alloy catalyst indicates that the ratio of Pt:Fe in electrocatalyst ≈1:1 (a.t.), which agree
well with the composition in the starting mixture.

306
Fig. 1. SEM images of (A) CCE surface immediately after polishing, (B) same electrode after Pt-Fe nanoparticles deposition (Pt-Fe/CCE) and (C)
Pt-Fe nanoparticles with high magnification.

Fig. 2. EDX data for the Pt-Fe nanoparticles modified CCE (A) and bare CCE (B).

In order to get information on the average particle size and the Pt–Fe alloy effect, the Pt–Fe/CCE
o
was characterized by XRD pattern presented in Fig. 3. The peak at approx. 55 (006) corresponds to the
0
supporting material (carbon-ceramic) [19]. No Fe or Fe2O3 peaks are observed in the XRD pattern of Pt–
Fe/CCE sample, indicating the absence of segregated metallic Fe phases [20]. The main characteristic
peaks of the face-centered cubic (fcc) crystalline Pt at 2θ = 26, 41.5, 46.8, 68.3, and 83 namely, the planes
(001) (111), (200), (220), and (311) [20]. The XRD patterns of Pt-Fe/CCE and Pt/CCE (not shown here)
indicate that Fe was dissolved into the fcc lattice of Pt and forming an alloy between Pt and Fe atoms. On the
other hand, the Pt-Fe/CCE catalyst preserved the fcc Pt lattice structure, with a slight shift of Pt diffraction
peaks.

Fig. 3. XRD patterns of Pt-Fe/CCE.

307
3.2. Electrochemical characteristics of Pt-Fe/CCE
To understand the electrochemical behavior of the Pt-Fe/CCE, the cyclic voltammetric responses of the Pt-
Fe/CCE and also Pt/CCE, smooth Pt and CC electrodes for comparison were recorded in 0.1 M H2SO4
-1
solution at a scan rate of 50 mV s . Fig. 4 shows the CVs of the Pt-Fe/CCE, smooth Pt and CCE (inset A,
curve 1 and 2, respectively) and Pt/CCE (inset above). Although the typical Pt-peaks for the hydrogen under-
potential deposition, the oxidation of hydrogen, formation of platinum oxide, and its reduction are present on
the Pt-Fe/CCE, they become ill-shaped compared to the smooth Pt (inset below) and even to the unalloyed
Pt modified CCE (Pt/CCE) (inset above). On the other hand, compared with the Pt/CCE, both peaks for
hydrogen adsorption and desorption of Pt-Fe/CCE decrease. In fact the decrease of hydrogen peak was
linear with the increase of amount of Fe in the Pt-Fe/CCE.

Fig. 4. CV of Pt-Fe/CCE in 0.1M H2SO4 at a scan rate of 50 mV sL1. The insets are CVs of Pt/CCE, bare CCE and smooth Pt in the same
conditions, respectively.

The actual active surface area of the Pt-Fe/CCE is equivalent to the number of Pt sites available for
hydrogen adsorption/desorption. The Pt-Fe/CCE produced QaH of 0.92 mC. Accordingly, the actual active
surface area (Ar) can be obtained from charge for hydrogen adsorption as: A r = QaH/Q0 (Q0 has been
2 2
commonly taken as 210 µC/real cm ) = 4.8 cm . The hydrogen adsorption charge of polycrystalline Pt was
calculated at 0.021 mC. These results show that the actual active surface area of the Pt-Fe/CCE is about
43.8 times larger than that of smooth Pt.
Inset (middle) of Fig. 4 shows the CV of bare CC substrate in 0.1M H 2SO4 solution. No
adsorption/desorption peaks of hydrogen appeared at the bare CCE.

3.3. Electrocatalytic oxidation of methanol on the Pt-Fe/CCE

Fig. 5 shows the VCs of Pt-Fe alloy electrocatalyst (blue line), Pt/CCE (brown line) and smooth Pt (dark blue
line) in 0.1 M H2SO4 + 0.5 M methanol. During the forward scan of the CV of smooth Pt, the first anodic peak
results from methanol oxidation and in the backward scan there is a small reduction current peak due to the
reduction of Pt oxide. With the clean Pt available, the methanol oxidation takes place more easily and
therefore the anodic current peak for methanol oxidation appears. While, it can be observed from the CV of
methanol on the Pt-Fe/CCE, that in the positive scan direction a main anodic peak at 0.61 V and a small
another anodic peak at about 1.2 V and in the negative scan direction a small cathodic peak at and a sharp
anodic peak at about 0.35 V were observed. The corresponding peaks on the Pt/CCE electrocatalyst are
located at 0.75 and 1.22V in the forward scan direction and at 0.41 and 0.35 V in the backward scan
direction. Comparison of the CVs in Fig. 1 and CV of methanol on the smooth Pt shows that the CV profile of
the Pt-Fe/CCE electrocatalyst has the usual characteristics of the smooth Pt except that the oxidation current
of main peak in the forward scan on the Pt-Fe/CCE is much higher than that on the smooth Pt, and
moreover, it is higher than that on the Pt/CCE. The anodic current of the main peak is significantly enhanced
after the introduction of Fe metal, e.g., smooth Pt<Pt/CCE<Pt-Fe/CCE. On the other hand, it is evident that
Pt-Fe/CCE electrocatalyst exhibits a lower onset potential than that of the Pt/CCE. Obviously, the Pt-Fe/CCE
exhibits the high electrocatalysis performance in methanol oxidation with respect to Pt/CCE, because it is

308
generally recognized that the onset potential can be an indicator in determining the electrochemical activity
for methanol oxidation. Another noticeable difference between the CV of methanol oxidation on the Pt-
Fe/CCE and Pt/CCE is the ratio of anodic current in the forward scan and backward scan, If/Ib. The ratio of
If/Ib for Pt-Fe/CCE in methanol oxidation is more than that of the Pt/CCE. Such a high value for Pt-Fe/CCE
electrocatalyst indicates that most of the intermediate carbonaceous species were oxidized to CO 2 in the
forward scan, suggesting that the interaction of Pt and Fe leads to the less poisoning of Pt by the CO-like
intermediates formed during methanol activation.

Fig. 5. CVs of 0.5M methanol on the Pt-Fe/CCE (blue line), Pt/CCE (brown line) and smooth Pt (dark blue line) in 0.1M H2SO4 at a scan rate of 50
mV s-1.

In order to reveal the correlation between methanol and Pt oxide species on the Pt-Fe/CCE, we have studied
the effect of upper limit potentials (EU) in cyclic potential scanning on the oxidation of methanol. Fig. 6 shows
the CVs of methanol oxidation on the Pt-Fe/CCE catalyst for EU of 1.0 to 1.6 V. As seen in Fig. 6, by
increasing the final positive potential limit, the anodic current of methanol oxidation in the PS remains
unchanged, but oxidation current in the backward scan is decreased. In lower limit potential, the Pt oxides
with high valence haven’t developed greatly, so the effect of the Pt oxides with high valence on methanol
oxidation in the backward scan is relatively small. It can be also seen that the potential of methanol oxidation
peak remains constant in the forward scan, while the potential of methanol oxidation peak shifts positively in
the backward scan. On the other hand, the peak current in the backward scan decreased as the EU
increased.

Fig.6. Effect of upper limit of potential scanning region on the electrooxidation of 0.5M methanol on the Pt-Fe/CCE in 0.1M H2SO4. (1) -0.3–1.0 V,
(2) -0.3–1.1 V, (3) -0.3–1.2 V, (4) -0.3–1.3 V, (5) -0.3–1.4 V, (6) -0.3–1.5 V, and (7) -0.3–1.6 V.

309
Indeed by increasing the final positive potentials the conversion of Pt to PtO is accelerated and caused a
decrease of oxidation current in the backward scan, which further demonstrates that methanol can only be
oxidized on clean metallic platinum nanoparticles surface. In order to investigate the kinetic characterization
of methanol oxidation on the Pt-Fe/CCE, we looked into the effect of scan rate on the behavior of methanol
oxidation. The obtained result clearly reveals that the peak current associated to the methanol
electrooxidation increases linearly with the scan rate and the current value of main anodic peak is liner vs.
1/2
υ . This behavior indicates that the electrocatalytic process under study is controlled by diffusion.

3.4. Electrocatalytic properties of Pt-Fe/CCE towards the oxidation of formic acid

Fig. 7 shows the CVs of the Pt-Fe/CCE and smooth Pt (inset A) in a 0.5M formic acid + 0.1M H2SO4 solution.
During the forward scan of the CV of smooth Pt , the first anodic peak results from formic acid oxidation,
while the second peak at 1.06 V can be attributed to CO oxidation and FA oxidation on sites that were
previously blocked by COads. This is because COads is generated as an intermediate during the
electrochemical oxidation of formic acid. The CV of Pt-Fe/CCE shows the usual characteristics of smooth Pt
except that both for forward and reverse scan directions the oxidation currents of formic acid on the Pt-
Fe/CCE are significantly higher than on the smooth Pt. At potentials with more than ca. 0.35 V, the reaction
becomes accelerated and maximum rate at ca. 0.70 V occurs. Another increase in current at potentials more
than ca. 1.35 is assigned to the reaction intermediates and poisonous adsorbed species to CO 2. Upon
reversing the potential sweep, a very steep increase of the reaction rate at ca. 0.5 V develops and a
maximum current is observed at ca. 0.38 V. After that, the current gradually decreases but the reaction rate
is still faster than in the forward scan. This large anodic peak in the reverse scan is attributed to the removal
of the incompletely oxidized carbonaceous species formed in the forward scan.

Fig. 7. CV of 0.5M formic acid on the Pt-Fe/CCE in 0.1M H2SO4 at a scan rate of 50 mV s-1. Inset is the CV of 0.5M formic acid on the smooth Pt in
the same conditions

The influence of upper limit potentials (anodic reversal potential) and effect of scan rate on the
electrooxidation of formic acid were studied and similar results with methanol were obtained.
Finally, from the practical view, the long-term stability of Pt-Fe/CCE was examined in 0.1M H2SO4
solution containing 0.5M methanol or 0.5M formic acid. It can be observed that the anodic currents remain
constant with an increase in the scan number at the initial stage and then start to decrease after 50 scans.
The peak current of the 250th scan is about 88% than that of the first scan. In general, the loss of the
catalytic activity after successive number of scans may result from the consumption of these fuels during the
CV scan.

4. CONCLUSION
A bimetallic Pt-Fe nanoelectrocatalyst with the Pt-Fe atomic ratio in surface layer 1:1 is synthesized by co-
depositing on highly prose carbon-ceramic from a solution of the corresponding metals. Electrooxidation of
methanol and formic acid was studied on the carbon-ceramic electrode modified bimetallic Pt-Fe
nanoparticles as a potent electrocatalyst in acidic media. The results were compared to those at the same
electrode modified by alone Pt nanoparticles and smooth Pt electrode. It was found that Pt-Fe electrocatalyst
was catalytically more active than platinum nanoparticles on the same substrate and Pt electrode. The high
electrocatalytic activity of the Pt-Fe electrocatalyst towards the oxidation of methanol and formic acid may be
directly related to effect of the Fe atoms in electrocatalytic activity of the Pt and its large surface electroactive

310
area. Moreover, it contributes to reduce the amount of the noble metal in the anode of DAFCs, which
remains one of the challenges to make the technology of DAFCs economically viable.

ACKNOWLEDGMENT
The authors gratefully acknowledge the research council of Azarbaijan Shahid Madani University for financial
support.

REFERENCES
[1] Klabunde K.J. 2001. Nanoscale Matrials in Chemistry, wiley-Interscience, New York.
[2] Faraji M., Y. Yamini, and M. Rezaee. 2010. Magnetic Nanoparticles: Synthesis, Stabilization, Functionalization,
Characterization, and Applications. Journal Iranian Chemistry Society 7:1-37.
[3] Stamatin I., A. Morzan, A. Dumitru. 2004. NANOSTRUCTURED MATERIALS USED FOR DESIGN OF
THE COMPONENTS OF FUEL CELLS. Romanian Journal Physics 49: 767-775,
[4] Zhong C.J., J. Luo, B. Fang, B.N. Wanjala, P.N. Njoki, R. Loukrakpam and J. Yin. 2010 Nanostructured
catalysts in fuel cells. Nanotechnology 21:062001(20pp)
[5] Guo S. and E. Wang. 2011. Noble metal nanomaterials: Controllable synthesis and application in fuel cells and
analytical sensors. Nanotechnology Today. 6:240-264.
[6] Kadirgam B.B, F.C. Lamy and J.M. Leger. 1981. Electrocatalytic oxidation of methanol on platinum-based
binary electrodes. Journal Electroanalytical Chemistry.127:75-85.
[7] Hoster H., T. Iwasita, H. Baumgartner, and W. Vielstich. 2001. Pt-Ru model catalysts for anodic methanol
oxidation: Influence of structure and composition on the reactivity. Physical Chemistry Chemical Physics.
3:337-346.
[8] Gupta S.S., S.S. Mahapatra, and J. A. Datta. 2004. Potential anode material for the direct alcohol fuel cell.
Journal Power Sources.131:169-174.
[9] Witham C.K., W. Chun, T.I. Valdez, and S.R.Narayanan. 2000. Performance of Direct Methanol Fuel Cells with
Sputter-Deposited Anode Catalyst Layers. Electrochemistry Solid-State Letter 3:497-500.
[10] Habibi B. and M.H. Pournaghi-Azar. 2010. Methanol oxidation on the polymer coated and polymer-stabilized
Pt nano-particles: A comparative study of permeability and catalyst particle distribution ability of the PANI and
its derivatives. International Journal of Hydrogen Energy.35:9318-9328.
[11] Niu L., Q. Li, F. Wei, X. Chen, and H.Wang. 2003. Electrochemical impedance and morphological
characterization of platinum-modified polyaniline film electrodes and their electrocatalytic activity for methanol
oxidation. Journal Electroanalytical Chemistry.544:121-128.
[12] Pournaghi-Azar M.H. and B. Habibi. 2007. Electrocatalytic oxidation of methanol on poly(phenylenediamines)
film palladized aluminum electrodes, modified by Pt micro-particles:Comparison of permselectivity of the films
for methanol. Journal Electroanalytical Chemistry. 601:53-62.
[13] Frelink T., W. Visscher, and J.A.R. van Veen. 2010. Particle size effect of carbon-supported platinum
catalysts for the electrooxidation of methanol. Journal of Electroanalytical Chemistry 382 (1995) 65-72
[14] Gan L., H. Du, B. Li, and F. Kang. 2010. The effect of particle size on the interaction of Pt catalyst particles
with a carbon black support. Carbon 48: 2125-31.
[15] Mukerjee S. and J. McBreen. 1998. Effect of particle size on the electrocatalysis by carbon-supported Pt
electrocatalysts: an in situ XAS investigation. Journal of Electroanalytical Chemistry. 448:163-171.
[16] Wang C. ,D. Li,M. Chi, J. Pearson, R.B. Rankin, J.ff Greeley, Z. Duan, G. Wang, D. van der Vliet, K.L. More,
N.M. Markovic, V.R. Stamenkovic. 2012. Rational Development of Ternary Alloy Electrocatalysts. Journal
Physical Chemistry Letter. 3:1668-1673.
[17] Stonehart P. 1990. Development of Advanced Noble Metal-Alloy Electrocatalysts for Phosphoric Acid Fuel
Cells (PAFC). Berichte der Bunsengesellschaft für physikalische Chemie. 94:913–921.
[18] Xiang L., A. Li, Z. Lijuan, L. Fan, W. Xiayan, and X. Dingguo. 2011. Pt-Based Intermetallic Compounds as
Electrocatalysts. Progress in Chemistry. 23:501-508.
[19] Habibi B. and N. Delnavaz. 2011. Carbon-ceramic supported bimetallic Pt-Ni nanoparticles as an
electrocatalyst for oxidation of formic acid. International Journal of Hydrogen Energy. 36:9581-9590.
[20] Xu J., K. Hua, G. Sun, C. Wang, X. Lv, and Y. Wang. 2006. Electrooxidation of methanol on carbon
nanotubes supported Pt–Fe alloy electrode. Electrochemistry Communications. 8:982-986.
[21] Zhang J.N., S.J. You, Y.X. Yuan, Q.L. Zhao, G.D. and Zhang. 2011. Efficient electrocatalysis of cathodic
oxygen reduction with Pt–Fe alloy catalyst in microbial fuel cell. Electrochemistry Communications.13: 903-905.

311
 A NEW CARBON NANOTUBE-SUPPORTED PT-RU ANODIC CATALYST BY REVERSE
MICROEMULSION FOR DIRECT METHANOL ELECTRO-OXIDATION

Jenshi B. Wang*, Charng-Ching Yeh, Han-Chang Gao

Department of Chemical Engineering, I-Shou University
Ta-shu District, Kaohsiung 840, Taiwan
jbwang@isu.edu.tw:

ABSTRACT
A multi-walled carbon nanotube (CNT)-supported Pt-Ru nanocatalyst prepared by reverse microemulsion was studied
for direct methanol electro-oxidation and compared with that made by polyol method. XRD and TEM characterization
showed that the reverse micelles-synthesized alloyed particles attained a smaller average particle size and displayed
a higher dispersion on the support with narrow size distribution. Results of cyclic voltammetry and
chronoamperometry in 1 M methanol+0.5 M sulfuric acid at room temperature revealed that the Pt-Ru/CNT catalyst
synthesized by reverse microemulsion exhibited a much higher electro-catalytic activity than that based on polyol
process.

INTRODUCTION
Direct methanol fuel cells (DMFC) have been attracting considerable interests in the applications to transportation and
portable devices because liquid methanol as fuel is more tractable than hydrogen. Although many different catalysts
have been investigated, Pt-Ru alloy system is among the most active anodic catalysts for the electro-oxidation of
methanol. Several methods have been developed for the preparation of multi-metallic nanoparticles (Toshima et al.,
1993; Link et al., 1999; Reetz et al., 1995). Among these, the synthesis through microemulsion appears as one of the
most promising methods since nanodroplets of water serve as nanoreactors favoring formation of monodispersed
nanoparticles. This method has advantages in that it does not require extreme conditions of temperature and
pressure, while providing a convenient control of size and composition (Bagwe et al.,1997; Bagwe et al., 2000;
Zhang et al., 2004).
Carbon black such as the commercial Vulcan XC has been employed as support of the nanosized metallic particles in
fuel cells. However, the presence of micropores limits its widespread applications because metallic particles may
easily become trapped in the pores. In this regard, in the search for highly active catalyst electrodes, carbon
nanotubes have been used recently as an alternative support, providing catalysts with better electrochemical
performance due to their unique morphology and high electrical conductivity (Wang et al., 2004; Li et al., 2002; Liu et
al., 2002). Although the benefit of CNT as electro-catalyst support has been recognized, still there is a dearth of works
devoting to the use of microemulsions to synthesize CNT-supported Pt-Ru nanoparticles as the anodic catalysts for
methanol electro-oxidation.
In this paper, the preparation of Pt-Ru/CNT nanocatalyst using a water-in-oil reverse microemulsion of
water/AOT/isooctane (AOT: sodium di-2-ethylhekyl sulfosuccinate) is reported. The structure and performance of the
catalyst prepared by this method in the methanol electro-oxidation are presented and compared with those of the
catalyst made by polyol process.

EXPERIMENTAL
Synthesis of nano Pt-Ru particles and supported catalysts
For microemulsion-synthesized nanoparticles, the reverse micellar solutions were prepared using AOT (Sigma) as
surfactant and isooctane (Mallinckrodt) as the oil phase. Aqueous solutions of H2PtCl6 (99.95%, Alfa Aesar), RuCl3
(99.99%, Alfa Aesar) and N2H5OH (Merck) were used to form the reverse micelles. The size of the particles was
controlled by fixing the molar ratio of surfactant to water (ω = [H2O]/[AOT]) at 5. All chemicals were analytically pure
and used as received without further purification. Deionized water was used throughout. An appropriate amount of
RuCl3 was added to the solution of H2PtCl6 to ensure the Pt:Ru atomic ratio of 1:1 before formation of reverse
micelles. Multi-walled carbon nanotube (CNT) (Sigma Aldrich AL-636509-10G) was taken as support. Before use, the
CNT was ultrasonically treated in nitric acid (65%, Fluka) at 50 oC for 4 h, and dried in vacuum oven at 60 oC for 12 h.
The acid pre-treated CNT was added to the solution of AOT/H2PtCl6 + RuCl3/isooctane under ultrasonic oscillation at
room temperature for 1 h. Afterwareds, another similar microemulsion containing hydrazine as aqueous phase, i.e.,
AOT/hydrazine/isooctane, was added into the CNT slurry under constant ultrasonic oscillation for 24 h. Platinum-
ruthenium nanoparticles were formed upon contact between the precursor containing droplets and the hydrazine
containing droplets. The concentrations of aqueous H2PtCl6 + RuCl3 and aqueous hydrazine were 0.1 and 1.0 M,
respectively. Ethanol was then employed to destabilize the microemulsion to provoke deposition of Pt-Ru
nanoparticles onto the support. The mixture was filtered and washed three times with 95% ethanol, and then the
resulting supported catalyst was dried in a vacuum oven at 60 oC for 12 h. The Pt loading of all catalysts was kept at
20 wt%. The catalyst is marked as Pt-Ru/CNT (emulsion).

312
For polyol-based nanoparticles, appropriate amounts of H2PtCl6 and RuCl3 with the Pt:Ru atomic ratio of 1:1 were
dissolved in ethylene glycol (EG) (99%, Panreac Quimica Sa). A glycol solution of NaOH (0.5 M) was added dropwise
into the glycol solution of precursors with stirring to bring the pH of solution to 12. The temperature was then
increased to 185 oC, and kept constant for 2 h so that the metals were reduced adequately. Afterwards, an
ultrasonically dispersed acid pre-treated CNT slurry was added dropwise with constant stirring for 4 h. The mixture
was separated and washed with deionized water for three times, and then the resulting filter cake was dried in a
vacuum oven at 60 oC for 12 h. Likewise, the Pt loading of the catalyst was kept at 20 wt%. The catalyst prepared by
this technique is designated as Pt-Ru/CNT (EG).
Experiments of transmission electron microscopy (TEM) were carried out on a Philips CM200 system to examine
surface morphology of the catalysts. Samples were first ultrasonicated in ethanol for 20 min and then deposited on
200 mesh Cu grids covered with Formvar/carbon films. X-ray diffraction (XRD) patterns were obtained on a Rigaku
MultiFlex diffractometer using Cu Kα radiation to study crystal structure of the supported catalysts.

Preparation of electrodes and electrochemical tests

Five milligram of catalyst was added to the mixture of 1 ml of methanol and 50 μl of perfluorosulfonic acid (5 wt%
Nafion, Aldrich) to formulate catalyst ink. A 25 μl ink slurry was then sampled and spread uniformly over a 3-mm
glassy carbon disk electrode (with an area of 0.0707 cm2). After methanol evaporation, the electrode was dried in
oven at 75 oC for 10 min.
All of the electrochemical performance studes were conducted at room temperature on the CHI 608B potentiostat (CH
Instrument) in a three-electrode cell, with SCE electrode as reference electrode and platinum wire as counter
electrode. All potentials in this work are quoted against reversible hydrogen electrode (RHE). The electrolyte was a
solution of 1 M CH3OH+0.5 M H2SO4, which was degassed with nitrogen for 20 min before the tests. The
electrocatalytic activity studies by cyclic voltammetry (CV) for methanol oxidation were carried out in the electrolyte at
10 mV s-1. To compare the long-term performance of the supported catalysts, chronoamperometrical (CA) studies
were undertaken in the 1 M CH3OH+0.5 M H2SO4 electrolyte under a constant potential of 0.947 V vs. RHE for 10
min.

RESULTS AND DISCUSSION

The structure of Pt-Ru/CNT catalysts
XRD patterns of the Pt-Ru/CNT catalysts prepared by the methods of polyol and microemulsion are shown in Fig. 1.
The peak at around 2θ = 25o in the XRD spectrum is ascribed to the diffraction of CNT. The peaks at about 40o, 46o,
68o and 81o can be assigned to the Bragg reflection of Pt (111), (200), (220), and (311) plane, respectively. A shift to
higher 2θ-values and an increase of peak broadening, with respect to the diffraction lines of pure Pt, were observed,
suggesting Pt-Ru alloy formation with a decrease in lattice parameter and mean particle size in these two catalysts.
The extensive peak broadening in the XRD patterns is attributed to the ultrafine dimensions of the crystallites formed.
No diffraction lines corresponding to those of pure Ru or Ru2O were detected, indicating that the catalysts synthesized
exhibited Pt face centered cubic structure. These observations are consistent with the results of (Gasteiger et al.,
1993) for a Pt50-Ru50 alloy, and also agreed with those in the literature (Chu et al., 1996; Zhou et al., 2003; Gojkovic
et al., 2003). The average crystallite size of Pt-Ru/CNT (emulsion) catalyst calculated from Scherrer formula is ca.
3.54 nm.

313
 Pt-Ru/CNT(EG)
Intensity (a.u.)

Pt-Ru/CNT(emulsion)

20 30 40 50 60 70 80 90

2 theta (degree)
Fig. 1. XRD patterns of Pt-Ru/CNT (EG) and Pt-Ru/CNT (emulsion) catalysts

Fig. 2 shows the TEM image depicting the morphology of the Pt-Ru/CNT (emulsion) catalyst. The Pt-Ru particles
obtained from the microemulsion method were small spherical ones which were fairly well dispersed on the CNT with
a uniform particle size distribution. The spherical particle morphology might be as a result of the redox reaction
between H2PtCl6 or RuCl3 and hydrazine occurring within the dispersed aqueous droplets. Also shown in Fig. 2 is its
size distribution histogram. Its average TEM particle size of 3.40 nm, which is very close to that calculated by the
Scherrer formula, is smaller than that of 4.46 nm for Pt-Ru/CNT (EG) made by polyol process.

20 nm
(a)

314
 35

30

25
Frequency (%)

20

15

10

0
0 1 2 3 4 5 6 7 8 9 10 11
Particle size (nm)

(b)
Fig. 2. (a) TEM micrograph, and (b) TEM particle size distribution of Pt-Ru/CNT (emulsion) catalyst

Electrocatalytic activities of electrodes

Cyclic voltammograms showing electro-catalytic activity of the electrodes prepared from Pt-Ru/CNT (emulsion) and
Pt-Ru/CNT (EG) are compared in Fig. 3. The CV curves exhibit typical features of methanol electro-oxidation on the
platinum alloy electrode. In the anodic scan, the onset of methanol oxidation occurred at ca. 0.367 V vs. RHE for Pt-
Ru/CNT (emulsion), which is much lower than the onset potential of 0.67 V for Pt-Ru/CNT (EG). Furthermore, the
peak current density is 17.98 mA cm-2 at the peak potential of 0.886 V for Pt-Ru/CNT (emulsion), while the
corresponding value for Pt-Ru/CNT (EG) is 2.79 mA cm-2 at 0.913 V. The Pt-Ru/CNT (emulsion) catalyst prepared by
microemulsion method exhibited a much higher current density than that of Pt-Ru/CNT (EG). It is also clearly
observable that methanol re-oxidation due to the reduction of platinum oxide occurred in the cathodic scan. These CV
findings demonstrate the advantages of employing reverse micelles microemulsion to prepare Pt-Ru nanocatalysts
that have high catalytic activity for methanol electro-oxidation.

20

Pt-Ru/CNT (Emulsion)
Pt-Ru/CNT (EG)
15
Current density (mA cm-2)

10

-5
0.0 0.2 0.4 0.6 0.8 1.0 1.2

V (V vs RHE)
Fig. 3. Cyclic voltamograms of Pt-Ru/CNT (EG) and Pt-Ru/CNT (emulsion) in 1 M CH3OH+0.5 M H2SO4 at a scan rate
of 10 mV s-1

Chronoamperometrical plots of current density vs. time at a constant potential of 0.947 V is displayed in Fig. 4. These
curves reflect the activity and stability of the catalysts in catalyzing the methanol electro-oxidation. The potentiostatic

315
currents decreased rapidly in the very initial period. This may be due to the formation of intermediate species, such as
COads, COOHads, and CHOads, during the methanol electro-oxidation (Gu et al., 2006). For clarity, the inset shows the
magnification of the very initial currents. Moreover, Pt-Ru/CNT (emulsion) maintained both higher initial and limiting
current densities over time, which implies a much better catalytic activity and stability of methanol electro-oxidation by
this catalyst. These results agree with those of CVs shown in Fig. 3. All these findings suggest that CNTs coupled with
microemulsion technique can improve the electrochemical performance of DMFC anodes. Similar results are obtained
for carbon fibers with cup-stacked-type structure (de Moraes et al., 2006).

100

95
80 71.63
Current density (mA cm-2)

60

0 15

40

Pt-Ru/CNT (Emulsion)
Pt-Ru/CNT (EG)
20

0 100 200 300 400 500 600

Time (s)
Fig. 4. Chronoamperometrical plot at 0.947 V vs. RHE for 10 min for Pt-Ru/CNT (EG) and Pt-Ru/CNT (emulsion) in 1
M CH3OH+0.5 M H2SO4, with the current magnification in the first initial 15 s shown in the inset.

CONCLUSIONS
A reverse micelles process coupled with multi-walled carbon nanotubes has been developed and proved as an
excellent means for the preparing of supported Pt-Ru nanoparticles as catalysts for direct methanol electro-oxidation.
The reverse microemulsion-synthesized Pt-Ru/CNT catalyst displayed a much higher electro-catalytic activity in
methanol oxidation than a polyol-based one. The higher activity thus obtained by the former catalyst is attributable to
the morphology of its unique nanostructure in that small spherical particles were homogeneously dispersed on the
surface of carbon nanotubes with narrow particle size distribution.

Acknowledgement
The authors are grateful for the financial supports of National Science Council, Taiwan under the grant NSC 100-
2221-E-214-044, and I-Shou University under the contract ISU 100-02-03.

REFERENCES
Bagwe, R. P. and K. C. Khilar. 1997. Effects of the intermicellar exchange rate and cations on the size of silver
chloride nanoparticles formed in reverse micelles of AOT. Langmuir 13:6432-6438.
Bagwe, R. P. and K. C. Khilar. 2000. Effects of intermicellar exchange rate on the formation of silver nanoparticles
in reverse microemulsions of AOT. Langmuir 16:905-910.
Chu, D. and S. Gilman. 1996. Methanol electro-oxidation on unsupported Pt-Ru alloys at different temperatures.
Journal of the Electrochemical Society 143:1685-1690.
Gasteiger, H. A., N. Markovic, P. N. Ross and E. J. Cairns. 1993. Methanol electrooxidation on well-characterized
Pt-Ru alloys. Journal of Physical Chemistry 97:12020-12029.
Gojkovic, S. L., T. R. Vidakovic and D. R. Durovic. 2003. Kinetic study of methanol oxidation on carbon-supported
PtRu electrocatalyst. Electrochimica Acta 48:3607-3614.

316
Gu, Y. J. and W. T. Wong. 2006. Nanostructure PtRu/MWNTs as anode catalysts prepared in a vacuum for direct
methanol oxidation. Langmuir 22:11447-11452.
Li, W. Z., C. H. Liang, J. S. Qiu, W. J. Zhou, H. M. Han, Z. B. Wei, G. Q. Sun and Q. Xin. 2002. Carbon nanotubes
as support for cathode catalyst of a direct methanol fuel cell. Carbon 40:791-794.
Link, S, Z. L. Wang and M. A. El-Sayed. 1999. Alloy formation of gold-silver nanoparticles and the dependence of
the plasmon absorption on their composition. Journal of Physical Chemistry B 103:3529-33.
Liu, Z. L., X. H. Lin, J. Y. Lee, W. Zhang, M. Han and L. M. Gan. 2002. Preparation and characterization of
platinum-based electrocatalysts on multiwalled carbon nanotubes for proton exchange membrane fuel cells.
Langmuir 18:4054-4060.
de Moraes I. R., W. J. da Silva, S. Tronto and J. M. Rosolen. Carbon fibers with cup-stacked-type structure: An
advantageous support for Pt-Ru catalyst in methanol oxidation. Journal of Power Sources 160:997-1002.
Reetz, M. T., W. Helbig and S. A. Quaiser. 1995. Electrochemical preparation of nanostructured bimetallic clusters.
Chemistry of Materials 7:2227-2228.
Toshima, N., T. Yonezawa and K. Kushihashi. 1993. Polymer-protected palladium-platinum bimetallic clusters -
preparation, catalytic properties and structural considerations. Journal of the Chemical Society-Faraday
Transactions 89:2537-2543.
Wang, C, M. Waje, X. Wang, J. M. Tang and R. C. Haddon and Y. S. Yan. 2004. Proton exchange membrane fuel
cells with carbon nanotube based electrodes. Nano Letters 4:345-348.
Zhang, P and L. Gao. 2003. Copper sulfide flakes and nanodisks. Journal of Materials Chemistry 13:2007-10.
Zhou, W. J., Z. H. Zhou, S. Q. Song, W. Z. Li, G. Q. Sun, P. Tsiakaras and Q. Xin. 2003. Pt based anode catalysts
for direct ethanol fuel cells. Applied Catalysis B-Environmental 46:273-285.

317
 ELECTROOXIDATION OF C1-C4 ON THE Pt-Sn ALLOY NANOPARTICLES MODIFIED
CARBON-CERAMIC ELECTRODE
Biuck Habibi*, Elaheh Dadashpour and Nasrin Delnavz
Electroanalytical Chemistry Laboratory, Department of Chemistry, Faculty of Sciences, Azarbaijan Shahid Madani
University, Tabriz, Iran E-mail: B.Habibi@azaruniv.edu

ABSTRACT
The electrooxidation of C1-C4 alcohols was studied on the platinum-tin nanoparticles carbon-ceramic
modified electrode (Pt-Sn/CCE) as a potent electrocatalyst in a mixed these alcohols and 0.1 M H 2SO4
solutions. The Pt-Sn/CCE catalyst shows good electrocatalytic activity for electrooxidation of these alcohols
in comparison with platinum nanoparticles carbon-ceramic modified (Pt/CCE) and smooth Pt electrodes.
Indeed, the presences of Sn atoms in alloy enhance the electrocatalytic activity of Pt towards the oxidation of
these alcohols and reduce the amount of Pt in the anodic material of direct alcohol fuel cells. On the other
hand, the Pt-Sn/CCE catalyst has satisfactory stability and reproducibility for electrooxidation of these
alcohols in acidic media when stored in ambient conditions or continues cycling making it more attractive for
fuel cell applications.

Keywords: Alcohols, nanoparticles, Alloy, Carbon-ceramic, Electrooxidation

1. INTRODUCTION
Nanoscience is one of the most important research and development frontiers in modern science and use of
nanoparticles materials offers many advantages due to their unique size and physical properties [1]. Owing
to these properties, the nanoparticle materials are being favorable considered for a wide variety of electronic,
medicine, batteries, and fuel cells applications. One the most research activities on production and
application of the nanoscience in fuel cells are focused on: synthesis and use of nanocatalysts as the anodic
materials for fuel cells. Metal and metal alloy nanoparticles attract many interests for application in catalysts
for fuel cell [2]. Previous reports have been shown that the best performing fuel cell electrodes as an anode
in the direct alcohol fuel cells (DAFCs) are either Pt or Pt-based materials. These catalysts have been one of
the most expensive materials in the fabrication of components for DAFCs.
As we know, a DAFCs that uses hydrocarbon liquid fuel has several merits, namely it has a high
specific energy, does not require a reformer, and it is easy to handle the liquid alcohol fuel. Methanol,
ethanol, 2-propanol and 2-butanol have all been considered as alcohol fuels. The general equation for
alcohol oxidation may be written as:
+ −
CnH2n+1OH + (2n−1) H2O → (n)CO2 + (6n)H + (6n)e

At present, methanol is one of the most widely studied organic fuels [3] and it is known to produce
high current densities in DAFCs with platinum based catalysts. Nevertheless, during the reaction, the Pt
surface becomes poisoned by strongly adsorbed intermediates such as carbon monoxide, due to the
dissociative chemisorptions of methanol. On the other hand, ethanol has much less toxicity and also has a
higher specific energy. Consequently, ethanol electrooxidation can be considered to be an important
research topic. It has been reported that ethanol is oxidized to acetaldehyde and CO 2 by a dual-path
mechanism in which oxidation to CO2 proceeds via adsorbed intermediates and to acetaldehyde by the
dehydrogenation of adsorbed ethanol.
Also, during the last decade, several researchers have reported the operation of a direct 2-propanol
fuel cell, and recently it has been also reported that the performance of a cell operating on 2-propanol is
substantially higher than when it is operating on methanol. On the other hand, four-carbon alcohols have
received particular attention because they can be produced in significant quantities from a variety of
renewable feed stocks and have several advantages over short-chain alcohols.
In our previous work [4], we have explored a simple method for the fabrication Pt-Sn alloy
nanoparticles on the carbon-ceramic electrode (CCE) which has several interested characteristics, such as
high conductivity, relative chemical inertness, good mechanical properties, physical rigidity and high porosity
as a substrate in catalysts deposition. The obtained catalyst (Pt-Sn/CCE) was characterized by scanning
electron microscopy, energy-dispersive X-ray spectroscopy, X-ray diffraction and cyclic voltammetery. Based
on the electrocatalytic behavior of Pt-Sn nanoparticles towards the formic acid [4], the present work is
focused on electrooxidation of C1-C4 alcohols on the bimetallic Pt-Sn alloy nanoparticles. This is the

318
continuation of our previous works on the role of metal or metal alloy nanoparticles in fuel cell applications
[5-6].

2. EXPERIMENTAL
2.1. Chemicals
Methyltrimethoxysilane (MTMOS), methanol, ethanol, 2-propanol, 2- butanol, H2PtCl6.5H2O, SnCl2.2H2O,
HCl, H2SO4 and graphite powder of high purity were obtained from Merck or Fluka. All solutions were
prepared with double distilled water.

2.2. Procedure of CCE preparation

The sol-gel processing method was used for fabricating CCE according to the following procedure: The
amount of 0.9 ml MTMOS was mixed with 0.6 ml methanol. After addition of 0.6 ml HCl 0.1M as catalyst, the
mixture was magnetically stirred (for about 15 min) until it produced a clear and homogeneous solution.
Then, 0.3 g graphite powder was added and the mixture was shaken for another 5 min. Subsequently, the
homogenized mixture was firmly packed into a Teflon tube (with 3 mm inner diameter and 10 mm length) and
dried for at least 24 h at room temperature. A copper wire was inserted through the other end to set up
electric contact. The electrode surface was polished with 1500 emery paper and rinsed with water.

2.3. Procedure of Pt-Sn/CCE preparation

The Pt-Sn/CCE was prepared as follow: After polishing of the CCE, the Pt-Sn nanoparticles were
electrodeposited on the CCE surface from an aqueous solution of 0.1 M H 2SO4 containing H2PtCl6.5H2O and
SnCl2.2H2O at -0.2 V versus the saturated calomel electrode (SCE). The Pt-Sn (50:50) was used as the
electrocatalyst in alcohols electrooxidation.

2.4. Instrumentation
The electrochemical experiments were carried out using an AUTOLAB PGSTAT-100
(potentiostat/galvanostat) equipped with a USB electrochemical interface and a driven GEPS software was
used for electrochemical experiments. A conventional three electrode cell was used at room temperature.
The smooth Pt or the nanoparticles modified electrode (Pt-Sn/CCE or Pt/CCE) was used as a working
electrode. A SCE and a platinum wire were used as the reference and auxiliary electrodes, respectively.
JULABO thermostat was used to control cell temperature at a constant value.

3. Results and discussion

3.1. Electrocatalytic oxidation of methanol on the Pt-Sn/CCE
Fig. 1 shows the cyclic voltammogram (CV) of 0.5 M methanol in the 0.1 M H 2SO4 solution on the Pt-Sn/CCE
-1
electrocatalyst at scan rate of 50 mV s . It can be observed from the CV of methanol on the Pt-Sn/CCE, that
in the positive scan (PS) direction a main anodic peak (I) at 0.86 V and a small another anodic peak (II) at
about 1.18 V and in the negative scan (NS) direction a small cathodic peak (III) at 0.35 V and a sharp anodic
peak (IV) at about 0.33 V were observed. Comparison of the CV in Fig. 1 and CV of methanol on the smooth
Pt [7, 8] and even Pt/CCE (not shown here) shows that the CV profile of the Pt-Sn/CCE electrocatalyst has
the usual characteristics of the smooth Pt except that the oxidation current of main peak in the PS on the Pt-
Sn/CCE is much higher than that on the smooth Pt, and moreover, it is higher than that on the Pt/CCE [9].
The anodic current density of the main peak is significantly enhanced after the introduction of Sn metal, e.g.,
-2 -2
Pt/CCE (0.65 mA cm ) < Pt-Sn/CCE (2.3 mA cm ). On the other hand, it is evident that Pt-Sn/CCE
electrocatalyst exhibits a lower onset potential of ca. 0.318 V than that of the Pt/CCE [9]. Obviously, the Pt-
Sn/CCE exhibits the high electrocatalysis performance in methanol oxidation with respect to Pt/CCE,
because it is generally recognized that the onset potential can be an indicator in determining the
electrochemical activity for methanol oxidation [10]. Another noticeable difference between the CV of
methanol oxidation on the Pt-Sn/CCE and Pt/CCE is the ratio of anodic current in the PS and NS, IPS/INS. The
ratio of IPS/INS for Pt-Sn/CCE in methanol oxidation is 2.8 which is more than 1.7 multiple higher than that of
the Pt/CCE (1.6). Such a high value for Pt-Sn/CCE electrocatalyst indicates that most of the intermediate
carbonaceous species were oxidized to CO 2 in the PS, suggesting that the interaction of Pt and Sn leads to
the less poisoning of Pt by the CO-like intermediates formed during methanol activation [11]. The different
ranges of potential over which, the formation and dissolution of surface oxides occur on the smooth Pt or on
the Pt based electrodes, is a striking feature of the electrochemical behaviors of these electrodes [12]. It is
also reflected in a kinetic irreversibility of most electrocatalytic oxidation even reduction reactions that
proceed on these electrodes [13]. In order to reveal the correlation between methanol oxidation and Pt oxide
species on the Pt-Sn/CCE, we have studied the effect of upper limit potentials (EU) in cyclic potential

319
scanning on methanol oxidation. Fig. 2 shows the CVs of methanol oxidation on the Pt-Sn/CCE catalyst for
EU of 1.0 to 1.8 V. As seen in Fig. 2, by increasing the final positive potential limit, the anodic current of
methanol oxidation in the PS remains unchanged, but oxidation current in the NS is decreased. In lower limit
potential, the Pt oxides with high valence haven’t developed greatly, so the effect of the Pt oxides with high
valence on methanol oxidation in the NS is relatively small. It can be also seen that the potential of methanol
oxidation peak remains constant in the PS, while the potential of methanol oxidation peak shifts positively in
the NS. On the other hand, the peak current in the NS decreased as the EU increased. Indeed by increasing
the final positive potentials the conversion of Pt to PtO is accelerated and caused a decrease of oxidation
current in the NS, which further demonstrates that methanol can only be oxidized on clean metallic platinum
nanoparticles surface [14].

Fig. 1. CV of 0.5 M methanol on the Pt-Sn/CCE in 0.1 M H2SO4 at a scan rate of 50 mV s-1.

Fig. 2. Effect of upper limit of potential scanning region on the electrooxidation of 0.5 M ethanol on the Pt-Sn/CCE in 0.1 M H2SO4. 0.0–0.8 V, 0.0–
0.9 V, 0.0–1.0 V, 0.0–1.1 V, 0.0–1.2 V, 0.0–1.3 V and 0.0–1.5 V. Scan rate 50 mV s-1.

3.2. Electrocatalytic oxidation of ethanol on the Pt-Sn/CCE

-1
Fig. 3 shows the CV of 0.1M ethanol in the 0.1 M H2SO4 solution on the Pt-Sn/CCE at scan rate of 50 mV s .
For the electrooxidation of ethanol on the smooth Pt electrode [15], the appearance of two main peaks on
the PS can be ascribed to the oxidation of the fuel by two kinds of chemisorbed oxygen species. Several
spectroscopic and electrochemical studies have been down [16, 17] and the mechanism of ethanol oxidation
in acid solution has been summarized as the following scheme of parallel reactions:
CH3CH2OH → [CH3CH2OH]ads → C1ads; C2ads → CO2 (total oxidation)
CH3CH2OH → [CH3CH2OH]ads → CH3CHO → CH3COOH (partial oxidation)
in which C1ads and C2ads represent the adsorbed intermediates on the electrode surface with one and two
carbon atoms, respectively.

320
Fig. 3. CV of 0.1 M ethanol on the Pt-Sn/CCE in 0.1M H2SO4 at a scan rate of 50 mV s-1.

Similar behavior is observed for the ethanol oxidation on the Pt-Sn/CCE electrocatalyst; the CV profiles of
the Pt-Sn/CCE electrocatalyst have the usual characteristics of the smooth Pt [15]. But, comparison of CV in
Fig. 3 shows that the oxidation currents on the Pt-Sn/CCE electrocatalyst, both for the PS and NS, are much
higher than those on the smooth Pt, and moreover, they are higher than those on the Pt/CCE [9]. Obviously,
Pt-sN/CCE exhibits the high electrocatalysis performance in ethanol oxidation. Additionally, it is noted that
the onset potential of ethanol oxidation on the Pt-Sn/CCE is even lower than those on the Pt/CCE and
smooth Pt. The obtained results again indicate that the addition of Sn into Pt catalysts can significantly
improve the catalyst performance for the ethanol electrooxidation at relatively less expense, signifying that
the Pt-Sn/CCE electrocatalyst is more active than smooth Pt and Pt/CCE catalysts.

3.3. Electrocatalytic oxidation of 2-propanol on the Pt-Sn/CCE

Fig. 4 curve red shows the CV of 0.15 M 2-propanol in the 0.1 M H2SO4 solution on the Pt-Sn/CCE
-1
electrocatalyst at the scan rate of 50 mVs . The inset of Fig.4 also shows the CV of 2-propanol oxidation on
the smooth Pt electrode. As can be seen in the inset, no CV features of hydrogen adsorption/desorption at a
Pt electrode appeared in the potential region of hydrogen adsorption/desorption, which indicated effectively
that a part of Pt surface was covered by adsorbates of 2-propanol like other alcohols [18, 19]. Higher than
these potentials, in the PS, the oxidation of 2-propanol occurred in two anodic peaks (a1 and a2) in the range
of potential between 0.0 and 1.5 V. While, in the NS the oxidation of 2-propanol took place in the relatively
wide ranges of potential and a current peak occurred at 0.25 V (a 3). However, a small negative current due
to the reduction of the surface Pt oxides which is formed at higher potentials in the PGPS, was observed in
the range of potential between 0.63 and 0.45 V (c 1). Based on some previous reports [20, 21], it is
considered that the main electrochemical reaction in the oxidation of 2-propanol on the Pt and Pt-based
electrocatalysts is dehydrogenation, conversion of 2-propanol to acetone, with the charge-transfer reaction of
hydrogen. Besides the above main reaction, the electrochemical reaction including an intermediate, 2-
propanol → intermediate → carbon dioxide, proceeds successively in parallel.

Fig. 4. CV of 0.15 M 2-propanol on the Pt-Sn/CCE in 0.1M H2SO4 at a scan rate of 50 mV s-1. Inset is the CV of 2-propanol at the smooth Pt
electrode.

Electrooxidation of 2-propanol on the Pt-Sn/CCE occurs at two oxidation peaks, the first and the second
anodic peaks in the PS and at another oxidation peak in the NS. The first anodic peak appears around 0.40

321
V with the onset potential (Eonset = -0.1 V), the second anodic peak appears around 1.1 V. Considering that
the OH adsorption by water decomposition on the Pt surface occurs at about 0.4V, the low onset potential is
probably due to presence of Sn metal in the alloy and dehydrogenation of 2-propanol. On the other hand, in
presence of Sn atoms, because the bifunctional mechanism could be operational at lower potential, both
mechanisms could occur at these potentials. The first and the second anodic peaks are related to the direct
oxidation of adsorbed species formed during the adsorption of 2-propanol. The appearance of the two peaks
on the PS can be ascribed to the oxidation of the fuel by the two kinds of chemisorbed oxygen species [22].
In order to reveal the correlation between 2-propanol oxidation and Pt oxide species, we have studied the
effect of the upper limit potentials (EU) in the cyclic potential scanning on the 2-propanol oxidation. The
reason, as reported in the literature, is the fact that the different ranges of potential over which the formation
and dissolution of surface oxides occur on the smooth Pt or on the Pt-based electrode, form a striking
feature of the electrochemical behaviors of these electrodes [12]. Fig. 5 shows the CVs of 2-propanol
oxidation on the Pt-Sn/CCE electrode for EU of 1.1 to 1.5 V. As seen in Fig. 5, by increasing the final positive
potential limit, the anodic current density of 2-propanol oxidation in the PS remains unchanged, but the
oxidation current density in the NS is decreased. In the lower limit potential, the Pt oxides with a high valence
do not develop greatly, so the effect of the Pt oxides with a high valence on the 2-propanol oxidation in the
NS is relatively small. It can be also seen that the potential of the 2-propanol oxidation peak remains
invariable in the PS, while the potential of the 2-propanol oxidation peak shifts positively in the NS. On the
other hand, the peak current in the NS decreased as the EU increased. Indeed by increasing the final
positive potentials, the conversion of Pt to PtO is accelerated and this causes a decrease of the oxidation
current density in the NS, which further demonstrates that 2-propanol can only be oxidized on a clean
metallic platinum nanoparticles' surface [23].

Fig. 5. Effect of upper limit of potential scanning region on the electrooxidation of 0.15 M 2-propanol at the Pt-Sn/CCE in 0.1 M H2SO4. 0.0-0.9 V,
0.0-1.0 V, 0.0-1.1 V, 0.0-1.2 V, 0.0-1.3 V, 0.0-1.4 V and 0.0-1.5 V. Scan rate 50 mV s-1.

3.4. Electrocatalytic oxidation of 2-butanol on the Pt-Sn/CCE

Fig. 6 shows the CVs of 0.15 M 2-butanol in the 0.1 M H2SO4 solution on the Pt-Sn/CCE and Pt/CCE
-1
electrocatalysts at scan rate of 50 mVs . Three oxidation peaks, 0.36 (A1), 0.95 (A2) and 0.25 V (A3) are
observed for the electrooxidation of 2-butanol on the Pt-Sn/CCE electrocatalyst. The appearance of these
two main peaks on the PS for the electrooxidation of 2-butanol on the Pt-Sn/CCE electrocatalyst can be
ascribed to the oxidation of this fuel by two kinds of chemisorbed oxygen species [24, 25]. In the NS, an
anodic peak appears and this is attributed to the renewed oxidation of the fuel. This behavior is very similar
to that shown by 2-propanol oxidation. A comparison of the CVs of Fig. 6 and the oxidation of 2-butanol on
the smooth Pt (not shown here) shows that the CVs profile of the Pt-Sn/CCE electrocatalyst has the usual
characteristic of smooth Pt except the fact that the oxidation current on the Pt-Sn/CCE electrocatalyst, both
for the PS and NS, is much higher than on smooth Pt and even Pt/CCE (curve red in Fig. 6). Furthermore,
Pt-Sn/CCE exhibits a high electrocatalysis performance in 2-butanol oxidation. Obviously, the anodic
currents densities of 2-butanol oxidation in PS are significantly enhanced after the introduction of Sn atoms
in the electrocatalyst. Additionally, it is noted that the onset potential of 2-butanol oxidation on the Pt-Sn/CCE
is even lower than that on the smooth Pt and Pt/CCE like as 2-propanol. These results indicate that the
addition of Sn into Pt catalysts can significantly improve the electrode performance for the 2-butanol
electrooxidation at a relatively less cost.

322
 In order to reveal the correlation between 2-butanol oxidation and Pt oxide species, we have studied
the effect of the upper limit potentials (EU) in the cyclic potential scanning on the 2-butanol oxidation. And
similar results as 2-propanol oxidation were obtained.

Fig. 6. CVs of 0.15 M 2-butanol at the Pt-Sn/CCE(blue line) and the Pt/CCE (red line) in 0.1 M H2SO4 at a scan rate of 50 mV s-1.

Finally the effect of some affecting parameters on the anodic peak currents of C1-C4 alcohols oxidation such
as the influence of electrodeposition method of Pt-Sn nanoparticles, effect of the ratio of metals in the alloy,
C1-C4 alcohols concentration and the long-term stability of the Pt-Sn nanoparticles were studied and
optimized.

4. CONCLUSION
Electrooxidation of C1-C4 alcohols was studied on the carbon-ceramic electrode modified bimetallic platinum-
tin nanoparticles as a potent electrocatalyst in acidic media. The results were compared with those at the
same electrode modified by platinum nanoparticles and smooth Pt electrode alone. It was found that Pt-Sn
electrocatalyst was catalytically more active than platinum nanoparticles on the same substrate and Pt
electrode. The high electrocatalytic activity of the Pt-Sn electrocatalyst towards the oxidation of C1-C4
alcohols may be directly related to the effect of the Sn atoms in electrocatalytic activity of the Pt and its large
electroactive surface area. Moreover, it contributes to the reduction of the amount of the noble metal in the
anode of DAFCs, which remains one of the challenges to make the technology of DAFCs economically
viable. Also, the Pt-Sn electrocatalyst exhibits satisfactory stability and reproducibility when stored in ambient
conditions and in constant potential applying or continuous cycling; which makes it attractive as the anode in
direct alcohol fuel cells (DAFCs) and the related applications.

ACKNOWLEDGMENT
The authors gratefully acknowledge the research council of Azarbaijan Shahid Madani University for financial
support.

REFERENCES
[1] Faraji M., Y. Yamini and M. Rezaee. 2010. Magnetic Nanoparticles: Synthesis, Stabilization, Functionalization,
Characterization, and Applications. Journal of Iranian Chemical Society 7:1-37.
[2] Guo S. and E. Wang. 2011. Noble metal nanomaterials: Controllable synthesis and application in fuel cell and
analytical sensors. Nanotechnology Today 6:240-264.
[3] Shudo T. and K. Suzuki. (2008). Performance improvement in direct methanol fuel cells using a highly porous
corrosion-resisting stainless steel flow field. International Journal of Hydrogen Energy 33:2850-2856.
[4] Habibi B. and N. Delnavaz. 2012, Electrosynthesis, characterization and electrocatalytic properties of Pt-Sn/CCE
towards oxidation of formic acid. RSC Advance 2:1609-1617.
[5] Habibi B. and N. Delnavaz. 2011. Electrocatalytic oxidation of formic acid and formaldehyde on platinum
nanoparticles decorated carbon-ceramic substrate. International Journal of Hydrogen Energy. 35: 8831-8839.
[6] Habibi B. and E. Dadashpour. 2013. Electrooxidation of 2-propanol and 2-butanol on the Pt-Ni alloy nanoparticles in
acidic media. Electrochimica Acta 88:157-164.
[7] Wang Z., G. Gao, H. Zhu, Z. Sun, H. Liu, and X. Zhao. 2009. Electrodeposition of platinum microparticle interface on
conducting polymer film modified nichrome for electrocatalytic oxidation of methanol. International Journal of
Hydrogen Energy.34:9334-9340.
[8] Becerõk I., Ş. Süzer, and F. Kadirgan. 2001. Electrooxidation of methanol on doped polypyrrole films in acidic media.
Journal of Electroanalytical Chemistry 502:118-125.

323
[9] Razmi H., Es. Habibi, and H. Heidari. 2008. Electrocatalytic oxidation of methanol and ethanol at carbon ceramic
electrode modified with platinum nanoparticles. Electrochimica Acta 53:8178-8185.
[10] Lin Y., X. Cui, and C.H.C.M. Yen. 2005. PtRu/Carbon Nanotube Nanocomposite Synthesized in Supercritical Fluid:
A Novel Electrocatalyst for Direct Methanol Fuel Cells. Langmuir 21:11474-11479.
[11] Zhao Y., E. Yifeng, L. Fan, Y. Qiu, and S. Yang. 2007. A new route for the electrodeposition of platinum-nickel alloy
nanoparticles on multi-walled carbon nanotubes. Electrochimica Acta 52:5873-5878.
[12] Angerstein-Kozlowska H., B.E. Conway, and W.B.A. Sharp. 1973. The real condition of electrochemically oxidized
platinum surfaces: Part I. Resolution of component processes. Journal of Electroanalytical Chemistry 43:9-36.
[13] Conway B.E., N. Marincic, D. Gilroy, and E.J. Rudd. 1966. Oxide Involvement in Some Anodic Oxidation Reactions.
Journal of Electroanalytical Chemistry Society 113:1144-1158.
[14] Pournaghi-Azar M.H. and B. Habibi. 2007. Electrocatalytic oxidation of methanol on poly(phenylenediamines) film
palladized aluminum electrodes, modified by Pt micro-particles: Comparison of permselectivity of the films for
methanol. Journal of Electroanalytical Chemistry 601:53-62.
[15] Gupta S.S. and J. Datta. 2006. A comparative study on ethanol oxidation behavior at Pt and PtRh electrodeposits.
Journal of Electroanalytical Chemistry 594:65-72.
[16] de Souza J.P.I., S.L. Queiroz, K. Bergamaski, E.R. Gonzalez, and F.C. Nart. 2002. Electro-Oxidation of Ethanol on
Pt, Rh, and PtRh Electrodes. A Study Using DEMS and in-Situ FTIR Techniques. Journal of Physical Chemistry
106:9825-9830.
[17] Iwasita T., B. Rasch, E. Cattaneo, and W. Vielstich. 1989. A sniftirs study of ethanol oxidation on platinum.
Electrochimica Acta34:1073-1079.
[18] Pournaghi-Azar M.H. and B.A. Habibi. 2005. Preparation of a platinum layer-modified aluminum electrode by
electrochemical and electroless cementations and its use for the electrooxidation of methanol. Journal of
Electroanalytical Chemistry 580:23-35.
[19] Mikhailova A.A., A.A. Pasynskii, V.A. Grinberg, Y.A. Velikodnyi, and O.A. Khazova. 2010. CO and methanol
oxidation at platinum-tin electrodes. Russian Journal of Electrochemistry 46:26-33.
[20] Otomo J., X. Li, T. Kobayashi, C.-j. Wen, H. Nagamoto, and H. Takahashi. 2004. AC-impedance spectroscopy of
anodic reactions with adsorbed intermediates: electro-oxidations of 2-propanol and methanol on carbon-supported
Pt catalyst. Journal of Electroanalytical Chemistry 573:99-109.
[21] Sun S.-G. and Y. Lin. 1996. In situ FTIR spectroscopic investigations of reaction mechanism of isopropanol oxidation
on platinum single crystal electrodes. Electrochimica. Acta 41:693-700.
[22] Vielstich W.. 1965. Fuel cells: Modern processes for the electrochemical production of energy (New Delhi: Wiley-
Interscience).
[23] Min M.-k., J. Cho, K. Cho, and H. Kim. 2000. Particle size and alloying effects of Pt-based alloy catalysts for fuel cell
applications. Electrochimica Acta, 45:4211-4217.
[24] Lamy C., E.M. Belgsir, and J.-M. Léger. 2001. Electrocatalytic oxidation of aliphatic alcohols: Application to the direct
alcohol fuel cell (DAFC). Journal of Applied Electrochemistry 31:799-809.
[25] Lee C.G., M. Umed, and I. Uchid. 2006. Cyclic voltammetric analysis of C1-C4 alcohol electrooxidations with Pt/C
and Pt-Ru/C microporous electrodes. Journal of Power Sources 160:78-92.

324
 PREPARATION AND CHARACTERIZATION OF NANOSIZED NiCo/Al2O3 CATALYSTS
PREPARED BY POLYOL METHOD FOR PARTIAL OXIDATION OF METHANE

*
Tuğçe Çınar, Tuba Gürkaynak Altınçekiç

Department of Chemical Engineering,Istanbul University,34320,Istanbul,Turkey

e-mail:tubag@istanbul.edu.tr

ABSTRACT
Small Ni-Co metal nanoparticles (< 10nm) supported on γ-Al2O3 were prepared by a simple polyol method in
ethylene glycol (EG) medium. The effects of Ni/Co ratio on the crystallite and particle sizes were determined by
XRD and TEM analysis. The activities of the catalysts were evaluated for partial oxidation of methane (CPOM).
The results indicated that in polyol process higher Ni/Co ratio showed smaller active bimetallic particle sizes and
good metal dispersion leading to high catalytic activity. The activity of the catalysts in CPOM reaction also
depended on the thermal treatment. H2-pretreated catalysts exhibited higher activity than non pre- treated
catalyst, suggesting that the presence of organic matrix hinders the access to the nickel-cobalt active sites. The
activities of the catalysts in terms of CH 4 conversion were found to be in the order of 10Ni/γ-Al2O3>5Ni5Co/γ-
Al2O3=7Ni3Co/γ-Al2O3>3Ni7Co/γ-Al2O3>10Co/ γ-Al2O3. 5Co5Ni exhibited the highest CH 4 conversion in partial
oxidation of methane. Among the catalysts tested, NiCo/ γ-Al2O3 catalysts with ratio of Ni/Co 7:3 and 5:5 exhibited
the higher activity (73%) and stability with a gas hour space velocity of 157,500 (Lkg-1h-1) at 800 ºC.

1. INTRODUCTION
In the development of Ni and Co based catalysts for partial oxidation of methane, the focus is always
directed on the effective suppression of carbon formation, or carbon deposition, on catalyst surface during
reaction (Ruchenstein et al., 2001; 2002; 2004) Although tremendous attempts have been made and a
number of papers have been published, carbon formation is still a great challenge for Ni and Co based
catalyst.
The main drawback of Ni catalysts is their deactivation by coke deposition as the reaction progress. Also
Co-based catalysts show higher resistance to carbon deposition , they have lower activities for the partial
oxidation or methane(Choudhary et al.,1998, Nishimoto et al.,2004, York et al.,2003) An attractive option of
a Ni-based catalyst for partial oxidation of methane is to use the Ni-Co bimetallic catalysts, in which
synergy between the two metals seems to take place (Zhang et al.,2007;2008)( San-Jose´ -Alonso et
al.,2009)
Alloying nickel with cobalt breaks the integrity of surface nickel ensembles, which decreases the apparent
nickel particle size to the extent, which is too small for carbon nucleation and growth (Petar Djinovi´c et al.,
2012). Additionally, formation of NiCo alloy is reported to improve catalysts stability by improving resistance
of active metal oxidation (Takanabe et al., 2005). Also recent reports revealed that bimetallic Co-Ni
catalysts had better catalytic performance than monometallic nickel or cobalt catalysts in methane reforming
where the Ni/Co ratio of the catalysts played an important role (Enger et al.2009; Chen et al., 2011;
Choudhary et al., 2006).

325
Chen et al. (2011) reported that the Co-Ni aerogel catalyst in methane oxy-reforming, 5Co5Ni exhibited the
best catalytic performance, while catalysts with other Co/Ni ratios were either less active or more prone to
coking.) Zhang et al. (2008) disclosed that Ni-Co-Al-Mg-O catalysts prepared using precipitation method
with lower Ni-Co content (1.83-3.61wt%for Ni and 2.76-4.53wt% for Co) had higher and more stable activity
with no deactivation. In contrast with those obtained by the authors Gonzalez et al. (2005) reported that in
bimetallic NiCo catalysts, the catalytic activity for methane dry reforming increases as the proportion of
cobalt decreases to 0.7.
These NiCo bimetallic catalysts can be synthesized from different methods (Wojcieszak et al.,2005 ; Wang
et al.,2009 Yu et al.,2013; Şahiner et al.,2011) among Which the polyol process has been proved to be
promising a versatile and an easy chemical reduction method for the preparation of nanoscaled catalysts.
In this method, the solid precursor is suspended in the liquid polyol, generally in ethylene glycol (EG),which
may be quite soluble (nitrate, chloride, acetate) or only slightly soluble(oxide, hydroxide). Then the solution
or suspension is stirred and heated to a given temperature which can reach the boiling point of the polyol in
order to reduce the metal ions (Fievet et al., 2013)
In our previous studies, we prepared Cu/ZnO catalysts for liquid phase hydroxylation of benzene to phenol
(Altincekiç et al.2011) and Ni/Al2O3 catalysts for partial oxidation of methane by the polyol method
(Bayrakdar et al., 2013). The Ni/Al2O3-5PVP catalysts (5.4%wt metal loading) obtained by this method
showed the highest activity and convert about 92% methane at 800ºC and were also quite stable for a long
time on stream with excess of methane. The results demonstrated that polyol method will be an alternative
method for nanosized catalyst preparation instead high temperature calcinations process requiring
traditional methods.
In this work, γ-Al2O3-supported Ni–Co bimetallic catalysts with different Ni/Co ratios were synthesized by
polyol method and characterized using X-Ray diffraction (XRD), High Resolution Transmission Electron
Microscopy (HRTEM), Atomic Absorption Spectroscopy (ICP-AES). The effect of the Ni/Co ratio on the
catalytic performance in partial oxidation of CH 4 to syngas at the high temperature 800ºC and atmospheric
pressure was investigated. The novelty of this work is characterized by a one-pot procedure which
combines the formation of nanoparticles, reduction and deposition onto support under mild conditions. We
hope that the views presented here will contribute to the progress in the field of research.

2. EXPERIMENTAL PROCEDURE
2.1. Synthesis of Catalysts
The preparation of the Alumina supported nickel cobalt bi-metallic particles was performed by a polyol
process. In this method an appropriate amount of alumina support, Co(Ac)2.4H2O and Ni(Ac) 2.4H2O were
mixed with 100 ml ethylene glycol at room temperature in a tree-necked flask under magnetic stirring. The
reaction flask was fitted with a reflux condenser and a thermometer for the control of the reaction
temperature. The Ni-Co concentration (0.05M) in the solution was calculated to obtain the total nominal
composition of 10 wt% Ni-Co. The dark grey suspension formed within 30min. and was kept at this
temperature for 2h and rapidly cooled down to room temperature with ice bath. After reduction gray solid
obtained was separated by centrifugation and washed several times with acetone and distilled water; then

326
dried at 90°C for 15-18 hr. Five different samples were prepared according to this procedure, varying the
[Ni]/[Co] ratio : (10/0), (7/3), (5/5), (3/7) and (0/10). Size and morphology of the Ni -Co nanoparticles were
examined on a high resolution transmission electron microscopy (HRTEM, JEOL, 2100 LaB6, Japa n). The
crystalline phase was identified by using an X-ray diffractometer (Rigaku, D/Max-2200, Japan) Cu/Kα1
radiation (λ=1.5405 Å). The average crystallite sizes of metallic Ni-Co was calculated by using the
Scherrer method. Actual Nickel-Cobalt loadings of the synthesized particles were obtained by using
Analyst 200 flame (Perkin Elmer) AAS equipment. The as-prepared catalysts were also tested for partial
oxidation of CH 4 to syngas in the high temperature 800ºC at atmospheric pressure. Before activity tes ts
the catalysts were pretreated at 450ºC for 4h under H 2 flow.
2.2. Catalyst Characterization
The catalysts were characterized by using Atomic Absorption Spectrophotometer (AAS), High Resolution
Transmission Electron Microscope (HRTEM) and X-ray diffraction (XRD). Actual Nickel-Cobalt loadings of
the prepared catalysts were obtained by using Thermo Elemental X series ICP-MS equipment. 50 mg of
each catalyst were dissolved in a strong acid mixture and the volumes of each sample were completed to
50 ml by using deionized water. Nickel-Cobalt particle sizes and dispersions of the catalysts were
investigated by using JEOL 2100 LaB6 HRTEM. Powder X-ray diffraction patterns of the as-prepared
catalysts were revealed by using Rigaku D/Max-2200 diffractometer, using a nickel filtered CuKα1 (λ=
0.15406 nm) radiation.
The step scans were taken over the range of 2θ angles from 10º to 90º. The average crystallite sizes of
metallic Ni-Co alloy was calculated by using the Scherrer method, where the constant K was taken as 0.9,
b was the FWHM of the Ni (200) at 51.7º.TPO experiments were carried out by using the Microreactor
(CATLAB)-MS(QIC-20) system (Hiden Analytical). TPO analysis was actualized to determine consisted
carbon species and deposition amounts on the catalysts after 10h stability tests. Experiments were carried
in a gas mixture of 10% O2/Ar (50 ml/min) over 10 mg of catalyst samples. Temperature was linearly
increased from 50 ºC to 900 ºC at 20 ºC/min CO2 formation was followed with on-line MS.
2.3. Catalytic Activity Measurements
The activity and selectivity measurements were conducted at atmospheric pressure in a quartz
microreactor (0.4 cm inner diameter and 19 cm length) equipped with a thermocouple in the centre of
catalyst bed. This system was integrated to the GC (HP 5890) for the experiments and integrated line was
heated with a heat tape to prevent the water condensation in the line. Before the effluent stream went into
the GC, it was cooled in a condenser to remove the formed water. Residue gas phase was analyzed on-
line by the GC equipped with TCD detector. Two columns, Porapak Q and Molecular Sieve 5A, were used
in a series/bypass arrangement for the complete separation of H2, N2, O2, CH4, CO and CO2. Then
reactant gas mixture (CH4/O2/N2=29/15/60) was fed to the reactor with the flow rate of 105 mL/min
-1 -1
(GHSV=157,500 L kg h ). Effluents were analyzed 800 ºC. During the activity tests, carbon balances
were always within 95-100%. The stability tests were carried out using excess methane in order to
accelerate the deactivation rate of the catalysts (CH4/O2/N2 = 58/15/58) at 800ºC for 10h. Effluents were

327
analyzed with 0.5h interval. Activity and selectivity results were calculated according to the equations
given below:
CH4 Conversion H2 Selectivity

CO Selectivity

3. RESULTS AND DISCUSSION

3.1. Metal Loading Efficiencies according to ICP-MS Analysis
In this study nickel acetate and cobalt acetate was chosen as the precursor to prepare the γ-Al2O3-supported

NiCo catalysts by polyol method. As is well known, the metal loading weight on the catalyst is an important
parameter for catalytic activity. In theory, activity enhanced by increasing the proportion of the active phase.
The ICP-MS Analysis of the Ni/γ-Al2O3 catalysts prepared by polyol method at different Ni/Co ratios are
shown in Table 1. It can be seen from Table 1 that the actual metal loading weight on γ-Al2O3 is lower than
the total nominal value (10 wt%) in all samples, whereas the actual Ni loading weights on γ-Al2O3 were 7.7,
3.8, 2.8 and 2 wt%, respectively.

Table1. ICP-MS Analysis Results for Ni and Co Loading

Catalyst % Ni % Co
(Loading Efficiency) %Ni (Theorical) (Loading Efficiency) % Co (Theorical)

10Ni/γ-Al2O3 7.7 10 - -

7Ni3Co/γ-Al2O3 3.8 7 1.5 3

5Ni5Co/γ-Al2O3 2.8 5 2.6 5

3Ni7Co/γ-Al2O3 2.0 3 3.8 7

10Co/γ-Al2O3 - - 6.8 10

3.2. X-Ray Diffraction (XRD) Results

XRD analysis results can explain the dispersion and crystal sizes of the nickel cobalt particles on γ-Al2O3.
However, it is difficult to distinguish Co or Ni or their alloys in the XRD patterns due to their similar structural
feature and close position of diffraction peaks. In terms of diffraction patterns of narrow range (42º-45º
Fig.1a-b), the diffraction peaks of Sample “a” at 2θ=44,4º, 51,8º, 76,3º are characteristic peaks of metallic
nickel with a fcc structure. The peaks appear at about 2θ=44,5º, 51,7º, 76,3º in samples “b”, “c”, “d”, which
are typical diffraction angles for NiCo alloy with a fcc structure. Sample “e” exhibited peaks at 2θ=44,3º and
76,3º, which are characteristic of metallic cobalt with a fcc structure. The peaks at 44.4º, 51.8º and 76.3º can
be assigned to (111), (200) and (220) crystalline plane diffraction, respectively. According to reduction
+2 +2
potentials of nickel and cobalt, the reduction of Ni is easier than Co so nickel nucleates first when the

328
reduction begins, then cobalt is reduced and precipitates with nickel on the performed nickel nuclei and the
nickel-cobalt alloy is produced ( Li et al.,1999) . This speculation is also supported by the d-spacing values of
the NiCo alloys. It was observed that the diffraction peak of the metal particles was in the intermediate
position between those attributable to metallic nickel (44.5º) and cobalt (44.3º) and the peak shifted to lower
angle with the increase of the Co/Ni ratio (Table 2), verifying the formation of Co-Ni metal alloy.
There are no peaks observed for Ni(OH)2, NiO, NiAl2O4 and Co3O4 among the synthesized catalysts,
indicating that the active metal species were totally reduced during the polyol process. The particles
crystallite sizes were estimated by the Scherrer’s law are shown in Table 3.

Table 2. Comparison of d-Spacing Values Corresponding to Three Major XRD Reflection Peaks of NiCo Alloys
a b
Co Ni
JPDS files 2.0462 1.7721 1.2530 2.0352 1.7625 1.2463
(hkl) (111) (200) (220) (111) (200) (220)
3Ni:7Co 2.0333 1.7698 1.2468
EG
5Ni:5Co 2.0346 1.7694 1.2456
EG
7Ni:3Co 2.0309 1.7659
EG
a b
From JCPDS files 89-4307(metallic Co) and files 4-0850 (metallic Ni)

Table3. Crystallite Sizes, Particle Sizes and Metal Dispersion

Catalyst 10Ni/Al2O3 7Ni3Co/Al2O3 5Ni5Co/Al2O3 3Ni7Co/Al2O3 10Co/Al2O3
Crystallite sizes by Scherrer Equation 13.10 8.74 7 7.6 11.05
(nm)
Average Particle Sizes from TEM 9 6.12 4.2 6.64 -
images (nm)
Metal dispersion by Eq.(1) (%) 7.3 10.8 15.7 10 -

16

14
Crystallite size(nm)

12

10

6
0/10 3/7 5/5 7/3 10/0
Ni/Co Ratio
Fig.1. Changes in NiCo crystallite sizes in relation to Ni/Co (w/w) ratio.

329
 
Ni or Co or NiCo alloy
* Al2O3

*

* 
a

20 40 60 80

2 (Degree)

(A)

Ni or Co or Ni-Co Alloy

b
c

d
e

42 43 44 45 46 47
2(Degree)

(B)
Fig.2. XRD patterns of the NiCo/ γ-Al2O3 catalysts with varying Ni/Co ratios (A)20-85º (B)42-47º. (a)10Ni,
(b)7Ni3Co, (c)5Ni5Co, (d)3Ni7Co, (e)10Co

Fig.3 shows the HRTEM images of the NiCo/γ-Al2O3 catalysts. The sample NiCo/γ-Al2O3 which the nickel
cobalt bimetallic particles (darker spots) have very small particle size and spherical. In these pictures, it is
observed that the Nickel cobalt bimetallic particles are loaded on the surface of the γ-Al2O3 support. For
the catalyst, it exhibits a uniform dispersion of NiCo-bimetallic particles with a size smaller than 10 nm in

330
diameter. Table 3 shows the average particle sizes of catalysts. With assumption that all the metals
reduced by polyol process are present on the surface of the alumina support in spherical particles of
corresponding sizes given by HRTEM, metal dispersion, DM, was calculated using the following equation
given by Vannice (2005)
(1)
3 2
Where VM is the bulk atomic volume of the metals (cm ), AM is the area of an atom (cm ), and d is the metal
-23 3
particle size in nm. For Ni and Co, their VM is very close to one another, so average 1.10 x10 cm is taken;
-15 2
and their AM is 1x 10 cm (Yu et al.,1995). The results of calculation are also given in Table 3.

24
7Ni3Co-Al2O3 7Ni3Co-Al2O3
22
20
18

Frequency (%) 16
14
12
10
8
6
4
2
0
2 3 4 5 6 7 8 9 10 11 12
Particle Size (nm)

20
5Ni5Co-Al2O3
5Ni5NCo-Al2O3
18

16

14
Frequency (%)

12

10

0
2 3 4 5 6 7
Particle Size (nm)

331
 3Ni7Co-Al2O3 24
%3Ni7Co-Al2O3
22
20
18
16

Frequency (%)
14
12
10
8
6
4
2
0
3 4 5 6 7 8 9 10 11 12
Particle size (nm)
Fig.3. HRTEM results of catalysts

3.3. Catalytic Activity, Selectivity and Stability Results

The catalytic activity, selectivity and stability tests of the as-synthesized catalysts were performed on GC-
Micro-reactor System. Before the catalytic activity tests, H2 pre-treatment was performed at 450ºC for 4h for
all catalysts. Boudjahem et.al (2004) prepared nickel nanoparticles supported on silica by hydrazine
reduction in aqueous medium for benzene hydrogenation and showed that the freshly prepared catalysts
are inactive without hydrogen pre-treatment. According to literature and our knowledge this is ascribed with
the presence of organic matrix incoming from the acetate fragment of the metal precursor, ethylene glycol
on the catalysts surface after the washing process. This matrix hinders the access of the nickel active sites.
After the hydrogen treatment, the organic matrix is removed and leads to the liberation of the metal active
sites, thus increasing the catalysts activity (Bayrakdar et al., 2013)

Table 4.Catalytic Activity, Selectivity and Stability Results

Catalyst CH4conversion H2 Selectivity CO Selectivity H2/CO

(%XCH4) (%SH2) (%SCO)
10Ni-Al2O3 79,5 91,2 88,3 2,0
7Ni3Co- Al2O3 73,0 89,1 85,8 2,0
5Ni5Co- Al2O3 73,4 78,3 85,0 1,84
3Ni7Co- Al2O3 67,6 86,2 82,1 2,0
10Co- Al2O3 25,7 1,7 1,9 1,7
−1 −1
Catalyst: 40 mg GHSV:157,500 Lkg h CH4:O2:N2=30:15:60 P:1 atm.

332
 100 3.5
10Ni/Al2O3
90 7Ni3Co/Al2O3
3.0
5Ni5Co/Al2O3
80 3Ni7Co/Al2O3
2.5
Methane Conversion (%)

10Co/Al2O3
70

60 2.0

H 2/CO Ratio
50 1.5

40
1.0
30 10Ni/Al2O3
7Ni3Co/Al2O3
20 0.5
5Ni5Co/Al2O3

10 3Ni7Co/Al2O3
0.0
10Co/Al2O3
0
600 650 700 750 800 600 650 700 750 800
o o
Temperature ( C) Temperature ( C)

100
100
90

80 80
70
CO Selectivity (%)

60
H 2 Selectivity (%)

60

50

40 10Co/Al2O3 40
10Ni/Al2O3
30 7Ni3Co/Al2O3 7Ni3Co/Al2O3
5Ni5Co/Al2O3 5Ni5Co/Al2O3
20 20
3Ni7Co/Al2O3 3Ni7Co/Al2O3
10 10Co/Al2O3 10Co/Al2O3

0 0

600 650 700 750 800 600 650 700 750 800
o o
Temperature ( C) Temperature ( C)

Fig.4. Activity, selectivity and H2/CO results of pre-treated catalysts between 600 and 800 °C. Catalyst: 40 mg
GHSV:157,500 Lkg−1 h−1 CH4:O2:N2=30:15:60 P:1 atm

60
5Ni5Co/Al2O3
7Ni3Co/Al2O3
50
10 Ni/Al2O3
% Methane Conversion

40

30

20

10

0
0 2 4 6 8 10
Time
Fig.5. Activity results of the catalysts after they exposed to stability test Catalyst:
−1 −1
40 mg GHSV: 157,500 Lkg h CH4:O2:N2=47:12:47 P: 1 atm. at 800 °C.

333
Table.5. Amount of deposited carbon and related coking rate after 10h stability tests
a
CH4 CH4 Deactivation Carbon deposition Average coking rate
conv. conv. amount
Catalysts (%) (g carbon/g catalyst h)
(90 min.) (6 h) (g carbon/g catalyst)

10Ni/ɣ-Al2O3 45,21 40,22 11 0,59 0,059

7Ni3Co/ɣ-Al2O3 43,32 41,7 3,73 0,62 0,062

-4
5Ni5Co/ɣ-Al2O3 33,1 33,1 0 0.002 2x10

Catalyst: 50 mg G HSV: 159,700 l kg−1 h−1, CH4:O2:N2=58:15:58, P:1 atm, T:800 °C, t:10 h.
a
Calculated as the difference between CH4 conversion at30 min. and 6 h, divided by conversion at 30 min and multiplied by 100

The effects of Cobalt addition on the performance of 10Ni/ɣ-Al2O3 in methane catalytic partial oxidation
(CPOM) were tested between 600-800°C and the results are shown in Fig.4. It is known that compared
with cobalt, nickel is very active component in methane catalytic partial oxidation. 10Ni/ɣ-Al2O3 catalyst give
79.5% CH4 conversion for 7.7 %( wt) actual Ni loading. In contrast 10Co/ɣ-Al2O3 catalyst prepared by polyol
method gives very low CH4 conversion for a 6.8% Co loading. We observed that the methane conversion
and yield of hydrogen at 800ºC are three times higher for 10Ni/ɣ-Al2O3 than for 10Co/ɣ-Al2O3 .The activities
in terms of CH4 conversion followed the order of 10Ni/ɣ-Al2O3 >5Ni5Co/ɣ-Al2O3 =7Ni3Co/ɣ-Al2O3 >3Ni7Co/ɣ-
Al2O3 >10Co/ɣ-Al2O3. It was reported that for the same bimetallic catalyst system, catalysts with higher
nickel contents usually had higher catalytic activities in methane partial oxidation (Koh et al., 2007). Our
results indicate that the activities of the Ni-Co bimetallic catalysts prepared by polyol method exhibit a
similar trend in methane catalytic partial oxidation the present study, the Co/Ni ratio showed a great
influence on the micro-structure of the catalysts, which might be one of the reasons for the varied catalytic
activities. More small particles mean more metal-support interfacial structures, thus more active sites
(Zhang et al.2008). TEM analyses revealed that the portion of smaller metal particles and the metal
dispersion increased when the Co/Ni ratio is 1, which should also be account for the varied activities of the
Ni-Co bimetallic catalysts. In addition, the poor activity of 10Co (25.7%) was mainly attributed to the
oxidation of cobalt, because cobalt is susceptible to oxidation in methane partial oxidation especially with
oxygen. Table 5 shows CH4 conversion at 90min. and after 10 h and the percentage of catalyst deactivation
(calculated as the difference between CH4 conversion at 90 min and 10 h, divided by conversion at 90 min
and multiplied by 100). The amount of carbon deposited after 10 h time on stream (in g of carbon per g of
catalyst, determined by TPO experiments) is also included in Table 5. It can be observed that deactivation
is generally same and is more significant for 10Ni/Al2O3 and 7Ni3Co/ɣ-Al2O3 catalysts with higher Ni
content. Surprisingly, the methane conversion of sample 5Ni5Co/ɣ-Al2O3 is lower than the 10Ni/ɣ-Al2O3,
7Ni3Co/ɣ-Al2O3 catalysts in stability tests. Despite the fact that high temperature and excess methane in the
reactant mixture were used for the catalysts, no carbon deposition was observed during 10 h reaction for
5Ni5Co/Al2O3 catalysts. This result is consistent with the previous report (Wang et al.,2001) that the Ni-Co
bimetallic catalysts is highly resistant to coking as compared to monometallic Ni catalysts which indicates
that the coexistence of Ni and Co will benefit each other.

334
CONCLUSIONS
Alumina supported Ni-Co bimetallic catalysts are prepared by a facile polyol method without using a
protective agent instead high temperature calcination processes requiring traditional methods for the first
time. Ni/Co metal alloys formed with a particle size in the range of 6.6-4.2 nm during the polyol process. CH4
conversion and H2/CO selectivity are notably changed with the variation in Ni/Co ratios. Among the Ni/Co
ratios studied, Ni/Co=1 is considered as the optimal candidate. The optimal loadings of Ni and Co are
2.8wt% and 2.6wt%. In our case when the Ni/Co ratio reached 1, the metal particle size is smaller than the
critical size, leading to a strong coke resistance. This can be mainly attributed to the synergic effect between
the two metals, the smaller metal particle size and stronger metal-support interaction of the catalyst with total
Ni-Co metal loading lower 5wt%.
Acknowledgements
This study was supported by The Scientific and Technological Research Council of Turkey, Engineering
Research Grant Committee (MAG) through project No: 110M446 and by Istanbul University Research Fund
through project No: 3642 and 10041.

REFERENCES
Bayrakdar E., Gürkaynak Altınçekiç T., Öksüzömer M.A.F,2013, Effects Of PVP On The Preparation Of
Nanosized Al2O3 Supported Ni Catalysts By Polyol Method For Catalytic Partial Oxidation Of Methane, Fuel
Processing Technology 110 167–175

Boudjahem,A.G, Monteverdi, S., Mercy, M., Bettahar, M.M., 2004, Nanonickel Particles Supported on
Silica.Morphology Effects on Their Surface And Hydrogenating Properties. Catalysis Letters Vol.97,(3-4) ,
177-183

Chen L., Qingshan Zhu, Rongfang Wu, 2011, Effect of Co-Ni Ratio on The Activity And Stability Of Co-Ni
Bimetallic Aerogel Catalyst For Methane Oxy-CO2 Reforming, International Journal of Hydrogen Energy, 36,
2128-2136

Choudhary V.R. , A.M. Rajput, B. Prabhakar, A.S. Mamman, 1998, Partial Oxidation of Methane to CO and
H2 Over nickel and/or Cobalt Containing ZrO2, ThO2, UO2, TiO2 and SiO2 Catalysts, Fuel, 77,15, 1803-1807

Choudhary VR, Mondal KC, Choudhary TV. , 2005, Methane Reforming Over a High Temperature Stable-
NiCoMgOx Supported On Zirconia-Hafnia Catalyst, Chemical Engineering Journal; 121:73-7.

Enger B.C., Rune Lødeng, Anders Holmen, 2009, Modified cobalt catalysts in the partial oxidation of
methane at moderate temperatures, Journal of Catalysis 262, 188–198

Enger, B.C., Lodeng, R., Holmen, A., 2008, A review of catalytic partial oxidation of methane to synthesis
gas with emphasis on reaction mechanisms over transition metal catalysts, Applied Catalysis A: General,
346, 1-27

Fievet F.,Brayner R., Nanometarials: A Danger or Promise? Chapter 1, The Polyol Process,Springer-Verlag
London 2013

Gonza´Lez O., Juan Lujan, Egle´ Pietri , Mireya R. Goldwasser , 2005, New Co-Ni catalyst systems used for
methane dry reforming based on supported catalysts over an INT-MM1 mesoporous material and a
perovskite-like oxide precursor LaCo0.4Ni0.6O3, Catalysis Today, 107–108, 436–443

Gurkaynak Altıncekic T., I. Boz, 2011, Liquid phase hydroxylation of benzene to phenol over Cu/ZnO
catalysts, Reaction Kinetics, Mechanisms and Catalysis, 102, 195–205

335
Li Y.D., L.Q.Li, H.W.Liao and H.R.Wang, 1999, Preparation of pure nickel,cobalt and nickel-copper alloys by
hydrothermal reduction, Journal of Material Chemical, 9, 2675-2677
Zhang J, Wang H, Dalai AK., 2008, Effects of metal content on activity and stability of Ni Co bimetallic
catalysts for CO2 reforming of CH4, Applied Catalysis A ; 339:121e9.

Koh ACW, Chen L, Leong WK, Johnson BFG, Khimyak T, Lin J., Hydrogen or synthesis gas production via
the partial oxidation of methane over supported nickel-cobalt catalysts. Int J Hydrogen Energy 2007;
32:725e30.

Nishimoto H A, Nakagawa K, Ikenaga N O, Nishitani‐Gamo M, Ando T, Suzuki T. ,2004, Partial oxidation of

methane to synthesis gas over oxidized diamond catalysts, Applied Catalysis A: General, 264, 1, 65-72

Petar Djinovi´C ,Ilja Gasan ,Osojnik ˇCrnivec ,Boˇstjan Erjavec, Albin Pintar, 2012, Influence of active Metal
loading and oxygen mobility on coke-free dry reforming of Ni–Co bimetallic catalysts, Applied Catalysis B:
Environmental, 125, 259–270

Ruckhenstein E., Y.H. Hu, 2002, Binary MgO-Based Solid Solution Catalysts for Methane Conversion To
Syngas, Catalysis Reviews, 44, 3, 423-453

Ruckenstein E., Wang H.Y., 2001, Partial Oxidation of Methane to Synthesis Gas over Alkaline Earth Metal
Oxide Supported Cobalt Catalysts, Journal of Catalysis, 199, 2, 309-317

Ruckenstein E., Y. H. Hu, 2004, Catalytic Conversion of Methane to Synthesis Gas by Partial Oxidation and
CO2 reforming, Advances in Catalysis, Volume 48, Pages 297-345

Sahiner N., Ozgur Ozay, Nahit Aktas, Erk Inger, Jibao He, 2011, The on demand generation of hydrogen
from Co-Ni bimetallic nano catalyst prepared by dual use of hydrogel: As template and as reactor
international journal of hydrogen energy, 36 ,15250-15258

San-Jose´ -Alonso D., J. Juan-Juan, M.J. Illa´n-Go´ mez, M.C. Roma´ n-Martı´nez , 2009, Ni, Co and
bimetallic Ni–Co catalysts for the dry reforming of methane, Applied Catalysis A: General , 371 ,54–59

Takanabe K.,K.Nagaoka,K.Nariai,K.-I.Aika, 2005, Titania-supported cobalt and nickel bimetallic catalysts for
carbon dioxide reforming of methane, Journal of Catalysis, 232, 2, 268-275

Vannice M.A, 2005, Kinetics of Catalytic Reactions, Springer, New York, 23–33.

Wang F., Yong Li, Weijie Cai, Ensheng Zhan, Xiaoling Mu, Wenjie Shen, 2009, Ethanol steam reforming
over Ni and Ni–Cu catalysts, Catalysis Today, 146, 31–36

Wang H.Y, RUCKENSTEIN E., 2001, Conversions of Methane to Synthesis Gas over Co/γ-Al2O3 by CO2
and/or O2 Catalysis Letters, 75 (1-2):13

Wojcieszak J.R., S. Monteverdi, M. Ziolek, M.M. Bettahar, 2005, Study of nickel catalysts supported on
Al2O3, SiO2 or Nb2O5 oxides, Journal of Molecular Catalysis A: Chemical 242, 81-90

York A. P. E., T. Xioa, and M. L. H. Green, 2003, Brief overview of the partial oxidation of methane to
synthesis gas, Topics in Catalysis, 22, 345-358.

Yu C. C., S. T. Oyama, 1995, Synthesis and characterization of new bimetallic transition metal oxynitrides:
MI-MII-O-N (MI, MII=V, Mo, W and Nb), Journal of Materials Science, 30, 16, 4037-4042

Yue Yu, Guoqiang Jin, Yingyong Wang, XiangYun Guo, 2013, Synthesis of natural gas from CO
methanation over SiC supported Ni–Co bimetallic catalysts, Catalysis Communications, 31, 5–10

Zhang J., Hui Wang, Ajay K. Dalai, 2007, Development of stable bimetallic catalysts for carbon dioxide
reforming of methane, Journal of Catalysis, 249, 2, 300-310

Zhang J., Hui Wang, Ajay K. Dalai, 2008, Effects of metal content on activity and stability of Ni-Co bimetallic
catalysts for CO2 reforming of CH4 Applied Catalysis A: General, 339, 2, 121-129

336
 CAPACITIVE PROPERTIES OF POLYPYRROLE/CARBON NANOTUBE COMPOSITES

A. Bayrakceken*, H. Bayrakceken, J. Naktiyok, and A. K. Ozer

Atatürk University, Faculty of Engineering, Department of Chemical Engineering, 25240, Erzurum, Turkey
ayse.bayrakceken@gmail.com, hatbay@atauni.edu.tr, jalenaktiyok@atauni.edu.tr, kadirozer@atauni.edu.tr

ABSTRACT
In this study, polypyrrole (PPy)/carbon nanotube (CNT) composites were synthesized by in situ chemical oxidative
polymerization of pyrrole monomer on CNT. A dopant named paratoluenesulfonic acid (p-TSA) was added during the
synthesis. Two different types of CNT having different properties were used. The composites were characterized by
using BET surface area analysis, FTIR, XRD and TGA techniques. The kinetics of thermal decomposition of PPy/CNT
composites were studied by means of thermal analysis techniques (TG/DTG) in nitrogen atmosphere at heating rates
of 2.5, 5, 7.5, and 10 K min-1. Kinetic parameters were determined from the TG and DTG curves for the composites by
using Kissinger–Akahira–Sunose (KAS) and Flynn-Wall-Ozawa (FWO). The kinetic triplets consisting of Ea, A and
g(α) model of the composites were determined. The electrochemical capacitive properties of the composites were
investigated by cyclic voltammetry (CV) technique.

INTRODUCTION
Carbon nanotubes (CNTs), after their invention, have attracted increasing attention in recent years because of their
excellent electrical and mechanical properties. These properties make them as a promising candidate utilized in the
fields of energy storage and energy conversion devices, supercapacitors, sensors, hydrogen storage, heterogeneous
catalysts and semiconductor devices (Fernández et al., 2009). The conducting polymers, and also the polypyrrole
(PPy) is gaining an increasing attention because of its excellent electrical properties obtained by doping that can be
used as the electrode material in supercapacitors (Zhang et al., 2009).

The increasing energy demand leads the energy vicinities to search for new energy storage devices having high
power density and long life cycle for portable electronic systems such as supercapacitors. Commonly used materials
used in supercapacitors are: i) high surface area carbon materials ii) redox metal oxides and iii) conducting polymers.
The main problem related with the utilization of the conducting polymers in supercapacitors is the poor cycle life and
this drawback can be eliminated with the addition of high surface area carbon material.

Mechanical and electrochemical properties of the PPy conducting polymer can be improved with the addition of CNTs
for supercapacitors. CNTs are of great interest due to their high surface area, high electronic conductivity, chemical
and mechanical stability. CNTs in these composites serve as the perfect backbone for the homogeneous distribution
of the PPy material in the matrix (An et al., 2002).

The thermal decomposition of the composites at higher temperatures can be analysed with TGA method. The TG-
DTG analysis of the composites offers the possibility of correlating the kinetic parameters of non-isothermal
degradation. There are different methods to analyze the non-isothermal kinetic data from TGA. According to the
mathematical model, there are two possible approaches: model-fitting and model-free (isoconversional) methods (P.
Simon, 2004). Model fitting methods use different models in order to choose a model that provides the best statistical
approximation during the calculation of the kinetic parameters. Model-free, non-isothermal methods require several
kinetic curves to perform the analysis. Calculations from several curves at different heating rates are performed on the
same value of conversion, which allows calculating the activation energy for each conversion point. The advantage of
the model-free analysis is found on its simplicity and on the avoidance of errors connected with the choice of a kinetic
model. In non-isothermal kinetics (model-free analysis), Kissinger-Akahira-Sunose (KAS) (H. Kissinegr, 1956), Flynn-
Wall-Ozawa (FWO) (Flynn and Wall, 1966) methods are the most popular representatives of the isoconversional
methods.

In this study, two different types of the CNTs were used and polypyrrole was grown over these supports by in situ
oxidative chemical polymerization. Then they were characterized by BET, FTIR, XRD, TEM and TGA techniques. The
decomposition kinetics of the composites were investigated by using the TGA apparatus at different heating rates and
the activation energy (Ea) is calculated by isoconversional methods. KAS and FWO are used to evaluate the value of
Ea at different values of α. The kinetic parameters determined by both models were compared and the mechanisms
for the decomposition reactions were predicted by the method suggested by (Ortega, 1996). Capacitive properties
were investigated by using cyclic voltammetry (CV) technique.

EXPERIMENTAL
Two different types of CNT were used, CNT (S) with a diameter of 9.5 nm and length of 1.5 µm (90% purity), CNT (P)
with a diameter of 6-9 nm and with a length of 5 µm (95% purity). CNT (S or P) PPy composites were synthesized by
in situ chemical oxidative polymerization of pyrrole monomer on carbon black at room temperature. Composites were
synthesized like the procedure given elsewhere (Memioğlu et al., 2012). Briefly, CNTs were treated with nitric acid and
hydrochloric acid. Required amounts of the CNTs were taken and then mixed with ethanol. After stirring for half an
hour distilled pyrrole solution was added to this mixture and stirred for 30 min further. At this step, the dopant

337
p-toluenesulfonic acid ( p-TSA) was added. Ammonium persulphate (APS) as the oxidant was added to this mixture
and stirred for 4 h. After stirring, the mixture was filtered, washed with ethanol, distilled water and acetone,
respectively.

Synthesized composites were characterized by using BET, FTIR, XRD, TGA and TEM techniques. Nitrogen
adsorption-desorption analyses were carried out by a Quantachrome Autosorb-1-C/MS instrument. Multipoint BET
surface area, pore size distribution and pore volume of the composites were analyzed by this technique. The IR
spectra of the materials were recorded in Perkin Elmer Spectrum One FT-IR spectrometer. XRD data of the samples
were obtained by a Rigaku Miniflex X-ray diffractometer. The kinetics of thermal decomposition of PPy/CNT
composites were studied by means of thermal analysis techniques (TG/DTG) in nitrogen atmosphere at heating rates
of 2.5, 5, 7.5, and 10 K min-1.JEOL JEM 2100F STEM instrument was used for TEM analysis.

Theoretical Models
An important step in the kinetic analysis is the knowledge of the dependence of activation energy, E, versus the
conversion degree, α. This allows the detection of multistep processes and gives information about the reaction
mechanism.

In thermogravimetric measurements, the degree of decomposition (conversion) can be calculated as follows:

where, α is extent of decomposition; W0, Wt, and W∞ are the initial, actual, and final mass of the sample,
respectively.
The rate of reaction of a substance is usually expressed as

where α represents the fraction of substance reacted at time t, f(α) is a function depending on the reaction
mechanism, and k is the specific rate constant given by the Arrhenius equation:

where A is the pre-exponential factor in min-1, E is the activation energy (kJ/mol), R is the universal gas constant
(J/mol/K), and T is thü1e absolute temperature (K). For a non-isothermal system, a constant heating rate β is applied
where

From Eq. 2, we get

Substituting value of dT/dt from Eq. 4 and the value of K from Eq. 3 into Eq. 5 and rearranging, we get

Ordinarily, at low temperatures, the rate of the reaction is assumed to be low enough to be neglected, and thus the left
and right-hand sides of Eq. 6 may be integrated from 0 to α and from 0 to T:

where g(α) is a new function and on the right-hand side of Eq. 7 is called the temperature
integral, which does not have an analytical solution.

If it is assumed that x = E/RT, then reduces to where P(x) is an

infinite function of x. Hence, substituting the values of temperature integral (P(x)) in Eq. 7 gives:

338
As the value of E/R is constant; the evaluation of is dependent on evaluating the function P(x). Doyle (1962)
has evaluated P(x) and suggested its value to be log(P(x)) = -2.315 - 0.4567x over the range of 20 ≤ x ≤ 60.
Kissinger–Akahira–Sunose (KAS) method

The Kissinger–Akahira–Sunose (KAS) method is obtained by differentiating Eq. 6 and rearranging the terms as
follows:

The activation energy, E, is determined from the slope of the plot of logarithm of heating rate, versus the inverse of
temperature, 1000/T, at the given α, in constant heating rate experiments.
Flynn–Wall–Ozawa (FWO) method

Using the Doyle [29] approximation and the logarithmic form of Eq. 7; the linear equation of Flynn–Wall–Ozawa
(FWO) method [19-20] is obtained:

On changing logarithm to the base 10 (log) to natural logarithm (ln), Eq. 9 can be rewritten as:

Considering small intervals and the same extent of reaction (a) for different heating rates; activation energy, E, is
calculated from the slope of lnβ versus 1000/T plot.

The mechanism of the reactions is predicted from the method suggested by Ortega [27]. Firstly, the activation energy
for the decomposition of the compoistes is calculated using Eq. 9, i.e. the FWO method, by plotting lnβ versus 1000/T,
and recording the slope. Next, in order to use this value of activation energy to predict the reaction mechanism, the
following procedure is adopted. Eq. 10 is rearranged to give:

The common forms of g(α) of some important mechanisms for solid state reactions are tabulated [27]. The
expressions for g(α) are inserted, one at a time, in Eq. 12; and by plotting ln(g(α)) versus 1000/T, for different β, the
value of activation energy for the chosen g(α) is calculated from the slope. Thus, different values of activation energy
are obtained for different possible mechanisms (g(α)). These values of activation energy are compared with the one
obtained by FWO method for the decomposition of the composites, to find the mechanisms of the decomposition
reactions for both material. Once the mechanism g(α) is known, then its value for each α is substituted in Eq. 12 along
with the value of activation energy to get the value of ln(A).

Electrochemical Characterization
The CV set up used was Pine Instrument combined with the Versastat 3 potentiostat. CV analyses were carried out in
a standard three electrode electrochemical cell system including working (glassy carbon; GC), reference (Ag/AgCl)
and counter (Pt wire) electrodes. The electrolyte was 1 M H2SO4. CNT or synthesized composites with deionized
water, 1,2-propandiol, and 20% Nafion solution (Aldrich) were mixed to be used as the electrode ink. The resulting
suspension was homogenized with a homogenizer (Ultraturrax T18). Before passing through the CV experiments, the
working electrode was put into a concentrated HNO3 solution in order to clear any impurities. Electrolyte solution was
saturated with nitrogen for 30 min to remove the oxygen before the experiments and the CV results were obtained
after 20 cycle with a scan rate of 50 mV/s in between 0-1 V vs NHE.

RESULTS AND DISCUSSION

Figure 1 (a) shows the liquid N2 adsorption isotherms of the different CNTs. The (raw) CNT samples adsorb much
more nitrogen, indicating a high specific surface area (BET) and high pore volume. It was observed from figure that
the nitrogen adsorbed by the CNT samples doesn’t change at the low values of P/P0, suggesting the presence of
scarcity micropores. On the other hand, a gradual uptake of N2 was observed at the intermediate P/P0, which
indicated the adsorption of N2 on the external surfaces of the CNT(S), CNT(P), and the CNT(S) PPy and CNT(P) PPy
composites. The isotherms show that the external surface area of the CNT(S) was greater than the CNT(S) PPy. The
isotherm of the CNT(S) is close to the CNT(S) PPy. The appearance of hysteresis in the N2 isotherms in the high P/P0
range indicates that the adsorption with polypyrrole had closed the micropores and declined the mesopores. It is seen
that the isotherms of the CNT(S) and CNT(P) samples show very little hysteresis. The structural properties of the
samples are summarized in Table 1. The addition of polypyrrole on the CNT samples causes the closure of the pores.

339
The pore size distribution (PSD) is an important parameter for porous materials. For mesopore region, the BJH
equation is conventionally used to calculate the PSD. BJH (for 2–50 nm region) methods was used to obtain PSDs of
the CNTs from the adsorption branch of the isotherm, and the results are shown in Figure 1 (b). The PSD of the
CNT(S) shows a peak in the micropore region at 0.92-2.54 nm, and a main peak in mesopore region at 4.57-63.01
nm. CNT(S)PPy sample has a main peak in mesopore region at 7.49-59.83 nm. It is seen that the PSD of the CNT(P)
is a peak in the micropore region at 0-3 nm, and a main peak in mesopore region at 5-48 nm.

1400
CNT (P) CNT (P) PPy
1200 CNT (S) CNT (S) PPy
Quantity adsorbed, cm3/g STP

1000

800

600

400

200

0.0 0.2 0.4 0.6 0.8 1.0

Relative Pressure, P/P0

(a)

(b)

Fig. 1. a) Nitrogen adsorption isotherms and b) Pore size distribution of the CNT samples

Table 1. The pore properties of CNT samples

SBET (m2/g) Pore Volume The Micropore Area (m2/g) The Micropore Volume (cm3/g)
(cm3/g)

CNT(S) 242.106 2.06649 28.2722 0.011205

CNT(S)PPy 223.384 1.73682 6.1699 0.000189
CNT(P) 275.916 2.04807 11.2543 0.001950
CNT(P)PPy 269.457 1.99167 10.8633 0.001966

340
FTIR results of the samples are given in Figure 2. CNT (S) and CNT (P) showed different characteristic which
indicates different functional groups over the materials. A clear peak formation was observed for the CNTs having PPy
at about 1450 cm-1 which shows the ring breathing with contribution from C=C/C-C and C-N (Turcu et al., 2006).

CNT (S) PPy

CNT (P) PPy

T (%), au

T (%), au
CNT (P)

CNT (S)

4000 3000 2000 1000 0 4000 3000 2000 1000 0

Wave number, cm-1 Wave number, cm-1

Fig.2. FTIR results of the samples.

XRD results of the catalysts are given in Figure 3. The The diffraction peaks at 26.5°, 42.4°, 54.7°, and 77.4°
observed in the diffraction of MWNTs can be attributed to the hexagonal graphite structures (002), (100), (004), and
(110) (Li et al., 2003).

CNT (S) PPy

Intensity, au

CNT (S)

CNT (P) PPy

CNT (P)

0 20 40 60 80 100
2 Theta, 0

Fig.3. XRD results of the catalysts.

TGA is used in order to determine the thermal behavior of the composite materials. Fig. 4 shows the TG curves
corresponding to experiments carried out at heating rates of 2.5, 5, 7.5 and 10 K/min under nitrogen atmosphere for
the decomposition of the CNT composites prepared by CNT. Fig. 4 indicates that the apparent decomposition
temperatures increase with the increase in heating rates. This can be explained on the basis of heat transfer
limitation. During the analysis, at low heating rate, a larger instantaneous thermal energy is provided to the system
and a longer time may be required for the purge gas to reach equilibrium with the temperature of the furnace of the
sample. While at the same time and in the same temperature region a higher heating rate has a short reaction time
and therefore the temperature needed for the sample to decompose is also higher. This causes the maximum rate
curve to shift to the right (30). TG measurements indicated that thermal behavior of CNT (S) PPy and CNT (P) PPy
has three-stage degradation. At the first stage, the thermal decomposition was occurred due to the mass loss of
moisture and of light volatile compounds at temperatures ranging from 383 to 523 K according to the heating rates.

341
The second stage proceeds from 533 K to the final temperature of 873 K, in which the main decomposition process
takes places. Most of the organic materials (hydrocarbon material and the evolution of gases and oil vapor) are
decomposed in this stage.

120
CNT (S) PPy 2.5 K/min 5 K/min
7.5 K/min 10 K/min
100

80
Mass Loss, %

60

40

20

200 400 600 800 1000 1200

Temperature, K
(a)

120
CNT (P) PPy 2.5 K/min 5 K/min
7.5 K/min 10 K/min
100

80
Mass Loss, %

60

40

20

0
200 400 600 800 1000 1200
Temperature, K
(b)
Fig.4. TGA results of the samples a) CNT (S) PPy b) CNT (P) PPy

Kinetic analysis
TG study consists of performing a kinetic analysis which includes mass loss curves obtained at different heating rates
in order to deduce the dependence of the kinetic parameters on conversion. The values of fraction converted (α) at
each temperature for CNT (S) PPy and CNT (P) PPy were calculated from the TG data. These values were used in
the FWO and KAS methods for calculating the kinetic parameters. Using the KAS method (Eq. 9), the activation
energies were calculated from the slope of the plot of ln (β/T2) versus 1000/T (Figure 5) in between 0.1–0.9 values of
fraction converted (α), with intervals of 0.1 for CNT (S) PPy and CNT (P) PPy. The value of the slope was equal to –
E/R. The calculated activation energies (Ea) and correlation coefficients (R2) are shown in Table 2. The activation
energies were also determined using the method of FWO, with Eq. 10. Here, a linear fit curve was obtained for ln β
versus 1000/T (Figure 6) at different fraction converted values. The same fraction converted (α) values was used in
the FWO method for CNT (S) PPy and CNT (P) PPy.

342
 (b)
(a)

Fig.5. Decomposition peak temperatures variation with heating rates at given conversion fraction (α) values by KAS method for decomposition of a) CNT
(S) PPy b) CNT (P) PPy

2.4 2.4
0.1 0.1
2.2 2.2 0.2
0.2
0.3 0.3
2.0 2.0 0.4
0.4
0.5 0.5
1.8 1.8
0.6 0.6
ln (

ln (

1.6 0.7 1.6 0.7

0.8 0.8
1.4 0.9 1.4 0.9

1.2 1.2

1.0 1.0

0.8 0.8
1.00 1.05 1.10 1.15 1.20 1.25 1.02 1.04 1.06 1.08 1.10 1.12 1.14 1.16
1000/T
1000/T
(a) (b)

Fig.6. Decomposition peak temperatures variation with heating rates at given conversion fraction (α) values by KAS method for decomposition of a) CNT
(S) PPy b) CNT (P) PPy

Table 2. Activation Energies (Ea) and correlation coefficient (r2) calculated by KAS and FWO methods for the decomposition of CNT (S) PPy and
CNT (P) PPy
CNT (S) PPy CNT (P) PPy

α KAS method FWO method KAS method FWO method

Ea/kJ mol-1 r2 Ea/kJ mol-1 r2 Ea/kJ mol-1 r2 Ea/kJ mol-1 r2
0.1 119.07±0.24 0.9972 133.38±0.24 0.9978 250.99±0.02 0.9998 265.75±0.02 0.9998
0.2 111.91±0.30 0.9965 126.37±0.31 0.9971 232.46±0.76 0.9921 247.41±0.75 0.9931
0.3 106.56±0.31 0.9960 121.21±0.33 0.9970 214.94±0.78 0.9918 230.00±0.77 0.9929
0.4 100.39±0.15 0.9982 115.19±0.44 0.9985 204.11±0.53 0.9943 219.28±0.52 0.9952
0.5 97.42±0.03 0.9995 112.37±0.04 0.9996 188.19±1.78 0.9857 203.49±1.32 0.9879
0.6 92.46±0.08 0.9999 107.55±0.08 0.9993 184.06±0.35 0.9962 199.46±0.34 0.9968
0.7 93.39±0.32 0.9960 108.66±0.30 0.9972 177.95±0.03 0.9997 193.42±0.03 0.9997
0.8 88.38±0.42 0.9947 103.76±0.44 0.9960 177.17±0.11 0.9988 192.74±0.11 0.9990
0.9 89.72±0.03 0.9996 105.25±0.02 0.9998 171.69±0.24 0.9973 187.38±0.24 0.9978
Av 99,92±0.24 0.9975 114.86±0.24 0.9972 200.17±0.51 0.9951 215.43±0.48 0.9958
Av: Average

Activation energy of decomposition of CNT (S) PPy and CNT (P) PPy composites
The linear fit plots were nearly parallel, which indicates that by using the KAS and FWO methods, unique values of
the kinetic parameters can be calculated for the decomposition of CNT (S) PPy and CNT (P) PPy. Moreover, as the
variation of calculated activation energy with conversion fraction, α (0.1–0.9), is fairly small. Some of the differences
observed between the values of Ea obtained using various linear integral methods (KAS and FWO) can be assigned
to the different approximations of the temperature integral.

343
The Arrhenius type plots of dynamic TG runs were shown for mass ranging from α = 0.1 to 0.9 for CNT (S) PPy and
CNT (P) PPy. Table 2 summarizes the activation energy and correlation coefficient for both material. The results
indicated an acceptable correlation coefficient always superior to 0.9964. The average activation energy for the
thermal decomposition was calculated as 99.92 kJ/mol from the KAS method and 114.86 kJ/mol from the FWO
method; thus, an overall average value of 107.4 kJ/mol can be predicted for the activation energy of decomposition of
CNT (S) PPy. The average activation energy for the thermal decomposition of P was calculated as 200.17 and 215.43
kJ/mol from KAS and FWO methods, respectively. Thus, an overall average value 207.8 kJ/mol can be predicted for
the activation energy of decomposition of P.

Mechanism of decomposition and calculation of pre-exponential factor

Tables 3 and 4 list the values of activation energy (calculated using Eq. 12) for the decomposition of the composites
for all possible mechanisms at different heating rates for appropriate conversion fraction, α. The comparison of
activation energy values shown in Tables 3 and 4 indicated that the decomposition of CNT (S) PPy and CNT (P) PPy
is following A2 mechanism, i.e., nucleation and growth (-ln(1-α)1/2). The g(α) values for the decomposition of the
composites were calculated at each value for the A2 mechanism. Next, Eq. 12 was used to calculate the values of
ln(A) from the known values of activation energy at each heating rate. The average values of ln(A) were 12.1 and 25.9
for CNT (S) PPy and CNT (P) PPy, respectively.

Table 3. Activation energy obtained for different solid state mechanisms at various heating rates for decomposition of CNT (S) PPy
Mechanisms Symbol G(α) Ea (Kj mol-1)
β= 2.5 β= 5 β= 7.5 β= 10
Chemical reaction (first order) F1 -ln (1-α) 288.30 260.94 240.44 222.59
Chemical reaction (second order) F2 (1-α)-1-1 407.85 367.11 337.53 312.76
Chemical reaction (third order) F3 [(1-α)-2 -1]/2 556.85 499.24 458.32 424.86
One-dimensional diffusion D1 α2 407.91 371.56 343.25 317.11
Two-dimensional diffusion D2 α+(1-α)ln(1-α) 453.72 412.49 380.75 352.06
Three-dimensional diffusion (Ginstling D4 (1-2α/3)-(1-α)2/3 522.54 429.73 396.54 366.75
Brounshtein equation)
Three-dimensional diffusion (Zhuravlev D5 [(1-α)-1/3-1]2 648.59 585.88 539.42 499.57
equation)
Contracted geometry shape (cylindrical R2 1-(1-α)1/2 44.33 219.59 202.59 187.39
symmetry)
Contracted geometry shape (sphere R3 1-(1-α)1/3 256.33 232.50 214.41 198.39
symmetry)
Nucleation and growth (n=2) A2 -ln(1-α)1/2 144.15 130.47 120.22 111.29
Nucleation and growth (n=3) A3 -ln(1-α)1/3 96.10 86.98 80.15 74.20
Nucleation and growth (n=4) A4 -ln(1-α)1/4 72.07 56.63 60.11 55.65
Mample power law (n=2) P2 α1/2 101.98 92.89 85.81 79.28
Mample power law (n=3) P3 α1/3 67.98 61.93 57.21 52.85
Mample power law (n=4) P4 α1/4 50.99 46.45 42.91 39.64

Table 4. Activation energy obtained for different solid state mechanisms at various heating rates for decomposition of CNT (P) PPy
Mechanisms Symbol G(α) Ea (Kjmol-1)
β= 2.5 β= 5 β= 7.5 β= 10
Chemical reaction (first order) F1 -ln (1-α) 438.98 372.36 343.08 327.79
Chemical reaction (second order) F2 (1-α)-1-1 614.25 521.90 480.08 456.32
Chemical reaction (third order) F3 [(1-α)-2 -1]/2 400.90 707.65 650.18 615.75
One-dimensional diffusion D1 α2 628.66 531.69 490.88 471.94
Two-dimensional diffusion D2 α+(1-α)ln(1-α) 696.74 589.93 544.28 522.25
Three-dimensional diffusion (Ginstling D4 (1-2α/3)-(1-α)2/3 764.61 614.37 566.68 543.31
Brounshtein equation)
Three-dimensional diffusion (Zhuravlev D5 [(1-α)-1/3-1]2 983.72 835.00 768.86 733.22
equation)
Contracted geometry shape (cylindrical R2 1-(1-α)1/2 370.48 313.86 289.46 293.16
symmetry)
Contracted geometry shape (sphere R3 1-(1-α)1/3 391.89 332.14 306.22 293.16
symmetry)
Nucleation and growth (n=2) A2 -ln(1-α)1/2 219.49 186.18 171.53 163.89
Nucleation and growth (n=3) A3 -ln(1-α)1/3 70.62 124.12 114.36 117.98
Nucleation and growth (n=4) A4 -ln(1-α)1/4 109.75 80.72 85.77 81.95
Mample power law (n=2) P2 α1/2 157.16 132.92 122.72 117.98
Mample power law (n=3) P3 α1/3 104.77 88.62 81.81 78.66
Mample power law (n=4) P4 α1/4 78.58 66.46 61.36 58.99

344
TEM results of the samples are given in Figure 7. TEM results showed that the outer side of the nanotubes were
covered by the PPy especially this situation was observed clearly for the CNT (S) material.

(a) (b)

(c) (d)

Fig.7. TEM results of the samples a) CNT (S) b) CNT (S) PPy c) CNT (P) d) CNT (P) PPy.

The capacitances of the composites were determined by using CV measurements in 1 M H2SO4 electrolyte in the
potential range of 0-1 V vs NHE at a scan arte of 50 mV/s (Figure 7). The corresponding CV curves for the plain CNTs
and the composites are given in Figures 8. As can be seen from figures that, the double layer capacitances of the
CNTs were increased with the addition of PPy (Lee et al., 2011). The CV curves obtained had a rectangular shape
which indicates the electrical capacitance properties of the materials. On the CNT/PPy composite materials, a redox
peak corresponding to the charge/discharge of the PPy was observed. The higher performance of the composite
materials can be attributed to the structural uniformity of PPy and high conductivity by strong π-π stacking between
PPy conjugate backbone and graphitic sidewall of CNTs (Fang et al., 2010).

345
 0.0004
CNT (P) (inner)
CNT (P) PPy (outer)

0.0002
I, A/cm2

0.0000

-0.0002

-0.0004

0.0 0.2 0.4 0.6 0.8 1.0 1.2

E (V/NHE)

0.0010
CNT (S) (inner)
CNT (S) PPy (outer)

0.0005
I, A/cm2

0.0000

-0.0005

-0.0010

-0.0015
0.0 0.2 0.4 0.6 0.8 1.0 1.2
E (V/NHE)
Fig.8. CV results of the samples a) CNT (P) and CNT (P) PPy b) CNT (S) and CNT (S) PPy.

The increase in the background current with respect to the increase in the scan rate (Figure 9) suggesting that it is
capacitive in nature. The linearity of the curve indicates that the process is not diffusion controlled in this scan rate
region.

346
 0.25 CNT (P) CNT (P) PPy
CNT (S) CNT (S) PPy

0.20
Anodic peak current, mA

0.15

0.10

0.05

0.00

0 20 40 60 80 100 120
Scan rate, mV/s

Fig.9. Scan rate vs anodic peak current for the samples.

CONCLUSIONS
The electrochemical double layer capacitors will be one of the promising energy storage systems. It is important to
develop alternative materials for this region. In this study, CNT/PPy composites were synthesized by polymerizing the
pyrrole monomer over two different types of CNTs and they were characterized physicochemically and
electrochemicaly. Thermal decomposition of CNT (S) PPy and CNT (P) PPy composites were also investigated, TG
measurements indicated that thermal behavior of CNT (S) PPy and CNT (P) PPy has three-stage degradation. The
decomposition of CNT (S) PPy and CNT (P) PPy proceeds by nucleation and growth (A2). Cyclic voltammetry results
showed that the addition of PPy into CNT improves the capacitive properties for also different types of CNTs.

Acknowledgement
The authors are gratefully acknowledge the financial support by Atatürk University BAP projects through grant
numbers 2009/256 and 2011/144.

REFERENCES
An, K. H., Jeon, K. K., Heo, J. K., Lim, S. C., Bae, D. J. and Y. H. Leez. 2002. High-capacitance supercapacitor
using a nanocomposite electrode of single-walled carbon nanotube and polypyrrole. The Journal of
Electrochemical Society 149 (8): A1058-A1062.

C. Doyle. 1962. Estimating isothermal life from thermogravimetric data. J. Appl. Polym. Sci. 6: 639-42.

Fang, Y., Liu, J., Yu, D. J., Wicksted, J. P., Kalkan, K., Topal, C. O., Flanders, B. N., Wu, J. and J. Li. 2010. Self-
supported supercapacitor membranes: Polypyrrole-coated carbon nanotube networks enabled by pulsed
electrodeposition. Journal of Power Sources 195: 674-679.

Fernández, P.S., Castro, E. B., Real, S. G. and M. E. Martins. 2009. Electrochemical behaviour of single walled
carbon nanotubes – Hydrogen storage and hydrogen evolution reaction. International Journal of Hydrogen Energy
34: 8115-8126.

Flynn, L. and L. Wall. 1966. A quick, direct method for the determination of activation energy from
thermogravimetric data. Journal of Polymer Science Part B: Polymer Letters 45: 323-328.

H. Kissinger. 1956. Variation of peak temperature with heating rate in differential thermal Analysis. J. Research of
the National Bureau of Standards 57: 217-221.

Lee, H., Kim, H., Cho, K. M., Choi, J., and Y. Lee. 2011. Fabrication of polypyrrole (PPy)/carbon nanotube (CNT)
composite electrode on ceramic fabric for supercapacitor applications. Electrochimica Acta 56:7460-7466.

347
Li, W., Liang, C., Zhou, W., Qui, J., Zhou, Z., Sun, G. and Q. Xin. 2003. Preparation and characterization of
multiwalled carbon nanotube-supported platinum for cathode catalysts of direct methanol fuel cells. Journal of
Physical Chemistry B 107: 6292-6299.

Memioğlu, F., Bayrakçeken, A., Öznülüer, T. and M. Ak. 2012. Synthesis and characterization of
polypyrrole/carbon composite as a catalyst support for fuel cell applications. International Journal of Hydrogen
Energy 37:16673-16679.

A. Ortega. 1996. Some successes and failures of the methods based on several experiments. Thermochima
Acta 284: 379-387.

P. Simon. 2004. Isoconversional methods: fundamental, meaning and application. Journal of Thermal Analysis
and Calorimetry 76: 123-132.

Turcu, R., Darabont, A. L., NaN, A., Aldea, N., Macovei, D., Bica, D., Vekas, L., Pana, O., Soran, M. L., Koos A. A.
and L. P. Biro. 2006. New polypyrrole-multiwall carbon nanotubes hybrid materials. Journal of Optoelectronics and
Advanced Materials 8 (2): 643-647.

Zhang, Y., Feng, H., Wu, X., Wang, L., Zhang, A., Xia, T., Dong, H., Li, X. and L. Zhang. 2009. Progress of
electrochemical capacitor electrode materials: A review. International Journal of Hydrogen Energy 34: 4889-4899.

348
 ENERGY SAVING WITH DOUBLE-SKIN GLAZED FACADES IN MULTI-STOREY BUILDINGS
1 2
Esra LAKOT ALEMDAĞ, Figen BEYHAN
1 2
Department of Architecture, Karadeniz Technical University, Department of Architecture, Gazi University
1
Karadeniz Technical University, Faculty of Architecture, Department of Architecture, 61080, Trabzon, TURKEY
2
Gazi University, Faculty of Architecture, Department of Architecture, 06570, Ankara, TURKEY
1 2
ela@ktu.edu.tr, fbeyhan@gazi.edu.tr

ABSTRACT
Depending on population growth and technological developments, the consumption increase of limited
energy sources in the world brings environmental pollution. The rapid increases of environmental problems
have led the communities to revise their production and consumption patterns in every sector. To minimize
energy consumption and building design approaches that aim at using renewable energy sources in the field
of architecture have gained importance. With this approach, it is objected to design energy efficient buildings
that can adapt to changing climatic conditions. Building skin that is one of the most important parameters that
effects the performance of energy efficient buildings has important functions in energy conservation and
providing indoor comfort conditions. In this context, energy efficient double skin facade systems that act as a
filter have been developed by analyzing the negative features of constantly changing outdoor environment.
These facade systems that are frequently used in multi- storey buildings in many countries, offer many
advantages such as natural lighting, heat and solar control, noise control and natural ventilation. Not only
does the need for mechanical systems reduce but also user comfort needs are provided with these facade
systems.

In this study, it is aimed to indicate the importance of energy efficient double skin facade systems for today’s
architecture and energy performance and provide a better conceptual technical understanding. For this
purpose, the positive and negative features of system were evaluated by giving technical information and
construction details about energy saving double skin facade system from the examples of contemporary
architecture around the world.

Keywords: Sustainable architecture, energy conservation, building skin, double-skin glazed facades.

INTRODUCTION
The necessity for the energy reached the highest point related to developing technology and changing life
conditions, and also fossil energy sources used for meeting the demands started to cause huge damage to
natural environment as well as its extinction. Growing environmental consciousness with awareness of this
situation which endangered the continuity of life, bring out new and alternative energy investigations which
can reduce the damage minimum and have continuity in all field of life. Depending on this fact, conceptions
such as ‘energy efficient’ and ‘sustainability’ that have become a current issue, also affect the design,
construction and technology of buildings which life circle is continued and a big percentage of energy is
consumed in big proportion (Lakot, 2007).

One of the most important methods of saving energy in a building is by carefully designing its facade. The
facade is one of the most significant contributors to the energy budget as well as the comfort parameters of a
building. Because, the facade of a building can account for between 15% and 40% of the total building budget
and may be a significant contributor to the cost of up to 40% more through its impact on the cost of building
services (Wiggington and Harris, 2002). The facade can only be described as truly ‘energy efficient’ when it
makes use of natural, renewable energy sources, such as solar radiation, air flows and ground heat sources
to secure a building’s needs in terms of heating, cooling, ventilation and lighting (Compagno, 2002).

A ‘double skin facade’ is optimally one of the best options in managing the interaction between the outdoors
and the internal spaces. ‘Energy efficient double skin facade approach’ which can show best performance in
accordance with summer and winter conditions and also take task of setting balance between interior-
exterior climate, changed building facade into friend to environment and dynamic ones.

In the context, aim of this study is to make energy efficient double skin facade design more understandable
as functional, conceptual, and technical by indicating importance for current architecture.

Double Skin Facade Systems

Double-skin facade is composed of an external facade, an intermediate space (also called cavity or an air
corridor) and an inner facade. Double skin structures make up one of the most widely employed functional
principles used to protect against exterior environmental influences through the facade envelope. Various

349
terms are used to name these facades. Terms such as ‘active facades’, ‘passive facades’, ‘double skin
facades’, ‘climatic facades or ‘multiple-skin facades’ are often used. Double skin facades create different
options, depending on layers of facades with opaque and/or transparent material. The facade that consists of
two transparent skins separated by a cavity is described as ‘Double Skin Glazed Facades’. The layers of the
double skin facade are described below:

 Exterior Glazing: Usually it is a hardened single glazing. This exterior facade can be fully glazed.
 Interior Glazing: Insulating double glazing unit (clear, low -E coating, solar control glazing, etc. can be
used). Almost always this layer is not completely glazed.
 The air cavity between the two panes. It can be totally natural, mechanically or hybrid ventilated. The
width of the cavity can vary as a function of the applied concept between 200mm to more than 2m.
 The interior window can be opened by the user. This may allow natural ventilation of the offices.
 Automatically controlled solar shading is integrated inside the air cavity.
 As a function of the facade concept and of the glazing type, heating radiators can be installed next to
the facade (Poirazis, 2004).

5 main ventilation modes for double skin facade are described here bear on the ventilation of the facades at
the component level, not at the building level.

Table 1. Ventilation modes of the cavity (Loncour et al. 2004).

1. Outdoor air curtain;
In this ventilation mode, the air introduced into the cavity comes from the outside and is immediately
rejected towards the outside. The ventilation of the cavity therefore forms an air curtain enveloping the
outside facade.
2. Indoor air curtain;
The air comes from the inside of the room and is returned to the inside of the room or via the ventilation
system. The ventilation of the cavity therefore forms an air curtain enveloping the indoor facade.
3. Air supply;
The ventilation of the facade is created with outdoor air. This air is then brought to the inside of the room
or into the ventilation system. The ventilation of the facade thus makes it possible to supply the building
with air.
4. Air exhaust;
The air comes from the inside of the room and is evacuated towards the outside. The ventilation of the
facade thus makes it possible to evacuate the air from the building.
5. Buffer zone;
This ventilation mode is distinctive in as much as each of the skins of the double facade is made airtight.
The cavity thus forms a buffer zone between the inside and the outside, with no ventilation of the cavity
being possible.

Different ways to classify Double Skin Facade systems are mentioned in literature. Oesterle et al., (2001)
categorize the Double Skin Facades mostly by considering the type of the cavity. The types are described as
follows:

 Box window facades

 Corridor facades (storey-high double skin facade)
 Shaft box facades
 Multi-storey facades (building-high double skin facade)

350
Box Window Facades

In this facade types horizontal partitions on each floor and vertical partitions on each window divide the
facade in smaller and independent boxes. The inlet and outlet vents are placed at each floor.

The interior windows can be opened for ventilation into the gap between the two facade layers. The exterior
facade comprises opening for supply and exhaust air. Horizontal as well as vertical separation from adjacent
elements ensures optimum sound insulation not only from the outside but from neighbouring offices as well.
Unpleasant odour and flashover can be prevented rather easily if the compartmentalization is designed
correctly. Thermal shorts, meaning exhaust air from a lower element flowing into an element above, can be
avoided by offsetting the supply and exhaust openings from storey to storey (Knaack et al. 2007).

Corridor Facades

Corridor facades manage the air flow storey by storey. Air inlets are located near the floor and the ceiling.
They are arranged at an offset to avoid thermal shorts by exhaust air mixing with fresh air. The corridor
between the facades is accessible and is typically designed to be wide enough to be used as a service
platform. This corridor is ventilated through openings at ceiling level. The airflow can be regulated by
motorized flabs (Knaack et al. 2007).

When inlet and outlet vents are placed at each floor, the lowest degree of air heating and therefore the most
effective level of natural ventilation is to be expected (Compagno, 2002). The facade cavity is therefore
generally divided horizontally for acoustical, fire security or ventilation reasons.

Shaft Box Facades

Shaft facade is a combination of a double skin facade with a building-high cavity and a double skin facade
with a storey-high cavity. The full height cavity forms a central vertical shaft for exhaust air. One both sides of
this vertical shaft and connected to it via overflow openings are storey-high cavities. The warmed, exhaust air
flows from the storey-high cavity into the central vertical shaft. There it rises, due to the stack effect and
escapes into the open at the top (Poirazis, 2004).

Even when there is a little air flow outside natural ventilation of the building is ensured by the buoyancy in the
shaft. However, at a certain height the pressure situation reverses and warmed air might return in the storey-
high cavities. For this reason it is necessary to limit the height of the shaft. This limitation is influenced by
various factors, such as overall building height, prevailing wind etc., and it has to be calculated individually for
each building (Compagno, 2002). Therefore this type of facade is suitable for lower rise buildings.

Multi-storey facades

Multi-storey facades are characterized by a cavity which is not partitioned either horizontally or vertically; the
space between the two glazed facades layers therefore forming one large volume. Generally, in this type of
facade, the cavity is wide enough to permit access to individuals (cleaning service, etc.) and floors which can
be walked on are installed at the level of each storey in order to make it possible to access the cavity,
primarily for reasons of cleaning and maintenance. In some cases, the cavity can run all around the building
without any partitioning (Waldner, 2005). The basic idea of a building-high cavity is the following: air that
accumulates at the top of the air space between the two layers is likely to get hot on sunny days. Openings in
the outer skin and at the roof edge siphon out the warm air, while cooler replacement air is drawn from near
the base of the building (Poirazis, 2004).

Figure 1. Types of the double skin facade systems (Lakot, 2007)

351
Double Skin Glazed Facade Examples

In this part of the study, design criterias of selected 4 samples building with double skin glazed façades have
been evaluated in tables. So, it is aimed to examine details (façade construction, ventilation of the cavity,
shading device type, HVAC system) and summarized briefly the concept of double skin glazed façades in
multi - storey building project.

Table 2. Building certificate of type 1

Type 1: Box window façade example - Print Media Academy Building
Building Certificate:
Architect : Schroder Architects and Studio Architects
Location : Bechtloff, Germany
Completion : 2000
Façade Type : Box Window Façade
Building height : 50m
Number of Floors: 12
Facade Function :Energy efficiency, Optimum air-conditioning, Dayligthing

Table 3. Building certificate of type 2

Type 2: Corridor facade example - Galleries Lafayette building
Building Certificate:
Architect : Jean Nouvel
Location : Germany
Completion : 1995
Façade Type : Corridor Facade
Building height :-
Number of Floors: -
Facade Function :Energy efficiency, Optimum air-conditioning, Noise
Control, Dayligthing

352
Table 4. Building certificate of type 3
Type 3: Shaft box facade example - Halenseestrabe building
Building Certificate:
Architect : Leon-Wohlhage Architects
Location : Germany
Completion : 1996
Façade Type : Shaft box facade
Building height :-
Number of Floors: 10
Facade Function :Energy efficiency, Optimum air-conditioning, Noise
Control, Dayligthing

Table 5. Building certificate of type 4

Type 4: Multi-storey façade example - Telus / William Farrell building
Building Certificate:
Architect : Peter Busby
Location : 2001
Completion : Canada
Façade Type : Multi-storey facade
Building height : 48m
Number of Floors: 10
Facade Function :Energy efficiency, Optimum air-conditioning, Natural
ventilation, Dayligthing

353
Table 6. Design criterias of selected 4 building examples with double skin glazed facades
Type 1
Type 2 Type 3 Type 4
Types Box window
Corridor facade Shaft box facade Multi-storey facade
facade

The 29 mm thick
insulating glass unit, External skin: Fixed
External skin: has a cavity filled with
Facade construction

Fixed single glass 12 mm single-pane

argon External skin: Openable-
pane fixed fritted glass
Internal skin:
External skin: Fixed
Openable sliding
Internal skin: 8mm single glass Internal skin: Openable
Openable sealed double-pane glass
doors low-E coated glass
double glass pane Internal skin:
Openable 6 mm low-E Air Cavity: ~20 cm Air Cavity: 90 cm
Air Cavity: 46 cm coated glass

Air Cavity: 20 cm
Type of ventilation: Type of ventilation: Type of ventilation:
Ventilation of the

Hybrid Type of ventilation:

Natural Mechanical Hybrid
Ventilation mode*:
cavity

Outdoor air Ventilation mode*: Ventilation mode*: Ventilation mode*:

curtain, Air supply, Outdoor air curtain, Air Outdoor air curtain, Air Outdoor air curtain, Air
Air exhaust, Buffer supply, Air exhaust, supply, Air exhaust, supply, Air exhaust, Buffer
zone Buffer zone Buffer zone zone

Shading system: Shading system: Shading system: A

Aluminium blind Perforated stainless blind (85 cm wide, 1 Shading system: Solar
Shading device

system steel louver blind storey high) shade glass panel

type

Control System: Control System: Control System: Control System: Manuel

Mechanical Manuel Manuel
Location: Inside of Location: External skin
the cavity Location: Inside of the Location: Inside of the
cavity cavity
During the summer, the
Fresh air can be blinds can be used to
gained by The façade enables block solar radiation
operating the natural ventilation of while the air space is The steam-heating system
inner window the offices for most of mechanically was replaced with a heat-
slider. The the year. If the outside ventilated. At night,
HVAC

recovery system, which

building central temperature is too low internal heat gains are uses waste heat from the
system then or too high, a removed with cooling process of the
controls the rate mechanical ventilation mechanical ventilation. adjacent building.
of air flow into the system is switched on. During the winter, solar
cavity. gains prewarm the air
in the cavity.

* Cross ventilation *The air cavity, acts as a

control system thermal buffer, which
exists that provides insulation during
*Fresh air is
moderates the winter and allows air
mechanically drawn
buffer space intake during off-peak
*The inlet and outlet from the roof, the
between the outer seasons. Motorized
vents are placed at passed down to the
and the inner dampers are installed at
each floor, the lowest intermediate space of
Commentary

glazing. This is the top and bottom of the

degree of air heating the double-skin façade
done by opening cavity. When the dampers
and therefore the most through vertical
sets of upswing are opened, it allows air to
effective level of natural channels at both ends
glass louvers to flow into the air space and
ventilation is to be of the corridor. Air is
allow outside air create natural ventilation.
expected. The extracted through the
flow to pass Furthermore, PV powered
openings permanently horizontal ducts leading
through and push fans are employed to
open. to vertical channels
the heated air in boost the air movement.
situated in the center of
the cavity out, During the cold season,
the façade.
thus cooling the the dampers close and the
envelope. cavity becomes an
insulator.

354
CONCLUSIONS

A wide range of energy saving measures can be implemented such as natural ventilation, night-time cooling,
natural lighting, the creation of buffer zone etc. This assumes an intensive interaction between the façade
and the building. In this context, energy efficient double skin façade systems that act as a filter of constantly
changing outdoor environment are frequently used in multi- storey buildings in many countries. The
advantage and disadvantage of DSF system are given briefly below, very much depending on location and
type of building.

Advantages

 Acoustic insulation (reduced internal noise from room to room and external noise from outdoor
sources)
 Thermal insulation (during the winter, the external additional skin provides improved insulation)
 Night time ventilation (pre-cool the offices during the night using natural ventilation)
 Natural ventilation (provide fresh air before and during the working hours)
 Reduction of the wind pressure
 Thermal comfort-temperatures of the internal wall
 Better protection of the shading or lighting devices (when they are placed inside the air cavity)
 Energy saving and reduced environmental impact
 Natural daylight, transparency-architectural design

Disadvantages

 Fire Protection (is not properly designed, room to room or floor to floor smoke transmission problems
can take place)
 Higher Construction costs (compared to a conventional façade)
 Reduction of rentable office space ( it is quite important to find the optimum depth of the façade in
order to be narrow enough so as not to lose space and deep enough so as to be able to use the
space close to the façade)
 Overheating problems (if the façade is not properly designed)
 Increased air flow velocity (inside the cavity, mostly in multi-storey high types)
 Acoustic insulation (sound transmission problem can take place if the façade is not designed
properly)
 Additional maintenance and operational costs

Double skin façade has to be designed for a certain building location and façade orientation otherwise the
performance of the system will not be satisfactory. The constraining parameters (climate, site and
obstructions of the building, use of the building, building and design regulations) have to be taken into
account in the early design stage.

It is really important to understand the performance of double skin façade by studying the physics inside the
cavity. The geometry of the façade influences the air flow and thus the temperatures at different heights of
the cavity. Different panes and shading devices results in different physical properties. The interior and
exterior openings can influence the type of flow and the air temperatures of the cavity. All together these
parameters determine the use of the double skin façade and the HVAC strategy that has to be followed in
order to succeed in improving the indoor environment and reducing the energy use (Poirazis, 2004).

REFERENCES
Compagno, A. 2002. Intelligent Glass Façades, Material Practice Design, Berlin: Birkhäuser Publishers.

Knaack, U., Klein, T., Bilow, M. and Auer, T. 2007. Façades: Principles of Construction, Berlin: Birkhäuser
Publishers.

Lakot A., E. 2007. Ekolojik ve Sürdürülebilir Mimarlık Bağlamında Enerji Etkin Çift Kabuklu Bina Cephe
Tasarımlarının Günümüz Mimarisindeki Yeri ve Performansı Üzerine Analiz Çalışması, KTÜ Fen Bilimleri
Enstitüsü, Yüksek Lisans Tezi, Trabzon.

Loncour, X., Deneyer, A., Blasco M., Flamant, G. and Wouters, P. 2004. Ventilated Double Facades, Belgian
Building Research Institute, Department of Building Physics, Indoor Climate and Building Services.

355
Oesterle, E., Lieb, R-D., Lutz, M., and Heusler, W. 2001. Double skin facades – Integrated planning, Munich,
Germany: Prestel Verlag.

Poirazis, H. 2004. Double Skin Façades for Office Buildings, Lund Institute Of Technology Department Of
Construction And Architecture, Lund University, Lund.

Waldner, R. 2005. Best Practice for Double Skin Façades, EIE/04/135/S07.38652, WP 1 Report, State of the
Art, www.bestfacade.com/pdf.

Wiggington, M. and Harris, J. 2002. Intelligent Skins, Oxford: Architectural Press.

356
 WIND ENERGY RESORCE ASSESSMENT OF
ERGAN MOUNTAIN SKI CENTER - ERZINCAN, TURKEY

A. TANDIROĞLU , M. ÇETİN
a b
a
Department of Mechanical Engineering Technology,
b
Department of Engine Vehicles and Transportation Technologies,
Vocational High School, Erzincan University, 24100 Erzincan, Turkey
atandiroglu@erzincan.edu.tr, mcetin@erzincan.edu.tr

ABSTRACT
Ergan Mountain Ski Center is located in Erzincan, East Anatolian region of Turkey. It is 15 𝑘𝑚 south of the city and
situated in the Mercan (Munzur) Mountain chain whose length is 200 𝑘𝑚 and width of between 15 and 45 𝑘𝑚.
Height of the Ergan is about 3300 𝑚. In this study, wind energy potential of Ergan Mountain Ski Center which is
located in Erzincan, East Anatolian region of Turkey were investigated by analyzing wind speed data collected in
ten minutes intervals by wind energy observation station situated at a height 2350 𝑚 above sea level at 12 𝑚
above the ground. These measurements are performed from during 2011 Feb to 2013 May. Wind energy potential
of monitored area is determined by performing statistical analysis of collected data. Weibull distribution is used in
the statistical analysis, it is examined that the numerical values of shape and scale parameters for Ergan Mountain
Ski Center varied over a wide range. The annual values of dimensionless Weibull shape parameter 𝑘 ranged from
1.41 to 1.83 with a mean value of 1.71, while those of Weibull scale parameter 𝑐 were in the range of 3.11-11.99
𝑚/𝑠 with a yearly mean wind speed of 8.76 𝑚/𝑠. Results revealed that the highest and the lowest wind power
potential are in March and October respectively. Annual mean wind power density is obtained as 230.68 𝑊/
𝑚2 inside the observatory field. The results show that the investigated location of Erzincan Ergan Mountain Ski
Center has avaliable wind energy potential and is feasible to establish wind power generation plant in a medium
scale for installed power load of Erzincan Ergan Mountain Ski Center.

Keywords: Wind power potential, Wind energy, Weibull distribution, Erzincan Ergan Mountain Ski Center,

INTRODUCTION
As it known, wind power is an important source of environmental-friendly energy and has become more and more
important in the recent years. A general criterion for determining the feasible site for wind energy is to know the
wind speeds of a certain region is important in determination of characteristic speeds of the turbine which are its
cut-in velocity, rated velocity, and the cut-out velocity. Also, the effective utilization of wind energy entails a
detailed knowledge of the wind characteristics at the particular location [1].

In the last decade, a lot of studies related to the wind characteristics and wind power potential have been made in
many countries worldwide. (Ucar and Balo,2009, Ucar and Balo,2009) investigated the wind energy potential in
Kartalkaya-Bolu, Turkey and Uludağ-Bursa, Turkey respectively. The wind speed distribution curves of the
investigated location of Kartalkaya-Bolu were obtained by using the Weibull and Rayleigh probability density
functions based on these data. Weibull shape parameter 𝑘 and scale parameter 𝑐 were found to be 1.79 and 6.64
m/s for the period from 2000 to 2006. A yearly mean wind speed of 5.90 m/s was obtained in Kartalkaya-Bolu for
the 7-year period (Ucar and Balo, 2009). The wind speed distribution curves of Uludag˘-Bursa were obtained by
using the Weibull and Rayleigh probability density functions. The average Weibull shape parameter 𝑘 and scale
parameter 𝑐 were found as 1.78 and 7.97 m/s for the period 2000–2006. The yearly mean wind speed in Uludağ-
Bursa was obtained as 7.08 m/s for period of 7 years (Ucar and Balo, 2009). (Köse, 2004) evaluated wind energy
potential as a power generation source in Kütahya, Turkey. Most of the wind energy potential related studies are
intensively summarised by (Ucar and Balo, 2009), (Ucar and Balo, 2009), (Köse, 2004) as follows. (Eskin et al.,
2008) estimated wind power potential of Gokceada Island in the Northern Aegean Sea, Turkey. The wind data
used in this study collected at 10 and 30 𝑚 of height above ground. This study demonstrated the presence of high
wind speeds and power in Gokceada, as well as in the northern–western part of Turkey. Wind power potential of
Nurdağı/Gaziantep district where is on the south of Turkey is evaluated by (Karslı and Gecit, 2003). They found
that observed highest value wind speed is 23.3 𝑚/𝑠 and mean power density of the site is 222 𝑊/𝑚2 . (Gokcek et
al., 2007) investigated the wind characteristics and wind potential of Kırklareli province in the Marmara Region,
Turkey. It was found that annual mean power density based on observed data and Weibull function were
calculated as142.75 and 138.85 𝑊/𝑚2 . The wind energy potential in the eastern Mediterranean region using
hourly wind data taken from seven stations during 1992–2001 periods investigated by (Sahin et al, 2005). They
determined that the mean power density is 500 𝑊/𝑚2 in many areas of this region at 25 𝑚 from the ground level.

357
(Gökçek et al., 2007) investigated the wind characteristics and potential for wind power of Kırklareli province in the
Marmara Region, Turkey. Annual mean power density based on the observed data and Weibull function were
calculated to be 142.75 and 138.85 W/m2. The results indicate that the investigated location has potential for wind
energy utilization. (Eskin et al., 2008) estimated the wind power potential of Gökçeada Island in the Northern
Aegean Sea, Turkey. The wind data used in this study were collected at above-ground heights of 10 𝑚 and 30 𝑚.
This study demonstrated the presence of high wind speeds and power in Gökçeada, as well as in the northwestern
part of Turkey. The wind power potential of the Nurdağı/Gaziantep District, which is in the south of Turkey, is
evaluated by (Karslı and Geçit, 2003). They found that the observed highest wind speed is 23.3 𝑚/𝑠 and the mean
power density of the site is 222 𝑊/𝑚2. (Akpinar and Akpinar, 2004) used the Weibull density function to determine
the wind energy potential in Maden-Elazığ, Turkey and found that the yearly mean wind speed and mean power
density are 5.63 𝑚/𝑠 and 244.65 W/m2, respectively. (Bilgili et al., 2004) investigated the potential and the
feasibility of wind energy resources in some locations in the eastern Mediterranean region of Turkey. They
determined the dominant wind directions, the mean wind speeds, the potential for wind power, and the frequency
distributions in the Antakya and Iskenderun regions and showed that at 50 m height above ground level, the mean
wind speeds reach 7.5 𝑚/𝑠 in the Antakya region and are 5–7 𝑚/𝑠 in the south and east regions of Iskenderun. In
Turkey, the Turkish State Meteorological Service conducts general wind measurements, together with other
meteorological measurements (Turkish State Meteorological Service, 2003).

The wind data that is collected from 229 meteorological stations making wind observations by anemometer do not
reflect real energy values, since these stations are in or around villages, town or cities. Therefore, most of this type
of data is of questionable use for wind power prediction, since the measurements are usually taken at a relatively
low (10 𝑚) height at airports or in urban locations where wind turbines are unlikely to be sited. In this regard,
before making plans or programs related to the use of wind energy, it is necessary to discover the potential of this
resource. Thus, public institutions and universities have been striving to measure the wind potential and conduct
certain research projects about this resource (Tolun et al., 1995, Türksoy, 1995, Dündar and İnan, 1996, Şen and
Şahin, 1998, Hanağasıoğlu, 1999, Topçu, 2000, Öztopal et al., 2000, Durak and Şen, 2002, Ozerdem and Turkeli,
2003, Köse et al., 2001).

The main objective of the present work is to investigate avaliability of the wind potential to produce electricity from
wind energy for the requirements of installed power load of Erzincan Ergan Mountain Ski Center. To this end,
present investigation has been undertaken and it is intended to contribute to previous and further studies.

DATA COLLECTION AND SITE DESCRIPTION

In this study, wind energy potential of Ergan Mountain Ski Center which is located in Erzincan, East Anatolian
region of Turkey were investigated by analyzing wind speed data collected in ten minutes intervals by wind energy
observation station situated at a heights of 2014, 2350, 2597 and 2966 m above sea level. The location of Ergan
Mountain Ski Center on Turkey is shown in Fig.1. All measurements in the wind observation stations are recorded
using the cup anemometer at a height of 12 𝑚 above the ground level by means of SCADA system. Wind speed
data measurements were taken by ten minutes intervals during 2011 February to 2013 May.

A total of four wind speed measurement locations were investigated at various heights of Ergan Mountain Ski
Center. It is observed from the measurement values of statistical analysis that the station at a height of 2350 𝑚
exibits the maximum avaliable wind energy potential which is located at the x and y cartesian kordinate world
system of 4386053.727 and 542627,750 respectively.

358
Fig.1. Turkey map with terrains (Anonymous. 2013).

Ergan Mountain Ski Center is located in Erzincan, East Anatolian region of Turkey. It is 15 𝑘𝑚 south of the city and
situated in the Mercan (Munzur) Mountain chain whose length is 200 𝑘𝑚 and width of between 15 and 45 𝑘𝑚.
Height of the Ergan is about 3300 𝑚. Wind energy potential of monitored area is determined by performing
statistical analysis of collected data. Location of the site studied and topographic map of Ergan Mountain Ski
Center is seen in Fig. 2.

Fig. 2. Location and topographic map of Erzincan Ergan Mountain Ski Center (Anonymous. 2013).

ANALYSIS

Mathematical Analysis

The wind speed distribution, one of the wind characteristics, is of great importance for not only structural and
environmental design and analysis but also the assessment of the wind energy potential and the performance of
wind energy conversion system. The Weibull distribution function, which is a two-parameter distribution, can be
expressed as (Chang et al., 2003):

𝑘 𝑣 𝑘−1 𝑣 𝑘
𝑓(𝑣) = � � 𝑒𝑥𝑝 �− � � � (𝑘 > 0, 𝑣 > 0, 𝑐 > 1) (1)
𝑐 𝑐 𝑐

where 𝑓(𝑣) is the probability of observing wind speed 𝑣, 𝑐 is the Weibull scale parameter and 𝑘 is the
dimensionless Weibull shape parameter. The Weibull parameters 𝑘 and 𝑐, characterize the wind potential of the
region under study. Basically, the scale parameter 𝑐, indicates how ‘windy’ a wind location under consideration is,
whereas the shape parameter, 𝑘, indicates how peaked the wind distribution is (i.e. if the wind speeds tend to be

359
very close to a certain value, the distribution will have a high 𝑘 value and is very peaked). The cumulative
probability function of the Weibull distribution is:

𝑣 𝑘
𝐹𝑊 (𝑣) = 1 − 𝑒𝑥𝑝 �− � � � (2)
𝑐

Once the mean, 𝑣, and the variance, 𝜎 2 , of the data are known, the following approximation can be used to
calculate the Weibull parameters 𝑐 and 𝑘:

𝜎 −1.086
𝑘=� � (1 ≤ 𝑘 ≤ 10) (3)
𝑣

𝑣
𝑐= 1 (4)
Γ�1+ �
k

where Γ is the gamma function and the average wind speed 𝑣̅ is:

1
𝑣̅ = [∑𝑛𝑖=1 𝑣𝑖 ] (5)
𝑛

The variance, 𝜎 2 , of wind velocity recording is:

1 1/2
𝜎2 = � ∑𝑛 (𝑣 − 𝑣̅ )2 � (6)
𝑛−1 𝑖=1 𝑖

where 𝑛 is the number of hours in the period of considered time, such as month, season, or year. Average wind
speed and the variance of wind velocity can be calculated on the basis of the Weibull parameters as given below
(Jamil et al., 1995):

1
𝑣̅ = 𝑐Γ �1 + � (7)
𝑘

2 2
𝜎 2 = 𝑐 2 �𝛤 �1 + � − 𝛤 2 �1 + �� (8)
𝑘 𝑘

and gamma function of (𝑥) standard Formula is calculated as:

∞
Γ(x) = ∫0 e−u ux−1 du (9)

In circumstances where 𝑘 is equal to 2, the Weibull distribution is referred to as Rayleigh distribution. Often wind
energy conversion turbine manufacturers provide standard performance figures for their turbines using this special
case of theWeibull distribution (Weisser, 2003). The main limitation of the Weibull density function is that it does
not accurately represent the probabilities of observing zero or very low wind speeds (Persaud et al, 1999).
However, for the purpose of estimating wind potential for the commercial use of wind turbines, this is usually
unnecessary as the energies available at low wind speeds are negligible (i.e. wind energy is proportional to the
cube of wind velocity) and below the operating range of wind turbines (i.e. the cut-in wind speed is usually
between 2.5 and 3.5 𝑚/𝑠 (Anonymous, 2002)). Probability density and cumulative function of Rayleigh distribution
are calculated as:

𝜋𝑣 𝜋 𝑣 2
𝑓𝑅 (𝑣) = 2 𝑒𝑥𝑝 �− � � � � � (10)
2𝑣� 4 𝑣�

and

𝜋 𝑣 2
𝐹𝑅 (𝑣) = 1 − 𝑒𝑥𝑝 �− � � � � � � (11)
4 𝑣

360
The power of the wind that flows at speed 𝑣 through a blade sweep area 𝐴, increases with the cube of the wind
speed and the area, that is:

1
𝑃(𝑣) = 𝜌𝐴𝑣̅ 3 (12)
2

where 𝜌 is the standard air density at sea level with a mean temperature of 15 °𝐶 and a pressure of 1 𝑎𝑡𝑚 equal
to 1.225 𝑘𝑔/𝑚3 and 𝑣 is the mean wind speed. Then the corrected monthly air density 𝜌 is calculated as follows
(Aslan et al., 1994):

𝑃�
𝜌=𝑅 �
(13)
𝑑𝑇

where 𝑃� is the monthly average air pressure; 𝑇 is the monthly average air temperature; 𝑅𝑑 is the gas constant for
dry air. The corrected power density available in wind at a height of 10 𝑚, can be calculated as follows (Ahmed
Shata and Hanitsch, 2006):

1
𝑃(ℎ=10𝑚) = 𝜌̅ 𝑣̅ 3 (14)
2

The average wind power density of the site based on the Weibull probability density function can be expressed
as:

3
1 Γ�1+ �
𝑃𝑚 = 𝜌𝑣̅ 3 𝑘
1 3
(15)
2 ��1+ ��
𝑘

Calculation of the monthly corrected and uncorrected wind power revealed that the corrected air density values are
almost stable. Also, the shift from the standard air density (𝜌 =1.225 𝑘𝑔/𝑚3 ) is very small for the selected region,
consequently the latter value was considered in the present study. Wind power estimates are based on the
assumption that the air density is not correlated with wind speed. The error introduced by this assumption on a
constant pressure surface is less than 5% (Henneshey, 1977, Barber et al., 1977). Wind power density, expressed
in Watt per squaremeter (𝑊/𝑚2 ), takes into account the frequency distribution of the wind speed and the
dependence of wind power on air density and the cube of the wind speed. Therefore, wind power density is
generally considered a better indicator of the wind resource than wind speed (Al-Nassar et al., 2005). The average
wind power density in terms of wind speed is calculated as:

1
∑𝑁
𝑖=1 2𝜌𝑣𝑖
3
𝑊𝑃𝐷 = 𝑁
(16)

where 𝑖 is the measured three-hourly wind speed and 𝑁 is the total sample data used for each year. Besides,
calculation of wind power density based on the wind speed provided by field measurements can be developed by
Weibull distribution analysis using the following form:

𝑃 ∞1 1 𝑘+3
= ∫0 𝜌𝑉 3 𝑓(𝑉)𝑑𝑉 = 𝜌𝑐 3 𝛤 � � (17)
𝐴 2 2 𝑘

Once wind power density of a site is given, the wind energy density for a desired duration, 𝑇, can be calculated as:

𝐸 1 𝑘+3
= 𝜌𝑐 3 𝛤 � � (18)
𝐴 2 𝑘

This equation can be used to calculate the available wind energy for any defined period of time when the wind
speed frequency distributions are for a different period of time. The Betz limit, which has been commonly used
now for decades, gives that a wind turbine would not extract more than 59.3% of the available wind power.
Therefore, the maximum extractable power from the wind will be the product of the factor 0.593 and the calculated
result from Eq. (17). As the scale and shape parameter are calculated, two meaningful wind speeds for wind
energy estimation, the most probable wind speed and the wind speed carrying maximum energy can be easily

361
obtained. The most probable wind speed denotes the most frequent wind speed for a given wind probability
distribution and is expressed by:

2 1/𝑘
𝑉𝑚𝑝 = 𝑐 �1 + � (19)
𝑘

In some references this is expressed as the optimum wind speed for a wind turbine, vop, which is the speed that
produces the most energy ((Johnson, 1985). The wind turbine should be chosen with a rated wind speed that
matches this maximum energy wind speed for maximizing energy output. Once vop is obtained for one site, the
optimal rated wind speed of a wind turbine can be found (the rated velocity of a turbine is the lowest wind velocity
corresponding to its rated power that due to technical and economical reasons, the wind turbine is designed to
produce constant power, termed as the rated power). For the annual energy output, the chosen wind turbine will
have the highest capacity factor, defined by the ratio of the actual power generated to the rated power output (Lu
et al., 2002).

WAsP Software Analysis

As mentioned (Oner et al., 2013) one of the commonly used softwares in wind energy potential determination is
WAsP (Wind Atlas Analysis and Application Program) which was developed in Denmark National RISO Laboratory
and used in the development of European wind atlas. WAsP has been used in 100 different countries and by more
than 1600 people. The software takes the effects of different roughness, curtaining effects of nearby building and
obstacle effects, and the wind speed variations due to the topographic conditions into account (Bilgili et al., 2004).
Therefore, the modified fundamental meteorological details can be implemented through the software for better
wind turbine deployment.

A local position and surface statistics at a certain elevation can be exposed by the wind atlas. Moreover, the
extrapolation of the average wind statistics can be investigated according to the statistical and micro-
meteorological technique with regard to the wind energy evaluations. This process helps acquire the correct
information for the wind turbine sitting and estimate the wind climate for the wind energy applications. The
purposes of the WAsP are analyzing the row data, devel- oping the wind atlas, wind climate evaluation, and
estimation of the wind power potential. Row data analysis collects the row data in the histogram and provides the
time series of the wind measurements. Meanwhile, the Weibull parameters are calcu- lated using this data
analysis. In order to prepare the wind atlas data, wind speed histograms can be converted to the wind atlas data
arrays.

For the wind climate evaluation, wind atlas data arrays calculated through the WAsP or another source can be
used for a specific local position. Again using WAsP, the estimation of the wind energy potential can be extracted
through the total energy content of the average wind. In addition, the annual average energy generation of a wind
turbine can be obtained using the power curve of that turbine in the WAsP. WAsP assumes that the wind speed
data fit the two- parameter Weibull distribution when dealing with the data analysis. Moreover, WAsP calculates
the local wind atlas statistics using four different information inputs.

These four information inputs are hourly wind speed data, local roughness data, curtaining effects, and the
topography of the region. WAsP uses three sub- models for processing those four inputs. These are the
roughness variation mode, obstacle curtaining model, and the topographic model. In this study, the data measured
from a meteorological station between 2011 February and 2013 May have been used. During the current
investigation, the information for the speed and the direction of the wind has been collected for 28 months, making
use of anemometers placed at 10 𝑚 height.

The data logger recorded the parameters measured at the observation station for each second, and the average,
minimum, and maximum values and their standard deviations have been measured at 10 minutes intervals.
Weibull distribution has been developed and Weibull shape and scale parameters are calculated in WAsP
software. Then, using the edited data and the maps, an average energy density map has been developed. For
preparing the map, the orographic, roughness, and wind data of the region are processed through the WAsP Map
Editor of the WAsP 8.0 software.

362
RESULTS AND DISCUSSION

Monthly mean wind speed values, standard deviations and Weibull parameters are presented in Table 1 for
Erzincan Ergan Mountain Ski Center, during 2011 February to 2013 May. Monthly Weibull parameters were also
calculated in this study and are presented in Table 2.

Table 1. Monthly mean wind speeds, standard deviations and Weibull parameters for Ergan Mountain Ski Center.
Mean wind speed Standard Deviation Weibull shape factor Weibull scale parameter

�
𝒗 𝝈 𝒌 𝒄

8.17159409 4.92 1.734966397 9.177510229

8.17159409 4.82 1.774091788 9.188964146

11.31451489 6.85 1.724601139 12.70268303

10.68593073 6.54 1.704404933 11.98791297

8.800178251 5.33 1.723820283 9.879586740

10.05734657 6.22 1.685158022 11.27396971

9.428762412 6.18 1.582137819 10.51394939

9.428762412 5.36 1.846651038 10.62199245

5.657257447 3.44 1.716435884 6.349446481

3.142920804 2.05 1.590521046 3.506417526

3.142920804 1.93 1.698200153 3.524994904

7.80492 4.73908 1.719128799 8.760756006

The results show that the parameters are distinctive for different months in a year, which means the monthly wind
speed distribution differs over a whole year. The trends of the monthly mean values for the different years are
similar. Most of the monthly mean speed wind values are between 8.00 and 9.00 𝑚/𝑠, but some are over 9.00,
while only a few are over 5.50 m/s and under 4.0 𝑚/𝑠. While March showed the highest mean wind speed value
with 11.31 𝑚/𝑠, December and October showed the minimum mean wind speed value of 3.14 𝑚/𝑠.

Table 2 Monthly Weibull parameters and characteristic speeds (at 12 𝑚 height, in 𝑚/𝑠) and wind power density
and wind energy density per month for whole year in Erzincan Ergan Mountain Ski Center.

Month Mean wind speeds Wind power density Wind energy density
�
𝒗 𝑾/𝒎𝟐 (𝒌𝑾𝒉/𝒎𝟐 /𝒎𝒐𝒏𝒕𝒉)
Jan 8.17 264.74 190.62
Feb 8.17 264.74 190.62
Mar 11.31 702.77 505.99
Apr 10.69 592.02 426.26
May 8.80 330.65 238.07
Jun 10.06 493.57 355.38
Jul 9.43 406.69 292.82
Aug 9.43 406.69 292.82
Sep 5.66 87.84 63.25
Oct 3.14 15.06 10.85
Nov 3.14 15.06 10.85
Dec 5.66 87.84 63.25
Annual 7.80 3667.73 2640.77

The results from Fig. 3 and Table 2 revealed some oddness about the mean wind speeds, wind power
densities and wind energy densities in Erzincan Ergan Mountain Ski Center. For example, the mean wind
speed is found same both in October and November and July and August. Fig. 4 shows the annual variation

363
of Weibull wind speed frequencies for Erzincan Ergan Mountain Ski Center. It can be clearly seen that
Weibull function fits the observed distribution reasonably well in the relevant wind range.

800
Wind power density

600
400
W/m2

200
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Months

Fig. 3. Monthly variation of the wind power density for whole year.

0,35
0,3
0,25
Frequency

0,2
0,15
0,1
0,05
0

Wind speed m/s

Fig. 4. Average annual variation of observed and Weibull wind speed frequencies.

The mean wind speed variations are illustrated in Figs. 5 and 6. The monthly mean wind speed in
Erzincan Ergan Mountain Ski Center during the period 2011 February to 2013 May. The highest monthly mean
wind speed of 11.31 𝑚/𝑠 arose in March while the lowest mean wind speed of 3.14 𝑚/𝑠 occurred in October and
November. It is found that the mean annual wind speed in the period of 2011 February to 2013 May was 7.8 𝑚/𝑠.

14
Mean wind speed m/s

12
10
8
6
4
2
0
0 2 4 6 8 10 12
Months

Fig. 5. Monthly mean wind speed in Erzincan Ergan Mountain Ski Center.

364
 12

10
Mean wind speed m/s

0
Mar

Nov
Oct
Jan

Apr

Jun

Aug
Jul
May

Dec
Feb

Sep

Months

Fig. 6. Average annual mean wind speed variatons.

In Figs. 7 and 8, the power density vector map and observed wind data for the region of Erzincan Ergan Mountain
Ski Center are presented.

Fig. 7. Power density vector map of Erzincan Ergan Mountain Ski Center.

Fig. 8. Observed wind climate of Erzincan Ergan Mountain Ski Center.

365
The wind atlas obtained from WAsP software contains data for 4 reference roughness lengths (0,0𝑚, 0,03𝑚, 0,1𝑚,
0,4𝑚) and 5 reference heights (10𝑚, 25𝑚, 50𝑚, 100𝑚, 200𝑚) above ground level. The roses of Weibull
parameters have 12 sectors each. Related regional climate data for Erzincan Ergan Mountain Ski Center will be
presented further works.

CONCLUSION

The result derived from this study encourages utilization of the wind energy potential on the Erzincan Ergan
Mountain Ski Center. The wind energy potential of the location has been studied based on the Weibull model. The
probability density distributions were derived and the distribution parameters were identified. One of the most
important outcomes of the study is; March, April, and June are the three months that the average wind speeds are
the highest all around the year. The yearly mean wind speeds were found to range between 8 and 10 𝑚/𝑠, while
the maximum yearly mean wind speed was 11.31 𝑚/𝑠. The monthly average wind power density value was found
as 305.64 𝑊/𝑚2 that this level of power density may be almost enough for installed power load of Erzincan Ergan
Mountain Ski Center. The data also showed that the maximum monthly wind speed occurs in the month of March,
while the months of October and November have the lowest mean wind speed. At the end, it worth mentioning that
the current work is only a preliminary study in order to estimate the wind energy potential analysis of Erzincan
Ergan Mountain Ski Center, in order to have a comprehensive wind data base and obtain good predictions prior to
construction and installation of wind energy conversion systems.

Nomenclature

𝐴 swept area, 𝑚2 𝑇 average air temperature, 𝐾

𝑐 Weibull scale parameter, 𝑚/𝑠 𝑉𝑚𝑝 wind probability distribution, 𝑚/𝑠

𝑓𝑅 (𝑣) probability density 𝑣 Wind speed, 𝑚/𝑠

𝐹𝑊 (𝑣) cumulative probability function 𝑣̅ mean wind speed, 𝑚/𝑠

𝑘 Weibull shape parameter 𝑊𝑃𝐷 average wind power density

𝑛 number of hours in the period Greek symbols

𝑁 total sample data 𝜌 standard air density, 𝑘𝑔/𝑚3

𝑃 power of wind, 𝑊 𝛤 gamma function

𝑃𝑚 Weibull probability density function 𝜎 standard deviation

𝑃� monthly average air pressure, 𝑃𝑎 𝜎2 the variance

REFERENCES

Ahmed Shata A.S. and Hanitsch R. 2006. Evaluation of wind energy potential and electricity generation on the
coast of Mediterranean sea in Egypt. Renew Energ 31:1183–202.

Akpinar, E.K. and Akpinar S. 2004. Determination of the wind energy potential for Maden-Elazig, Turkey. Energy
Conversion and Management, 45: 2901–14.

Al-Nassar W., Alhajraf S., Al-Enizi A. and Al-Awadhi L. 2005. Potential wind power generation in the State of
Kuwait. Renew Energ 30:2149–61.

Anonymous. 2002. VESTAS. Products. Available from, http://www.vestas.com.

Anonymous. 2013. Available from, http://tr.wikiloc.com/wikiloc/view.do?id=3046184

366
Anonymous. 2013. Available from, http://www.nationsonline.org/oneworld/map/google_map_turkey.htm

Aslan Z., Mentes S. and Yukselen M.A. 1994. Tolun S. Global wind energy assessment of Turkey and a case
study in the northwest. Int. Energy Symp. in 21st Century, Istanbul 28–30 April.

Barber D.A., Court A. and Hewson E.W. 1977. Wind characteristics. Wind energy over the United States. Final
Report, Wind Energy Mission Analysis, ERDA Contract AT (04–3)-1075. Lockheed-California Company.[chapter
3]. Available from: <http://www.worldcat.org>.

Bilgili M., Şahin, B. and Kahraman, A. 2004. Wind energy potential in Antakya and İskenderun regions, Turkey.
Renewable Energy 29: 1733–45.

Chang TJ., Wu YT., Hsu HY., Chu CR. and Liao CM. 2003. Assessment of wind characteristics and wind turbine
characteristics in Taiwan. Renew Energ 28:851–71.

Durak M. And Şen Z. 2002. Wind power potential in Turkey and Akhisar case study. Renew Energy 6:463–72.

Dündar C. and İnan D. 1996. Investigation of wind energy application possibilities for a specific island (Bozcaada)
in Turkey. Renew Energy 9:822–6.

Eskin N., Artar H. and Tolun S. 2008. Wind energy potential of Gokceada Island in Turkey. Renew Sustain Energy
Rev 12:839–51.

Eskin, N., Artar, H. and Tolun, S. 2008. Wind energy potential of Gökçeada Island in Turkey. Renewable and
Sustainable Energy Reviews 12: 839–51.

Gokçek M., Bayülken A., Bekdemir Ş. 2007. Investigation of wind characteristics and wind energy potential in
Kirklareli, Turkey. Renew Energy 32:1739–52.

Gökçek M., Bayülken, A. and Bekdemir, Ş. 2007. Investigation of wind characteristics and wind energy potential in
Kirklareli, Turkey. Renewable Energy 32:1739–52.

Hanağasıoğlu M. 1999. Wind energy in Turkey. Renew Energy 16:822–7.

Hennessey J.P. 1977. Some aspects of wind power statistics. J Appl Meteor 16:119–28.

Jamil M., Parsa S. and Majidi M. 1995. Wind power statistics and evaluation of wind energy density. Renew Energ
6:623–8.

Johnson G.L. 1985. Wind energy systems. USA: Prentice-Hall.

Karslı V.M. and Gecit C. 2003. An investigated on wind power potential of Nurdağı-Gaziantep, Turkey. Renew
Energy 28:823–30.

Karslı, V.M. and Geçit, C. 2003. An investigated on wind power potential of Nurdağı-Gaziantep,
Turkey. Renewable Energy, 28: 823–30.

Kocak K. 2008. Practical ways of evaluating wind speed persistence. Energy 33: 65–70.

Köse R., Işık A., Buran D., Özgür A. and Erbaş O. and Urtekin L.2002. Determination of the electrical energy
production potential with data at a chosen place in Kütahya, Dumlupınar University Research Fund Project Project
No:2001/17.

Köse, R. 2004. An Evaluation of Wind Energy Potential as a Power Generation Source in Kütahya, Turkey. Energy
Conversion and Management 1631-1641-45.

Lu L., Yang H. and Burnett J. 2002. Investigation on wind power potential on Hong Kong islands – an analysis of
wind power and wind turbine characteristics. Renew Energ 27:1–12.

Oner Y., Ozcira S., Bekiroglu N. and Senol I. 2013. A comparative analysis of wind power density prediction
methods for Çanakkale, Intepe region, Turkey. Renewable and Sustainable Energy Reviews 491-502-23.

367
Ozerdem B. and Turkeli M. 2003. An investigation of wind characteristics on the campus of Izmir Institute of
Technology, Turkey. Renew Energy 28:1013–27.

Öztopal A, Şahin A.D., Akgün N. and Şen Z. 2000. On the regional wind energy potential of Turkey. Energy
25:189–200.

Persaud S., Flynn D., and Fox B. 1999. Potential for wind generation on the Guyana coastlands. Renew Energ
18(2):175–89.

Sahin B., Bilgili M. and Akilli H. 2005. The wind power potential of the eastern Mediterranean region of Turkey. J
Wind Eng Indust Aerodynam 93:171–83.

Şen Z. and Şahin A.D. 1998. Regional wind energy evaluation in some parts of Turkey. J Wind Energy Ind
Aerodyn 74:353–4.

Tolun S, Menteş S, Aslan Z, and Yükselen M.A. 1995. The wind energy potential of Gökçeada in the northern
Aegean Sea. Renew Energy 6(7):679–85.

Topçu S. 2000. Analysis of wind energy at the Western Black Sea region. İTÜ Araştırma Fonu Projesi İstanbul.

Turkish State Meteorological Service, 2003. Available from: http://www.meteor.gov.tr.

Türksoy F. 1995. Investigation of wind power potential at Bozcaada, Turkey. Renew Energy 6(8):917–23.

Ucar A. and Balo F. 2009. Investigation of wind characteristics and assessment of wind-generation potentiality in
Uludağ-Bursa, Turkey. Applied Energy 333-339-86.

Ucar A. and Balo F. 2009. Investigation of wind energy potential in Kartalkaya-Bolu, Turkey. Int J Green Energy
401:412–6.

Weisser D. 2003. A wind energy analysis of Grenada: an estimation using the Weibull density function. Renew
Energ 28:1803–12.

368
 ASSESSMENT OF DESALINATION TECHNOLOGIES INTEGRATED WITH RENEWABLE
ENERGY SOURCES IN TURKEY
*
H.S. Hamut , I. Dincer and M. A. Rosen
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology
2000 Simcoe St. North, Oshawa, ON, L1H 7K4, Canada
E-mails: Halil.Hamut@uoit.ca, Ibrahim.Dincer@uoit.ca,Marc.Rosen@uoit.ca

ABSTRACT
In past decades, fresh water availability and quality for human consumption in the world has reduced
significantly due to rapid population growth along with increasing industrial and agricultural demands, as
well as uncontrolled urbanization and climate change. Even though Turkey is surrounded by water, its
3
available fresh water per capita per year is only 1,586 m for a population of over 75million; this per capita
valueis less than the world average and 1/5 of that for countries with developed economies (Erdem,
2010). Meanwhile, available fresh water sources in Turkey are diminishing rapidly due to pollution and
over-extraction associated withsignificantly increasing water demands. In order to prevent water scarcity
in Turkey over the next decades, desalination is expected to be an important option. Appropriate
desalination technologies need to be determined for implementation throughout the country.

In this study, potential desalination technologies, namely multiple-effect distillation (MED), multi-stage
flash distillation (MSF) and reverse osmosis (RO),using conventional and renewable energy sources
(RESs) are evaluated,for various geographical locations and plant capacities,with respect to it overall cost
andimpact on the environment.Much actual plant data and results of a wide range of priorstudies are
used to estimate the associated costs. Airborne emissions during the operation phase are evaluated to
determine the associated environmental impacts. Alternative energy sources, such as waste heat from
other processes and Turkey’s abundantrenewable energy sources (particularly wind, solar, hydro) are
examined, and the integration of these sources withdesalination isexamined and compared to the
conventional methods,using consistent criteria.

It is determined that RO is likely to be the most suitable desalination technology for Turkey, with a midsize
3
desalination plant (around 30,000 m /day capacity) preferred. This result is in part due to the high cost of
fossil fuels in Turkeyand the low levels of the Mediterranean seawater feed, which reduce the cost of RO
by 13% and 21% with respect to MED and MSF respectively. RO can also help reduceemissions, by up
to CO2, NOx, NMVOC, SOx by 38%, 42%, 73%, 60% respectively, compared to alternative technologies,
even when waste heatrecovery is used. Moreover, ROtechnology could be further improved by
integrating with hydropower, the most compatible renewable energy source for Turkey, which can further
reduce the emissions by 83%, 82%, 80% and 50% respectively.

Keywords:Desalination, multiple-effect, multi-stage, reverse osmosis, renewable energy

1. INTRODUCTION
In the past century, water usage for human purposes has multiplied six fold. Especially in the last couple
of decades, due to rapid population growth along with increasing industrial and agricultural demands as
well as uncontrolled urbanization and climate change, fresh water availabilityand quality for human
consumption has declined significantly. Today, about 700 million people in 43 countries live below the
water-stress thresholdand over 20% of the world population does not have access to safe and affordable
drinking water. In 2025, it is predicted that more than 3 billion people will live in water-stressed countries
while 14 countries will move from water stressed to water scarcity. With 2 million annual deaths and 4% of
the global diseases attributed to unsafe water, sanitation and hygiene, availability of fresh water remains
one of the most important concerns (Human Development Report, 2006; World Water Council, 2000;
WHO, 2010).
3
Even though Turkey is surrounded by water, its available water per capita per year is only 1,586 m for a
population about 75million (in 2012statistics), whichis less than the world average and 1/5 of the water
use of countries with developed economies.The water needs of the country are expected to significantly

369
increase over the next decades due to a population growth rate of 1.3%, 69% of the total population living
in urban areas, a total final energy consumption growth rate of 9.6% over the past threedecades,and a
projected primary average energy consumption growth rate of 5.9% per year overthe next decade
(Erdem, 2010).These are some of the most reliableindicators of water demand (American Associate for
the Advancement of Science, 2010).Meanwhile available fresh water sources in Turkey are diminishing
rapidly due to pollution of fresh water supplies and over-extraction. Turkey has a pollution rate of 1.1 tons
3 st
of industrial organic pollutant per available km of fresh water, which ranks it as31 in the world
(Environmental Statistics, 2001). Moreover, even though theWater Exploitation Index has decreased
around 10% in 21 European countries overthe past 10-15 years, Turkey was one of the seven countries
that had an increasing rate.Turkey’sextraction has increased by 1/2 for public water supply and by 1/3 for
irrigation (the twolargest contributors of water extraction in the country) duringthe past decade,
representing the highest upward trend amongcountries in Europe(European Environment Agency,
3
2009).It is estimated that by2023 annual freshwater per capita will decline to under 1,000m per capita,
which is considered below water scarcity levels(General Directorate of State Hydraulic Works, 2009; CIA
website, 2010).

The fresh water problems of Turkey can be addressed inmany ways. While wastewater reclamation,
urban runoff and storm water capture can create significant impact on increasing the sustainability of
available water sources, the process ofdesalination seems more appropriate for the country due toits
geographical proximity to seawater andthe negative public perception of waste water reclamation,
primarily due to concernswith respect to human health (Dolnicar and Shafer, 2006). Desalination is a
relatively mature technology thatis free from variations in rainfall. It avoids the need fordual piping
networks, andhas a virtually unlimited supply of the primary feedstock (seawater).Desalination is likely to
play a major role in augmenting fresh water resources globally (the capacity is estimated to be in excess
3
of 5000 m /d with an annual growth rate of 12% overthe last 5 years) (GWI, 2006). Growth has
beennotablein Mediterranean countries.

The objective is to improve understanding of desalination options for the provision of fresh water in
Turkey in the future. More than 90% of commercial desalination technology is based on multistage flash
(MSF), multi-effect evaporation (MED) and reverse osmosis (RO), so these are the main technologies
selected here for evaluation for utilization in Turkey.Moreover, due to Turkey’s geographical proximity to
three major seas, only seawater desalination is taken into consideration. The main evaluation criteria
areenvironmental impact, cost and compatibility (in terms of geographical, structural and political factors)
with the country.

2. DESALINATION TECHNOLOGIES
Industrial desalination uses seawater (as well as brackish water) and a form of energy (most commonly
thermal, mechanical or potential) to produceseparate streams of nearly salt free water that can be
consumed by humans (or used forirrigation).The main waste is rejected brine. Desalination processes
can be phase change/thermal or membrane separation. Thermal desalination is one of the oldest ways of
producing potable water and MSF accounts for over 60% of all desalinated water, mostly in Gulf States.
3
The majority of plant unitsproduceover 5000 m /day. Thermal desalination requires little pretreatment,
ishighly robust and can produce water with very low salt content (no more than 10 to 20 PPM). At the
same time there has been a significant increase in RO plants, mostly in the US, Spain and Japan. RO
technology consumes less energy than thermal desalination methods but usually requires more
pretreatment than thermal methods (especially forseawater with high salinity) and produces water with
salinities of around 200 to 500 PPM. In recent decades RO technology use has become significant world-
widedue to its lower energy requirements, environmental compatibility, ease ofcontrol and scale-up,
flexibility and simplicity (Water Desalination International Website, 2010;Cipollina et al., 2009).

2.1 Multiple-Effect Distillation

Multiple effect distillation, one of the oldest methods for desalination, involves boiling seawater in
evaporators (called effects) and condensingthe released steam to produce nearly pure water. The feed
seawater is preheated in the evaporator tubes and to the boiling point. Tubes are heated by externally
supplied steam which can be recycled as feed water and condenses at the other side of the tubes. Next
the seawater is sprayed onto the surface of evaporator tubes in a thin film to promote rapid boiling and

370
evaporation of the portion of the seawater applied.These effects are connected together in a sequence
with decreasing levels of pressure where the vapor formed in one effect istransferred to be utilized in the
next to evaporate a portion of the remaining seawater.Therefore the seawater intake is allowed to
undergo multiple boiling steps without the need of adding heat after the initial effect, leading to a high
thermodynamic efficiency. This system allows for low brine output and prevents rejection of large
amounts of energy, which is the main drawback of the single effect distillation system(El-Dessouky et al.,
2002). MEDoperates at low concentration and low temperature (less than 70°C), and can produce
steadily high purity distillate (between 1 to 50 PPM total dissolved solids). The MED process usually
islimited to smaller system capacities (Cipollina et al., 2009).The steam economy of the system increases
with increased number of effects, but islimited to the total temperature range available and minimum
temperature difference between the effects. Commonly 8 or 16 effects are carried out in a typical large
plant with performance ratios ranging between 10 and 18(Khawaji et al., 2008; Raluy and Uche, 2005).
3
The specific electricity consumption of typical MED plants is below 1.8 kW h/m . However this
consumption is insignificant compared the thermal energy consumption of the process.

2.2 Multi-Stage Flash Distillation

Unlike MED, MSF utilizes the liquid bulk as opposed to the surface of hot tubes in order to form vapor(El-
Dessouky et al., 2002). Seawater is flashed into steam in multiple stages by reducing the pressure in
each stagefor flashing at progressively lower temperatures (typically 2-5°C in each stage), without the
need of adding more heat.Seawater initially flows through heat exchanger tubes where it isgradually
heated to the inlet temperature of the brine heater by the latent heat of seawater flashing in each
chamber. Then, the seawater is further heated by the brine heater and enters the first stage of evaporator
flash chambers, with high temperature and pressure, where it immediately flashes and partially vaporizes
to reach the equilibrium conditions with the stage. The flashed vapor is drawn to the cooler tube bundle
surfaces where it is condensed and collected as distillate. The remaining seawater is then transferred into
the next stage where the same process is repeated about 19-28 times (in modern large MSF plants).The
produced desalinated water contains 2 to 10 PPMtotal dissolved solids and is re-mineralized through a
post-treatment process.Some portion of the cold seawater is used as acooling medium for the heat
rejection section and returned to the sea after applying scale-control measures. The plant efficiency
increases with the number of stages, although this increases plant capital cost. The upperoperating
temperature of brine is usually 90°C to 120°C. Increasing the temperature increases the efficiency of the
system, butalso increases the probability of scale formation and accelerated corrosion of metal surfaces
that are in contact with seawater (Khawaji et al., 2008;Raluy et al., 2006). The specific
3
electricityconsumption of the MSF plants istypically 4kW h/m ,which is significantly higher than for MED
plants but still lowcompared to thermal energy consumption.

There have been major technological advancements in MSF distillation plants over the past decades.
Plants have significantly higher reliability and life expectancies at lower costs due to technical
optimization (including experienced capitalization and better material alternatives) and less stringent
technical restraints, mostly regardingprocesses, materials and plant accessories(Borsani and Reblagliati,
2005).

2.3 Reverse Osmosis

Reverse osmosis is one of the fastest growing desalination methods and is a good alternative to thermal
desalination. It is a membrane separation process that produces aflow of near pure water by rejecting
large molecules, ion and the dissolved salts from seawater, as a result of applying an external pressure
higher than the osmotic pressure on the seawater (commonly with the use of high pressure stainless steel
electric pumps). During the process a small percentage (1% for new membranes) of the water leaks
around the seals, but this leak is acceptable for the water quality needed for human consumption and
industrial applications. Pretreatment of seawater is usually required to remove undesirable elements in
the seawater and to prevent their accumulation in the membrane (membrane fouling). The latter
phenomenon significant reducesthepermeate flux. Post-treatment usually includes PH adjustment,
addition of lime, and removal of dissolved gases and disinfection (Khawaji et al., 2008). The performance
of RO is characterized in terms of water flux, salt rejection and recovery rates (Cipollina et al., 2009)

371
Since there is no heating or phase separation associated with RO, the energy input for the process is
significantly less than for MED orMSF and mostly associated with pressurizing the seawater feed, which
is proportional to feed water salinity. An increase of 100 PPM TDS of the input water salinity increases
the feed pressure requirements by approximately 1 psi, increasing the energy requirement (Koklas and
Papathanassiou, 2006). Operating under low ambient temperatures also reduces problems due to
corrosion. A common large seawater RO plant includes feed water pre-treatment, high pressure pumping,
membrane separation and permeate post-treatment. The majority of the problems faced by the RO plants
occur duringpretreatment, due to the inadequacy of conventional filtration methods.

Over the past two decades, the RO process has undergonesignificant technological developments and
cost reductions. Most of the improvements are associated with the membranes, whichnow have better
resistance to compression, longer life, higher possible recovery, improved flux and improved salt
3
passages. The typical energy consumption of ROplants is 6 to8 kW h/m without energy recovery, but can
likely be reduced significantly usingcurrent innovations in energy recoverydevices.

3. RENEWABLE ENERGY POTENTIAL OF TURKEY

Even though desalination can help reduce water scarcity significantly, especially in areas where other
water supply alternatives are not available, it requires a great amount of energy which is mainly obtained
from fossil fuels at present. This imposes problems since fossil fuels are both harmful tothe environment
and are costly in Turkey; therefore alternative energy sources are worthanalyzing. Turkey is a country
with an abundance of many types of renewable energy sources, and the three major RESs of the country,
namely wind, solar and hydro,are analyzed in this study. The availability, technological and economic
readiness and compatibility of RESsfor use withdesalination methods are examined in this sectionin order
to provide the background, current trends and future predictions forRES utilization in Turkey,which is
requiredbeforethe most appropriate desalination method and energy usage can be assessed.

RESs have increased in importance with concernsoverenvironmental and sustainability issues associated
with conventional energy sources. RESsoften exhibit a relationtoregions where desalination is most
needed, since these tend to be the regions where RESs like wind or solar are most abundant (e.g.,
coastal areas with warm climates and/or remote regions where connection to the electrical grid is not
feasible or cost effective). However, several additionalcriteria, like the amount of water needed, feed-
water salinity and technical infrastructure, also need be taken into account when matching RESs with
desalination technologies (Karaghouli et al., 2009).

Renewable energy sources are characterized by intermittent and variable intensity as opposed to
desalination plants that require steady energy inputs. However, in many cases, a combination of multiple
RESs can be used to account for the variations in energy inputs ofeach source if multiple sources are
available. RESscan also be used for storing the excess electricity produced that is not used by
desalination in the grid, forlater use when the RES power decreases, if “grid-connected” systems make
more economic sense (Zejli et al., 2004).

Today conventional energy sources are expectedto have an increasing cost trend due to on-going
depletion of available fossil fuels as well as possible introduction of fiscal penalties for power plants
emitting carbon dioxide (Sanden and Azar, 2005). On the other hand, RESs have the opposite trend as a
result of various monetary incentives (including tax cuts) and continuous advancements in technology
that reduce their investment costs. The trend in low/high investment and generations costs (based on
lower and upper cost estimation ranges) of RESs in the past decade are shown in Figure 1.

Although all RESs account for only 12% of the total energy consumption of Turkey, RESsin Turkey have
the biggest potential in Europe, especially for hydro, wind and geothermal energy, and very good
potential in terms of biomass and solar energy (Evrendilek and Ertekin, 2006; Yuksel, 2008; Ediger and
Kentel, 1999; Ozgur, 2008). Moreover, although the current utilization of RES low, many suggest Turkey
has both the potential and the political willingness to utilize RESs more in the future. The renewable
energy potential of the country can be seen in Table 1.

372
In Turkey the average capital costs (the most significant costs associated with RESs) (Raluy and Uche,
2005; Karaghouli, 2009) are even lower than the world average due to the abundance of economically
exploitable sources, recent legal incentives, low construction and labor costs. Average capital costs for
RESs arecompared in Table 1.

3.1 Wind Energy

Wind-power technology has been improved significantly over the last decades and has begun to be used
to supply electricity or mechanical power to small scale desalination plants in various locations in the
world. In addition to technological readiness, it is also a highly compatible method since fresh water can
be produced when there is an available power supply and can be easily stored when there is not,
counteracting with the fluctuations in wind speeds without the need for expensive back-up systems.

5000
4500
Investment Cost ($/kW)

4000
3500
3000
2500
2000
1500
1000
500
0
Wind power Solar - Solar – Small hydro Bio-power Geothermal
Photovoltaic Thermal power power

2002 2010

(a)
20
18
Generation Cost ($/kW)

16
14
12
10
8
6
4
2
0
Wind power Solar - Solar – Small hydro Bio-power Geothermal
Photovoltaic Thermal power power

2002 2010

(b)

373
 8000
Investment Cost ($/kW)
7000
6000
5000
4000
3000
2000
1000
0
Wind power Solar - Solar – Small hydro Bio-power Geothermal
Photovoltaic Thermal power power

2002 2010

(c)
60

50
Generation Cost ($/kW)

40

30

20

10

0
Wind power Solar - Solar – Small hydro Bio-power Geothermal
Photovoltaic Thermal power power

2002 2010

(d)
Fig. 1.Low scenario (a) investment and (b) generation costs (in $/kW) and high scenario (c) investment
and (d) generation costs forelectricity production using RESs between the years 2002 and 2010 (data
obtained from Erdogdu, 2009).

Table 1: Turkey’s RESpotential and share of renewablein the country’s total production and consumption for major energy
sources as well as capital costs associated with the RESs(data used from Erdogdu, 2009; Ediger and Kentel, 1999; Toklu et al.,
2010).
Share of RES in Share of RES in World Turkey’s
Usage Natural Tech- Econo-
Energy type 1 Turkey’s total Turkey’s total capital capital cost
purpose capacity nical mical 2 2
production (%) consumption (%) cost (avg) (avg)

Wind energy Energy 0.06 Energy 0.96

Electric 400 110 50 920 900
(land) Exergy 0.05 Exergy 0.01
Electric 977,000 6105 305 Energy 2.41 Energy 0.32
Solar 5400 N/A
Thermal 80,000 500 25 Exergy 2.42 Exergy 0.31
Energy 22.15 Energy 2.96
Hydropower Electric 430 215 124.5 1000 650
Exergy 23.99 Exergy 3.02

374
 14.82
Electric 109 -- 1.4 Energy Energy 1.98
Geothermal 1230 1440
Thermal 31,500 7500 2.843 Exergy 4.64 Exergy 0.58
Energy 13.00 Energy 1.74
Biomass Total 120 50 32 N/A NA
Exergy 14.77 Exergy 1.86
1
The units for electricityis measured in Billion kWh and thermal energy is measured in Mtoe.
2
Capital costs are in terms of €/kW

When the desalination technologies considered in this study are examined, RO technology seems to be
most compatible with wind energy due to its modular nature that allows the process to adapt to
fluctuations in wind conditions and its high energy efficiency that leads to low specific energy
3
requirements (as low as 2 kW h/m ). The technology also has great variety of production capacities and
fast start up and stop times that are compatible with wind energy (Koklas and Papathanassiou, 2006).
Moreover, it is also possible to provide the RO system with valves that can dynamically
activate/deactivate RO vessels to adjust the energy demand tothe available wind energy (Forstmeier,
2007).

Wind energy has the potential tobe an important substitute forconventional energy sources for Turkey,
since Turkeyhas the highest share with 166 TW/year in technical wind energy potential among European
countries, enough in fact to meet all of Turkey’selectricity needs. The energy costs associated with
desalination plants decreases significantly with the implementation of wind energy (since there are no fuel
costs), and the capital cost(about 75% of total plant cost) is mainly associated with the wind turbines
(50% of the capital cost) (Koklas and Papathanassiou, 2006).Even though the cost of electricity and water
produced through this technology depends significantly on plant size and location and as well as various
economic, political and site-specific factors, it is estimated that the cost of electricity produced by a 500
kW wind turbine with a maximum efficiency of 30% in a geographical location similar to Turkey would cost
$0.07 per kWh (Evrendilek and Ertekin, 2003).

3.2 Solar Energy

Solar energy is one of the most promising RESs and includes solar thermal andphotovoltaics (PV)
technologies. In solar thermal systems, afluid is heated by solar radiation as it circulates along the solar
collectors (flat or parabolic) through an absorber pipe. It can be stored in an insulated tank or used in the
high temperature end of a standard thermal distillation process. PV technology convertssolar energy
directly into electricity by using sunlight toexciteelectrons in the cells and generate electrical charges
thatconducted away as direct current (Garcia-Rodriguez et al., 2002; Lamei et al., 2008). Solar thermal
energy systems are starting to be used with MED and MSF processesin large scale applications (from 5
MW to 345 MW) as well as with many RO systems. PV systems are used more often with RO methods in
3
relatively small applications (under 5 MW, corresponding to desalination plants under 15,000 m /d
capacity).

Photovoltaic technology has many advantages such as the capability ofintegrating with small desalination
plants, limited maintenance costs and ease of transportation and installation (Karaghouliet al., 2009),and
is usually coupled with RO systems due to the ease of integration. The main drawback comes from
storing the energy when solar energyis insufficient and sustaining the system operation using batteries.
PV technology still iscostly compared to other RESs.Turkey has substantial gross solar potential of 88
billion toe per year, with an average annual sunshine duration of 2640 h and an average solar intensity of
-2 -1
3.6 KWh h m day (Ediger and Kentel, 1999).
3
In cost analyses of RO plants in the Middle East with a capacity of 10,000 m /d and specific electricity
demand of 3 kWh/m, the price of electricity produced is calculated to be under $0.09 for solar thermal
systems and $0.50 for PV (Lamei et al., 2008). Environmentally, even though solar energy has
significantly less impact than conventional energy sources, there are still some potential negative effects
due to the materials and chemicals utilized in the technology. Solar thermal receivers use toxic chemicals
in theheat transfer system, which pose threats in cases of accidental or emergency releases to the
environment. PV systems also use highly toxic chemicals that require rigorous disposal and recycling
procedures.

375
3.3 Hydropower Energy
Hydropower contributes 20% of the world’s electricitygeneration, providing the majority of supply in 55
countries (Yuksel, 2010). It is derivedfrom the energy produced through natural flow of water, turning the
blades of turbines that are connected to a generator. The energy produced is proportional to the volume
of the water flow and the height from which it is falls. It can beneficiallycomplement power sourcesthatare
more intermittent like wind and solar since the generating capacity can be reserved during peak supply
periods(through regulation of water flow) and/or when the capacities of other RESs are limited (The
Pembina Institute, 2010). Hydropower is one of the most efficient technologies for electricity production
since modern turbines can convert as much as 90% of the available energy into electricity, which is
significantly higher than the 50% efficiency of modern fossil fuel plants.During its lifetime a hydropower
plantcan produce 200 times the energy needed to build it.

Hydropower technology is more compatible with RO desalination than other renewable energy sources.
Instead of converting the potential energy of water into electricity through the use of turbines and a
generator (which have energy loses of 16% and 5% of theoretical hydro-potential respectively) and then
converting the electricity to hydraulic pressure, the potential energy in the trunk main can be used to
provide hydraulic pressure in the pressure pumping unit of the RO system seamlessly. Since the energy
requirement of the pressure pumping system is one the major costs of RO plants, removing this stage
canlower initial capital costs, construction timesandenergy requirements and costs(Murakami, 1995).But
hydropower systems are generally built in areas with natural waterfalls, where freshwater is usually
abundant, and hence desalination maynot be needed (Raluy et al., 2005). Turkey has the second largest
gross hydroelectric potential in European countries after Norway, with 16% of total theoretical hydropower
potential of all Europe. It is the most important RES used for electricity production in Turkey, comprising
96% of all electricity produced by RESs in 2008 (Erdem, 2010).

Aside from the abundant availability of hydropower, there are also financial incentives created by the
Turkish government. The May 2005 Renewable Energy Law (Law No. 5346) provides an 85% discount
on all forest land acquisition to build small hydropower,on top of other incentives created for all RESs,
reducing capital costs significantly. Moreover, the operating and investment cost of small hydropower is
also significantly lower in Turkey than in most of the European countries, mostly due to lower construction
and labor costs. Hydroelectric plants have significantly longer economic lives (about 75 years) than most
other plants, making them even more economical in the long run. The technology is highly
environmentally benign (except for potential fish passage blockages and sediment transport interruption,
as well as issues with dams) with additional environmental advantages from the prevention of river
erosion in most areas (Kucukali and Baris, 2009).

3.4 Other Renewable Energy Sources

Biomass energy is a major renewable energy source in Turkey, especially in the rural regions, with
estimated recoverable bioenergy potential of 35.4 Mtoe (Yuksel, 2008). In crop residues, Turkey is
th
estimated to havethe 4 highest total energy potential (9.5 Mtoe) among the OECD countries. However
this energy source is not consideredinthis study due to a significant declining trend in biomass use in
Turkey,with its total energy consumption contribution reduced from 20% to 8% between 1980 and 2005.
This is primarily due to scarcity-of-supplyand environmental concerns, the latter becausebiomassis a
contributor of air pollution, e.g.,biomass combustion releases over 200 chemical pollutants including 14
carcinogens and 4 co-carcinogens into the atmosphere (Enrendilek and Ertekin, 2003)and contributes
toover-timbering anddeforestation. The share of biomass in RES useis expected to continue to decrease
in the future (Yuksel, 2010).

Turkey also has significant geothermal energy potential and is among the topseven countries in potential,
th
possessing 1/8 of the world’s total geothermal potential (Erdogdu, 2009). However, it is ranked last
ininstalled capacity (28 MW e) (Ozgur, 2008), and only 6% of this potential is of suitably high temperature
and enthalpy for electricity production (i.e.,temperatures over 150°C). Thus, the geothermal electricity
generation potential of Turkey is estimated to be only 16TWh/year. Few reliable studies are currently
available regarding economic and environmental effects of using geothermal energy withdesalination
technologies.

376
4. SELECTION CRITERIA
When the global distribution of various desalination plants and their respective energy sources are
examined, one of the major deciding factors in application is the availability of resources (environmental
and/or technical) and cost. The majority of desalination plants in the Middle Eastern region usethermal
desalination technologiesincorporatingconventional energy sources (fossil fuels, natural gas and
coal),due to the lowerenergy costs and simpler and more reliable nature of the technologies (compared to
membrane separation).Mostof these countries have not implemented significant renewable energy or
nuclear energy technology atlarge scales. Europe on the other hand has a significantly wider range of
energy production methods (43% thermal, 40% nuclear and 17% hydropower (Raluy et al., 2006)) and
has started to incorporate renewable energy resources. Turkey does not seem to fit in either option. Even
though geographically Turkey lies in the Middle East and like nearby countries reliesmostly on
conventional sources for energy production (31.8% natural gas, 29.9% fossil fuels and 27.3% coal, which
makes up about 90% of total energy consumption), the country does not have significant
sourcesavailable and dependsheavily on importing.Currently 97% of Turkey’s natural gas, 93% of its oil
and 20% of its coal use is imported (Erdem, 2010).General trends regardingproduction, import and
consumption suggestthat Turkey will import 99% of its oil within 20 years (Evrendilek and Ertekin, 2003).
Therefore the cost associated with desalination using conventional energy resourceswill likely be
considerably higher compared to most Gulf States. Even though Turkey has the largest potential in
Europe for RESs such as hydro, wind and geothermal and very good potential in terms of biomass and
solar, the utilization of these resources islow(12% of energy total energy consumption combined (Ediger
and Kentel, 1999)). Here, weexamine and compare desalination technologies and various energy sources
with respect to its environmental impact, cost and other criteria.

4.1 Environmental Impact

Over the past century, during an era of industrialization(Fiksel, 2009), humans have left a significant
footprint on the environment. With unplanned industrial growth and the false notion of unlimited
resources, we have reduced many of Earth’s important natural resources significantly and
negativelyimpactedthe environment. These actions have resulted in problems such as climate change
and global warming, decreases of biodiversity. In past decades,people have increasingly becomeaware
of environmental concernsacross the world. Today these concerns often play a role in people’s decision
making and many studies are being conducted to understand the negative environmental impacts of our
actions and ways to prevent them. Due to the increased complexity of today’s processes and the mix of
materials and energy sources utilized, understanding the exact relationship between actions and
associated environmental impacts isincreasingly difficultyand methodsare required to analyze these
relationships further to improve understanding of the technologies.

In determining the environmental impact of the studied technologies,the main focus is on the operation
phase of the life cycle of the plant since many studies show that the environmental load associated with
this stage is 89 to 99% higher than that of assembly and final plant disposal phases (Borsani and
Rebagliati, 2005; Karaghouli et al., 2009). Since the main contributor to environmental effects inthis stage
is the amount and source of energy used bydesalination technologies, the primary focus is the energy
usage of the desalination technologies andthe potential of alternative energy sources. The associated
environmental impact forMED, MSF and RO technologies are shown inFigure 2. Note that only airborne
emissions are taken into considerationsince these are typically associated with energy production
systems, are oftenregulatedby theeach country and are emitted in higher quantities than most other
substances (Raluy and Uche, 2005).

Figure 2 shows that MSF desalination has the highest impact on the environment in all categories
considered, followed my MED with slightly lower airborne emissions and RO with significantly less impact
than the first two. A strong correlation is observed between energy consumption and airborne emissions
3
in the first 3 columns. Thermal desalination methods usually consume 50-70 kWhth/m of distilled water
(thermal equivalent of producing 8 to 12 kg of freshwater per kg of heating steam) and have an order of
magnitude larger impact on airborne emissions than RO. The significantly higher energy consumption for
thermal desalination technologies creates an order of magnitude larger environmental impact than RO
technology.

377
 Emissions per m3 desalted
30
25
20
water

15
10
5
0
MSF MED RO (high MSF (w/ waste MED (w/ waste RO (low capacity)
capacity) heat) heat)

CO2 (Kg) NOx (g) NMVOC(g) SOx(g)

1 1 2
Fig. 2. Airborne emissions produced by MSF , MED and RO desalination systems.Data are obtained for
MSF and MED from the research project REN 2001-0292 that took part in Spanish Plan for Scientific
Research and Technological Development and Innovation (R & D & I) (Raluy et al., 2006; Borsani and
Rebagliati, 2005; Koklas and Papathanassiou, 2006). Values for RO are based on potential energy
consumption with respect to current technological improvements (Cipollina, 2009).

The high energy consumption of MSF and MED plants can be reduced by utilization of waste heat from
other processes, since desalination plants and production plants can often integrate in mutuallybeneficial
ways for this purpose. Most production plants require large water intakes for cooling waterand
rejectconsiderable amounts of waste heat to the atmosphere. Desalination plants can use the waste heat
to provide water,some of which can be used for the production plant cooling. This canreduce the
environmental impact of the thermal desalination methods significantly, often enough to
makethemcomparable to RO technology.However, it should be noted that even though utilization of waste
heat can potentially reduce the environmental effects of the desalination technologies, they require the
plant to be built in the vicinity of production plants, which limits the potential locations, especially since in
some cases fresh water is needed in rural regions away from industrialized areas.

Since RO plants consume almostno thermal energy (compared to MSF and MED) for desalination, waste
heat utilization is not compatible with this technology to reducethe environmental effects. However,
because of the latest improvements in energy consumption technology in RO plants and commercial
introduction of “isobaric devices”, the associated consumption can be reduced as much as 46%, from 3.9
3 3
kWh/m to 2.1 kWh/m (with a 6.0MPafeed water pressure and 50% recovery rate) (Cipollina et al., 2009),
reducing the overall impact on the environment.
5
Emissions per m3 desalted water

4,5
4
3,5
3
2,5
2
1,5
1
0,5
0
Wind Energy PV Energy2 Hydro-Power Energy

CO2 (Kg) NOx(g) NMVOC(g) SOx(g)

(a)

378
 10
Emissions per m3 desalted water
9
8
7
6
5
4
3
2
1
0
Wind Energy PV Energy2 Hydro-Power Energy

CO2 (Kg) NOx(g) NMVOC(g) SOx(g)

(b)
Fig. 3. Environmental Impact of RO technology incorporating various RESsfor (a) low and (b) capacity
(data obtained from Raluy et al., 2006; Raluy et al., 2005).RO data are obtained from the research project
REN 2001-0292 that took part in Spanish Plan for Scientific Research and Technological Development
and Innovation (R & D & I). The environmental load associated with RO technology using PV is
considered based on solar radiation data extrapolated from a plant used in Spain. The minimum and
maximum emission values are considered for RESs with respect to a conventional RO plant with 4
3
kWh/m energy consumption.

The overall environmental impact of the desalination technologies can be reduced further by integrating
them with of renewable energy sources. The extensive and increasing use of fossil fuels within the
current energy infrastructure is considered the largest cause of CO2 emissions, which are believed to be
directly tied to global warming, climate change and other environmental concerns. These negative
impacts can be reduced significantly through the utilization of various abundant RES in Turkey that have
little negative impact compared to conventional energy sources. Due to its low energy requirement and
modular nature, integration of RESswith desalination is primarily studied/used for RO since, unlike MED
and MSF technologies, there exists much reliable data regarding environmental impact of the integrated
technology. Therefore, integration of RES is only analyzed with respect to RO in this study. The
improvements possible,for the previously discussed emissions,are shown in Figure 3.It is seen that even
though the environmental impact of RO technology is significantly lower with respect to thermal
desalination technologies, the emissions can be reduced further, by up to 95% in some categories, via
integration of renewable energy sources. It also is observed that, among the RESs selected, hydropower
provides the greatest reduction in environmental impact in every category considered, with a reduction of
over 90% on average.

4.2 Cost
Another important criterion for determining the most suitable desalination technology (and associated
energy source)is based on the cost of the overall system. Due to insufficient studies and lack of reliable
data on cost estimates for MED, MSF and RO technologies in Turkey, the analysis is based onreal data
for various industrial desalination plants around the world as well as studies conducted for similar
geographical regions and countries that have similarenergy and economic potential toTurkey. The main
comparison arebased on unit water production costs. Some recent estimates of water costs of various
desalination technologies selected around the world are shown in Table 2.The unit water production cost
3
of MED/MSF and RO technologies are seen to be very similar ($0.81/m on average for MED/MSF and
3
$0.86/m for RO technologies based on plants provided in the table). The data suggests that, even
though RO plant unit water costs have dropped significantly in the last decades, they still are slightly
higher on average than for thermal plants on a unit water cost basis.

379
Table 2: Total water costs of selected industrial desalination plants(data obtained from Cipollina et al., 2009; El-Dessouky and
Ettouney, 2002; Khawaji et al., 2008).
Plant (location) Technology Water cost Plant capacity
3 3
($/m ) (m /d)
Shuweihat (UAE) MSF 1.13 454,610
RasLaffan (UAE) MSF 0.80 272,520
Hidd (UAE) MSF 0.69 400,00
Tenes (Algeria) SWRO 0.59 200,00
Taunton (Massachusetts, US) SWRO 1.53 18,925
Palmachim (Israel) SWRO 0.86 83,270
OuedSebt (Algeria) SWRO 0.68 100,00
Hadera (Israel) SWRO 0.86 330,00
Ashkelon (Israel) SWRO 0.78 326,144
Tianjin (China) SWRO 0.95 150,000
Dhekelia (Cyprus) SWRO 0.88 40,000
Carlsbad (California, US) SWRO 0.77 189,250
Pert (Australia) SWRO 0.75 143,700
Marafiq (Saudi Arabia) SWRO 0.83 758,516
Shoaiba 3 (Saudi Arabia) MSF 0.57 881,150
Reliance refinery (India) MED 1.53 14,400

Note that, although unit water costs provide an important tool for comparing different technology costs,
other factors should not be neglected. The majority of large capacity thermal desalination plants are in the
Middle Eastern region. Thus all MSF plants in Table 2 are in either the United Arabic Emirates or Saudi
Arabia due to the availability of data for these plants. The primary energy sources for these plants are
generally fossil fuels, which are significantly cheaper than for most other locations in the world;the unit
water costs would be drastically higher if the plants were in most parts of Europe. Therefore these costs
are considerably higher in Turkey when powered by fossil fuels.

Another factor is that, due to the nature of MED and MSF technologies, plant capacity is significantly
higher than for most RO plants. Since higher capacity is associated with lower unit costs due to
economies of scale, capacity should be considered when comparing unit water costs. Larger plants
require significantly higher investment costs and usually pose higher initial risks than relatively smaller
ones. For these reasons, the costs show a bias towards MED and MSF technologies and therefore the
actual gap between RO and MSF/MED technologies is considerably smaller in reality when implemented
in Turkey.

Beyond data from actual plants, many studieshave been conducted regarding estimating the cost
associated with these threedesalination technologies. These studies facilitateuseful comparisons based
on diverse geographical locations and similar capacities,allowingfair assessments. Table 3presents costs
databased on combinations of several major independent studies in the literature.

380
Table 3: Analysis of unit production costs based on various studies in the literature
Reference Process Capacity Unit production
3 3
(m /day) cost ($/m )
Borsani and MED 205,000 0.52
1
Rebegkiati (2005) MSF 205,000 0.52
RO 205,000 0.45
Karagiannis and MED 12,000-55,000 0.95-1.95
Saldatos (2008) MSF 23,000-528,000 0.52-1.75
RO 12,000-60,000 0.44-1.62
2
N. Wade (2009) MED 31,822 0.95
MSF 31,822 1.04
RO 31,822 0.82
G. Fiorenza (2003) MED 10,000 0.8
MSF 25,000 1.10
RO 6,000 0.7
Frioui and MED 1,000 1.38-1.45
3
Oumeddour (2008) MSF 1,000 1.2-1.34
RO 1,000 1.8
1
Cost estimates are based on 20 year plant life, recovery ratio of 40% and membrane replacement of 5 years (for
RO), steam factor time around 85% (for MED and MSF) and fuel price of natural gas $4.5/MWh(th), and for seawater
desalination.
2
Cost estimates are based on Mediterranean seawater feed of around 37,000mg/l total TDS.
3
Cost estimates are based on 30 year plant life, 100% production rate and steam factor time around 85%.
3
It can be seen in Table 3that the average unit cost estimated for MED, MSF and RO plants are $1.03/m ,
3 3
$1.01/m and $0.96/m respectively, which are similar. The table also shows that, at higher capacities,
RO technology can not only compete, but be cheaper, than MED and MSF technologies based on unit
production costs.

The studies also suggestthat RO technology can be economical whenincorporating RESs. Based on
analyses in several studies, the range of unit production costs of the examined RESs with various
capacities are shown in Fig. 4 (Karaghouli et al., 2009; Karagiannis and Soldatos, 2008; Mohameda et
al., 2005). Note that the high end production costs refer to experimental systems.

The large gap between high and low cost estimates with respect to each RES utilized with RO
desalination technology is due to factors like quality of renewable energy sources, government incentives,
transportation costs and production capacity. Most of these systems are built as prototypes and are hard
to compare on a cost basis. Even though these costs are still higher than utilizing conventional energy
source, the associated costs of using RESs have been declining significantly over the past decades and
are expected to continue to do so. It can be seen that, among the RESs, RO can provide the lowest cost
when integrated with hydropower. This cost average is expected to be even lower for Turkey due to the
large availability of economically exploitable hydro sources, cheap labor and land costs for construction
(capital cost of hydro is overwhelmingly larger than all other costs) and available technical expertise (96%
of current RES usage already comes for hydropower (Erdem, 2010)).

381
 16
14
Unit Production Cost ($/m3)

12
10
8
6
4
2
0
Wind Energy Solar Energy - Solar Energy – Hydro Energy
Photovoltaics Thermal

Low Capacity High Capacity

Fig. 4.Unit production cost ranges for RO technology integrated with various RESs.

4.3 Other Considerations

Through analyzing the environmental impacts and overall costs associated with building various
desalination technologies in Turkey, it is seen that using RESs provides significant advantages with
respect to both criteria. However, the benefits of implementing RESs for desalination technologies are far
greaterthan the studied criteria for the country. As previously mentioned, the dependence of fossil fuels is
one of the biggest concerns of Turkey, due to it comprising a considerably large portion of the country’s
energy use(90%) and very small portion of its energy reserves (under 1% of the world). This leads to
importing almost all of its energy needs from outside the borders (97% of natural gas, 93% of oil and 20%
of coal), the majority of them being a single sources. Historically Russia has been the largest gas supplier
(63% of the total natural gas imports (Akgul et al., 2008)) and Saudi Arabia has been the largest supplier
for crude oil (with 110,000 b/d (Turkish Market Case, 2012)). Dependence on a single energy resource
that dominates the country’s consumption ispoliticallyriskyand poses national security concerns.
Therefore a transition to RESscanhelp reducethis threat. Meanwhile as the country seeks energy sources
within its own boundaries, such actionsalso help boost thelocal economy and createlocal jobs.

In the past decade, the use of RESs as an alternative to fossil fuels has been promoted and encouraged
in Turkey through several legislations and such measures will be continued (Kaya, 2006; Ozyurt, 2010).
The Electricity Market Law (EML, No. 4628, enacted in February 2001) helps legal entities applying for
licenses for construction facilities based on RESs, by permitting them to pay only 1% of the total licensing
fee and avoidsthe requirement to pay the annual license fee for the first 8 years following the facility
completion date in their licenses. In May 2005 the government passed the Renewable Energy Law
No.5346 aiming to expand the use of RESs for providing electricity by guaranteeing the purchase of the
electricity produced from RESs (at a price of 7.4 cent/kWh with a 20% price increase at the beginning of
each year (Kucukali and Baris, 2009; Erdogdu, 2009)) and making electricity licensees purchase at least
8% of their annual electricity sales based on these RESs. In addition, in the case of the public land and
forest use for the electricity production with RESs, the lands could be rented or awarded access by the
Ministry of Environment and Forestry and/or Finance Departments. The Law also gives incentives
toresearch and investment of energy generating facilities. Finally, Turkey is to be recipient of a US$202
million renewable energy loan by the World Bank for building renewable energy sourced electricity
generation; the loan is expected to finance up to 40% of the capital costs (Toklu et al., 2010).

On the other hand there are also certain barriers for utilizing RES in Turkey, the main one being the lack
of co-ordination and co-operation within and between various ministries, agencies, institutes and other
stakeholders (Kaya, 2006). However with the on-going depletion of conventional energy sources and

382
ever-rising oil prices and governments approach to increase the utilization of RES, these issues would be
resolves as RES becomes a major contributor to its overall energy consumption.

5. CONCLUSIONS
The study analyzedvarious actual MED/MSF and RO desalination plants around the world as well
asmany studies conducted on these technologies andtheir integration withRESs. The evaluation was
primarily based on costs and environmental impactsof the desalination technologies and associated
energy sources. The analysis shows that RO technology would be the best option for Turkeyfor the
specified plant capacity based on the following points:
• The RO method is highly competitive with MED and MSF methods, reducing the cost by 13% and
21% for MED and MSF respectively, for the midsize plant capacity and could cost even less in high
capacity plants. Moreover, continuous improvements in technology for both RO systems (especially
the membranes) and integration of RES indicate that RO technology will be much more competitive
with MED and MSF technologies in terms of cost.
• The cost of thermal desalination using fossil fuels in Turkey would be much more expensive than
Middle Eastern regions presented due to cost of fossil fuels in Turkey being significantly higher than
in the Middle East.
• The cost of RO desalination in Turkey will be significantly cheaper than in the Middle East due to
Mediterranean seawater feed (around 37,000 mg/l TDS content) being considerably less than for the
Arabian Gulf Region (65,000 – 75,000 mg/l TDS).
• RO desalination technology can reduce emissions of CO2, NOx, NMVOC andSOx by as much as
38%, 42%, 73%, 60% respectively compared to alternative technologies, even when waste heat
recovery is used. ROis highly compatible with RESs due to its low energy consumption (5-6 times
lower than other desalination methods) and modular nature.
To reduce the long term cost and environmental impactfurther, hydropower energy source is determined
to be the best option for integrating with RO based on the following points:
• Hydropower has seamless integration with RO technology and has the lowest potential unit
production cost and the least environmental impact on airborne emissions in every emission category
which can further reduce the emissions of CO2, NOx, NMVOC, SOxby 83%, 82%, 80% and 50%
respectively.
• Theactual cost of production is estimated to be even lower for Turkey due to the high number of
existing economically exploitable resources, construction and labor coststhat are much lower
thanaverage,experienced capitalization and various government incentives and can help reducing the
country’s dependency on fossil fuel imports and gain economic and political strength in the region.

Acronyms
CO2 Carbon dioxide
MED Multiple-effect distillation
MSF Multi-stage flash distillation
NMVOC Non-metal volatile organic compounds
PV Photovoltaics
RO Reverse osmosis
SOx Sulfur oxide
MSF Multi-stage flash distillation
RO Reverse osmosis

REFERENCES
Akgul D., Cakmakci M., Kayaalp N., Koyuncu I. 2008. Cost Analysis of seawater desalination with reverse
osmosis in Turkey. Desalination 220: 123–131.

383
Bilgili M., Sahin B., Kahraman A. 2004. Wind energy potential in Antakya and Iskenderun regions, Turkey.
Renewable Energy 29: 1733–1745.
Borsani R., Rebagliati S.2005. Fundamentals and costing of MSF desalination plants and comparison
with other technologies.Desalination 182: 29–37.
Central Intelligence Agency.The World Factbook.www.cia.gov.tr, accessed on 20/10/10.
Cipollina A., Micale G. and Rizutti L. 2009.Green Energy and Technology.Seawater desalination.
Conventional and renewable energy processes. Springer-Verlag Berlin Heidelberg
Ediger V., Kentel E. 1999. Renewable energy potential as an alternative fossil fuels in Turkey. Energy
Conversion & Management 40: 743-755.
El-Dessouky T. and Ettouney M. 2002.Fundamentals of Salt Water Desalination. Elsevier Science B.V.
Dolnicar S., Schäfer A. I. 2006. Public perception of desalinated versus recycled water in Australia.
Journal of Environmental Management. 90: 888–900.
Erdogdu E. 2009. A snapshot of geothermal energy potential and utilization in Turkey.Renewable and
Sustainable Energy Reviews 13: 2535-2543.
Erdogdu E. 2009. On the wind energy in Turkey.Renewable and Sustainable Energy Reviews 13: 1361–
1371.
Evrendilek F., Ertekin C., 2003. Assessing the potential of renewable energy sources in
Turkey.Renewable Energy 28: 2303–2315.
nd
Fiksel J. 2009. Design for Environment. A guide to sustainable product development.2 Edition.
Fiorenza G., Sharma V. K., Braccio G. 2003. Techno-economic evaluation of a solar powered water
desalination plant.Energy Conversion and Management 44: 2217–2240.
Frioui S., Oumeddour R., 2008. Investment and production costs of desalination plants by semi-empirical
method.Desalination 223: 457-463.
Forstmeier M., Mannerheim F., D’Amato D., Shah M., Liu Y., Baldea M., Stella A. 2007. Feasibility study
on wind-powered desalination.Desalination 203: 463–470.
Garcia-Rodriguez L., Palmero-Marreroa A., Gbmez-Camachob C. 2002.Comparison of solar thermal
technologies for applications in seawater desalination.Desalination 142: 135-142.
General Directorate of State Hydraulic Works 2009.Turkey Water Report.http://www.dsi.gov.tr, accessed
on 18/10/10.
Human Development Report 2006.Beyond scarcity: Power, poverty and the global water crisis. United
Nations Development Programme(UNDP), New York, USA.
Environmental Statistics. 2001. Fresh Water Pollution by Countries. Based on World Development
Indicators.http://www.Nationmaster.com.
Erdem Z. 2010. The contribution of renewable resources in meeting Turkey’s energy-related
challenges.Renewable and Sustainable Energy Reviews 14: 2710–2722.
European Environment Agency Website.Use of freshwater resources (CSI 018).Published on
01/28/2009.http://www.eea.europa.eu.
GWI Market Profile.2007. Desalination Markets Report, accessed on 12/10/10.
Karagiannis I. C., Soldatos P. 2008. Water desalination cost literature: review and assessment.
Desalination 223: 448-456.
Kaya D. 2006.Renewable energy policies in Turkey. Renewable and Sustainable Energy Reviews 10:
152–163.
Khan W. Z. 2009. Desalination of raw water using a polyamide hollow fiber membrane.Desalination 244:
59–65.
Karaghouli A., Renne D., Lawrance L., Kazmerski L. 2009. Solar and wind opportunities for water
desalination in Arab Regions. Renewable and Sustainable Energy Reviews 13: 2397–2407.

384
Khawaji A., Kutubkhanah I. and Wie J. M. 2008.Advances in Seawater Desalination
Technologies.Desalination221: 47.
Koklas P., Papathanassiou S. 2006. Component sizing for an autonomous wind-driven desalination
plant.Renewable Energy 31: 2122–2139.
Kucukali S., Baris K. 2009. Assessment of small hydropower (SHP) development in Turkey: Laws,
regulations and EU policy perspective. Energy Policy 37: 3872-3879.
Lamei A., van der Zaag P., von Munch E. 2008. Impact of solar energy cost on water production cost of
seawater desalination plants in Egypt. Energy Policy 36: 1748–1756.
Mohameda E., Papadakis G., Mathioulakis E., Belessiotis V. 2005. The effect of hydraulic energy
recovery in a small sea water reverse osmosis desalination system; experimental and economical
evaluation.Desalination184: 241- 246.
Murakami M. 1995. Managing Water for Peace in the Middle East: Alternative Strategies Experimental
seawater reverse-osmosis desalination.United Nations University Press.
Murakami M. 2008. Hydro-powered Reverse Osmosis Desalination for Co-generation. Inaugural
Ceremony International Seminar on Efficient Water Use.Uruguay.
Ozyurt O. 2010. Energy issues and renewable for sustainable development in Turkey.Renewable and
Sustainable Energy Reviews 14: 2976-2985.
Raluy R. G., Serra L., Uche J. 2005. Life cycle assessment of desalination technologies integrated with
renewable energies. Desalination 183: 81-93.
Raluy R., Uche J. 2005. Life Cycle Assessment of Water Production Technologies.Desalination 183: 81 -
93.
Raluy G., Serra L., Uche J. 2006. Life Cycle Assessment of MSF, MED and RO desalination
technologies.Energy 31: 2361–2372.
Ozgur M., 2008. Review of Turkey’s renewable energy potential. Renewable Energy 33: 2345-2356.
Ozturk M., Bezir N., Ozek N. 2009. Hydropower-water a renewable energy in Turkey: Sources and policy.
Renewable and Sustainable Energy Reviews 13: 605-615.
Sanden B., Azar C. 2005. Near Term technology policies for long term climate targets-economy wide
versus technology specific approaches.Energy policy 33: 1557-1576.
Somers B. 2010. Energy Experts Say Water Use and Energy Consumption Linked, Urge Conservation,
American Association for the Advancement of Science.
The Pembina Institute.Sustainable Energy Solutions.Renewable Energy.Hydropower
www.pembina.org/re/sources/hydro-power accessed on 20/11/10.
Toklu E., Guney M. S., Isik M., Comakli O., Kaygusuz K. 2010. Energy production, consumption, policies
and recent developments in Turkey.Renewable and Sustainable Energy Reviews 14: 1172–1186.
Turkish Market Case. 2012. APS Review Oil Market Trends. www.AllBusiness.com.
Wade N. M. 2001. Distillation plant development and cost update. Desalination 136: 3–12.
Water Desalination International website.Key Desalination Facts
http://www.waterdesalination.com/factssample1.htm, accessed on 25/10/10.
WHO website.World Health Organization.Facts and Figures on Water Quality and Health
www.who.int/water_sanitation_health/facts_figures/en/print.html, accessed on 18/10/10.
World Water Council. 2000. World Water Vision. Making Water Everybody’s Business.Earthscan
publications Ltd. London, UK.
Yuksel I. 2008.Global warming and renewable energy sources for sustaining development in
Turkey.Renewable and Sustainable Energy Reviews 12: 372–396.
Yuksel I. 2010.Renewable energy hydropower for sustainable development in Turkey.Renewable and
Sustainable Energy Reviews 14: 2113-3219.
Zejli D., Benchrifa R., Bennouna A., Zazi K. 2004. Economic analysis of wind-powered desalination in the
south of Morocco.Desalination 165: 219-230.

385
 TORREFACTION OF AGRICULTURE AND FORESTRY BIOMASS USING TGA-FTIR-MS

N. Y. Harun and M. T. Afzal

Department of Mechanical Engineering,
University of New Brunswick, Fredericton, NB E3B 5A3 Canada
mafzal@unb.ca

ABSTRACT
The torrefaction of agriculture (switchgrass and timothy) and forestry (spruce and pine) biomass was studied using
simultaneously thermogravimetric analyzer (TGA) coupled with Fourier transform infrared (FTIR), and mass
spectrometer (MS). The chemical functional groups present in the gases were identified by FTIR and the
quantification of gaseous products was determined using MS at different torrefaction temperatures ranging from
200 to 290 oC. TG-FTIR and TG-MS techniques are paired to refine the identification of gases. TGA results
showed that the behaviour of the agricultural and forestry biomass was not the same due to their composition
variation. The decomposition of switchgrass took place at a lower temperature than other biomass. Both
switchgrass and timothy have two peaks of degradation rate compared to only one peak present for forestry
biomass. The FTIR analysis indicated that most of the chemical compositions present in the biomass are
decomposed at torrefaction temperature of 290 oC. The mass spectrometric analysis at torrefaction temperature
200 °C and 230 °C quantified the degradation of combustible gases: CH4, C2H4, CO and O2 around 20-30%.
Whilst at torrefaction temperature 260 °C and 290 °C the degradation of combustible gases were more than 30%.
Moreover, all gaseous products evolved from the torrefaction of agricultural and forestry biomasses were almost
similar in characteristics, but vary in proportions.

Keywords: Agricultural biomass, Forestry biomass, Torrefaction, TG-FTIR

INTRODUCTION
Canada is one of the countries in the world which generates abundant amount of biomass from agriculture and
forestry sector. Canada has about 42% of forest land and 6.8% of agricultural land from which about 42% of
residue is produced entirely from forestry and agricultural (Wood and Layzell, 2001). Hence, there exists large
potential to convert this biomass into valuable products for various applications. Thermo-chemical treatment is an
effective method to handle the issue of this huge biomass resource. This process has great advantages since it is
flexible in feedstock, produces different fuels and is environmental friendly.

Torrefaction is also emerging as one of thermo-chemical method to produce energy fuel by improving the biomass
properties (Acharya et al., 2012). The biomass is heated between temperatures 200 to 300 °C in an inert
condition.Torrefaction has been reported (Acharya et al., 2012) to improve the hygroscopic behavior, decay
resistivity, grindability, and higher heating value. Moreover, it can help to reduce the greenhouse gas emissions.
Assuming net calorific value of 6.2 MWh/tonne, torrefied pellets (4,533,724 metric tons) is estimated to save CO2
emissions of 21,081,816 tons, if the former is replaced by traditional pellets (5,736,549 metric tons) (Melin, 2012).
Despite this saving in emissions, the behavior of biomass torrefaction still needs serious research attention.

With regard to this, TGA–FTIR technique is the most widely used method to observe the thermal decomposition of
materials under various environments. Understanding about the kinetics and in depth analysis of mass loss with
the type of release of gas products (Fasina and Littlefield, 2012) is essential for the fundamental knowledge,
design, operation and control (Miranda et al., 2007) of torrefaction equipment and process. Therefore, the
influence of parameters on chemical composition of the torrefied product has been extensively studied. In
torrefaction, the lignin network is modified (Alen et al, 2002) (Tjeerdsma and Militz, 2005) (Nguila et al, 2007), and
hemicelluloses are strongly degraded (Nuopponen et al, 2004) leading to formation of carbonaceous material
within the wood (Nguila et al, 2007). The anhydrous mass loss during torrefaction could be a reliable and accurate
indicator to predict dimension stability and decay resistivity of pellet fuel (Hakkou et al, 2006) (Welzcher et al,
2007).

Some studies were found related to torrefaction of biomass using TG analysis to determine the torrefaction kinetic
of cellulose, hemicellulose, lignin and xylan (Chen and Kuo, 2011). The same author has also conducted
torrefaction of various biomasses (bamboo, willow, coconut shell and wood) using TG analysis (Chen and Kuo,
2010). Another study also reported which was focused on torrefaction of deciduous wood (beech and willow),
coniferous wood (larch) and straw using TG analyser (Prins et al., 2006). Very recently, a study on torrefaction of
wheat straw but using TGA/DSC configuration mode was performed (Shang et al., 2013). From this and other
literature survey, none has attempted to torrefy biomass using TGA coupled with FTIR. Since TGA coupled with
FTIR can be advantageous for continuous online gas analysis, the torrefaction thermal decomposition can be
studied in depth. It was also difficult to find torrefaction of agriculture (switchgrass and timothy) and forestry

386
(spruce and pine) biomass using TG-FTIR-MS technology. Keeping this in view, the study contributes new
knowledge about the characteristics of the gas evolved during torrefaction of various types of biomass. The main
objective was to determine the torrefaction behavior as well as simultaneously investigate the chemical functional
groups and quantification of the gas product. TG-FTIR-MS analyses of gas were carried out at different
temperature of the torrefied biomass. Comparison of torrefaction of agriculture and forestry biomass is presented
in this study.

MATERIALS AND METHOD

Four types of biomass were selected for torrefaction. Two were agriculture (switchgrass and timothy) and other
two were forestry (spruce and pine). The agricultural biomass samples were provided by NB Department of
Agriculture. The forestry biomass chips of size ranging from 10 to 50 mm were acquired from local sawmill. Prior to
the analysis, biomasses were first dried in the oven at 105 oC for 24 h, and then grounded to about 500 μm
particle size.

TG-FTIR experiment
A Nicolet 6700 FTIR Thermo Scientific connected to TGA Q500 TA Instrument was used to perform torrefaction
experiments. Approximately 10 mg of sample was loaded in a TGA crucible. The temperature was initially raised to
105 oC at heating rate of 20 oC/min and held for 5 min in order to remove moisture. At second stage the samples
were torrefied using same heating rate to a desired temperature. The inert atmosphere was maintained using
nitrogen gas at flow rate of 100 ml/min. Vapors released from biomass torrefaction were subjected to the FTIR
spectrometer through a transfer tube, which is heated to about 200oC to prevent the condensation of vapors on the
tube wall. The IR spectra were recorded with a temporal resolution about 2 s. The resolution of the collected
spectra was set tobe 1 cm−1 and the spectral range was set from 4000 to
−1
400 cm . At the carrier gas flow rate of 100 ml/min, it took about 50 s for the vapors to reach the FTIR cell from
the thermogravimetric analyzer. Therefore, there was time delay of about 50 s.

TG-MS experiment
The volatile products generated from TGA that was first scanned by FTIR spectrometer to detect the functional
group, and then were furthered transferred to the ionization source of the mass spectrometer, SRS RGA 200, in
order to determine specific chemical compound. The transfer line was heated and maintained at a temperature of
200ºC to prevent the condensation of the volatile gases released during the torrefaction process. The mass
spectrometer is operated at electron energy of 70 eV. The gases were analysed by using the RGA software. A
scan of the m/z was carried out from 1 to 100 amu to determine which m/z has to be followed during the TG
experiments. The intensities of 10 selected ions (m/z = 15, 16, 27, 28, 29, 32, 43, 44, 94, and 96) were monitored
with the thermogravimetric parameters. An absolute quantification in this work was performed for each monitored
gases: CH4, CO2, CO, C2H4, NH3, O2, C6H6O, CH2O, C2H4O2 and CH5O3P. The concentrations of CH4, CO2, CO,
C2H4, NH3, O2, C6H6O, CH2O, C2H4O2 and CH5O3P were directly deduced from the partial pressure of gases by
using Eq.1.

[Ai] = (Pi / PT) * 1 000 000 (1)

Where, [Ai] is the concentration of the gas A, Pi is the partial pressure for m/z = i representing of gas A and PT is
the total pressure for all species i.

RESULTS AND DISCUSSION

Proximate analysis
Table 1 shows the proximate analysis of the biomass used in this study. Moisture content, volatile matter and fixed
carbon were determined using TGA analysis. However, ash content was investigated as per ASTM Standard
Method E1755-01. Low moisture content in the present biomass suggests them as a good candidate for
torrefaction process. Switchgrass showed highest amount of volatile matter and lowest fixed carbon content
compared to other biomass. Agricultural biomass showed high amount of ash content compared to forestry
biomass. The same comparison in the ash content was reported for agricultural biomass and woody biomass
(Raimie et al., 2012).

Table 1: Proximate analysis data of selected biomass materials (% weight dry).

Spruce Pine Switchgrass Timothy
Moisture (%) 2.92 4.63 2.23 2.82
Volatile matter (%) 73.75 73.05 68.46 79.55
Fixed Carbon (%) 23.33 22.32 25.70 13.57

387
 Ash (%) trace trace 3.61 4.06
TG-DTG analysis
Fig. 1 depicts the thermogravimetric curve for different biomasses. The degradation reactivity for forestry biomass
(spruce and pine) was almost similar to each other. However the degradation behavior of switchgrass differed from
other biomass. The initial decrease in the weight loss of biomass was attributed to removal of moisture content up
to temperature 110 °C. The initial degradation temperature for spruce, pine and timothy was about 210 °C. Whilst
for switchgrass it was about 175 °C. This shows that the decomposition of switchgrass took place earlier
compared to other biomass. From the TGA, the final degradation temperature for spruce, pine, switchgrass and
timothy was about 375, 375, 360 and 380 °C, respectively. Between temperatures 200 to 500 °C higher
decomposition of cellulose and hemicellulose might have occurred, while small amount of lignin degradation is
expected to take place. This observation was reported similarly for various woody biomasses (Chen and Kuo,
2010). No significant weight loss was observed after temperature of 500 °C. The residual or biochar content for
spruce, pine, switchgrass and timothy was about 16, 15, 21 and 11 wt %, respectively.

Fig. 1. Thermo-diagram of biomass in nitrogen atmosphere, the loss of moisture, volatiles and char.

Apparently, switchgrass showed a highest total weight loss at temperature 800 °C. With respect to the torrefaction
temperature of about 300 °C, switchgrass showed highest weight loss of about 30 wt% compared to spruce
(17 wt%), pine (20 wt%) and timothy (20 wt%). This shows that switchgrass can be an attractive raw material for
torrefaction. According to previous studies (Antal, 1983) and (Mansaray and Ghaly, 1998), the thermal
decomposition of hemicellulose, cellulose and lignin occurs at temperatures ranging from 150-350 °C, 275–
350 °C, and 250-500 °C, respectively. Similarly, the decomposition of lignocellulosic materials in four biomasses is
presented in Fig. 2.

Fig. 2. DTG curves of biomass.

388
Fig. 2 shows that agricultural biomass has two peaks as contrast to one peak for the forestry biomass. The first
peak is attributed due to degradation of hemicellulose. However, second peak represent the decomposition of
cellulose. The first and the second peak for switchgrass were found at temperature of about 230 and 330 °C,
respectively. In case of timothy it was found at temperature of about 310 and 355 °C, respectively. The DTG peaks
for forestry biomass (spruce and pine) was at temperature 350 °C. From this observation, it can be concluded that
the hemicellulose content in the softwood biomass is less compared to agricultural biomass. A kinetic study on
willow and agricultural biomass was found that willow has less hemicellulose content compared to reed canary
grass and wheat straw (Bridgeman et al., 2008).

FTIR analysis
Fig. 3 to 6 illustrates the possible chemical functional groups of the torrefied biomass. The FTIR spectra shown in
Fig. 3, Fig. 4, Fig. 5 and Fig. 6 correspond to the gas analysed at different torrefaction temperature of 200, 230,
260 and 290 °C, respectively. Overall, the peaks for different functional groups became more sharp and visible
when the temperature was increased from 200 to 290 °C. This shows that at higher temperature around 300 °C
more degradation of biomass took place. The band around 3600–3400 cm−1 is apparently due to hydroxyl (O-H)
groups, which was found in all torrefied biomass. The C-H stretch band around 2970-2780 cm-1 is the indication of
an organic compound. A sharp band around 2300-2200 cm-1wave number is due to the formation of CO2 during
torrefaction of biomass. It is observed that the two major bands in the region of 1740–1710 cm−1and 1250–1220
cm−1, might be because of C–O and C–O–C functional groups, respectively. Organic phosphates
(P–O) functional groups can be found in range of 1350–1250 cm-1 stretch band, and the presented of aliphatic
phosphates (P-O-C) functional groups is more obvious at around 1050-990 cm-1 than organic phosphates (P-O)
functional groups. A very sharp peak around 730-550 cm-1 might be due to acid chlorides (C-Cl) functional groups
present (John Coates, 2000)].

Torrefaction 200 °C
Same procedure was employed for all torrefaction condition; the change of transmittance intensities indicates a
variation in the gas concentration (Bassiliakis et al, 2001). Fig. 3 shows, pine and swicthgrass are more reactive
than timothy and spruce at torrefaction temperature 200 ºC, particularly for O-H, C-C and C-O-C of functional
groups. C-H and O=C=O functional groups in all biomass, which are obviously methane (CH4), and carbon dioxide
(CO2) and carbon monoxide (CO), significantly transmit as they are reactive between 150 to 300 ºC (Tihay and
Gillard, 2010). However, different biomass gives out CO2 and CO at different concentration as shown in Fig. 3 to
Fig. 6. CO2 from spruce is the lowest in concentration compared to pine, switchgrass, and timothy at torrefaction
temperature 200 ºC. Switchgrass shows significant C-O-P and O-P groups than the others. Organochlorine
compounds (C-Cl functional groups) are normally present in plants and wood. Its reactivity varies enormously, but
most are relatively inert. However, this functional group has shown different concentration from different biomass.
The intensity at torrefaction temperature 200 ºC is at the highest in pine compared to switchgrass and timothy, and
seemingly none from spruce.

Fig. 3. Decomposition of functional groups at torrefaction temperature 200 °C.

Torrefaction 230 °C
As at torrefaction temperature 200 ºC, most functional groups significantly transmitted at torrefaction 230 ºC but at
different level of intensity. Fig. 4 shows, the chemical compounds in switchgrass are the most reactive compared
to the one in pine, spruce and timothy. O-H, C-C and C-O-C of functional groups in particular, an increase in the
intensity were transmitted from torrefaction temperature 230 ºC. The release of CO2 at torrefaction temperature
230 ºC is about the same as at torrefaction temperature 200 ºC for pine, switchgrass and timothy. However, the

389
CO2 release from spruce has increased at this torrefaction temperature. At torrefaction temperature 200 ºC, as well
as torrefaction temperature 230 ºC, the phosphate compounds released significantly only from switchgrass, and
apparently none from the other biomass. Organochlorine compounds (C-Cl functional groups) are released from
all biomass at this torrefaction temperature, but the transmittance still at different intensity. Pine and switchgrass
released almost of this compound at torrefaction temperature 230 ºC.

Fig. 4. Decomposition of functional groups at torrefaction temperature 230 °C.

Torrefaction 260 °C
From Fig. 5, pine and timothy were found has less reactive as compared to spruce and switchgrass. The
transmittance intensity for spruce has shown significant increased at torrefaction temperature 260 ºC from the
torrefaction temperature 200 °C and 230 °C. The intensities of all the functional groups of chemicals for spruce
were just a slight lesser than switchgrass. This can be anticipated that the degradation of spruce, pine and
timnothy were engaged at later temperature than that of switchgrass. This might be almost hemicellulose and
partially cellulose from switchgrass easily released at lower torrefaction temperature (200 °C, 230 °C and
260 °C) due to less lignin content to bind them from devolatilization and depolymerisation (Tumuluru et al., 2010)
(Demibras, 2009) (Mohan et al.,2006). Lignin decompose from temperature 280 ºC and difficult to dehydrate. Thus
it converts to more char and loose the covalent bond to cellulose or hemicelluloses (Tumuluru et al., 2010). The
releases of CO2 from switchgrass and spruce at torrefaction temperature 260 ºC were higher than that of lower
torrefaction temperature (200 and 230 ºC). Switchgrass was also identified giving off the gases (functional groups
O=C=O) the highest among the selected biomasses at torrefaction temperature 260 °C. The phosphate
compounds were gradually degraded at this torrefaction temperature from torrefaction temperature 200 °C and
230 °C for all biomass. Organochlorine compounds (C-Cl functional groups) released from spruce and switchgrass
significantly greater than that from pine and timothy.

Fig. 5. Decomposition of functional groups at torrefaction temperature 260 °C.

390
Torrefaction 290 °C
It is clearly shown from Fig. 6, the transmittance intensities of all the selected functional groups were significantly
high which means almost gases and organic compound were degraded at torrefaction temperature 290 ºC. The
released intensities of O-H, C-H, O=C=O, C-C, C-O-C, and O-P functional groups were found greater at
torrefaction temperature 290 °C than that of at lower torrefaction temperature. The lignocellulosic biomass has
converted to carbonaceous materials due to generous hemicellulose and cellulose was degraded at this
torrefaction temperature. The organochlorine compounds (C-Cl functional groups) and gases (O=C=O functional
groups) released from swicthgrass were still significantly greater than that of spruce, pine and timothy.

Fig. 6. Decomposition of functional groups at torrefaction temperature 290 °C.

Gas quantification
FTIR and mass spectrometric analyses have advantages and disadvantages. Using information provided by the
FTIR spectrum is possible to have an identification of the gases emitted by each biomass. However, quantification
of each emitted gases require specific tool or software to measure. TG-FTIR and TG-MS techniques are therefore
complementary since they can refine the identification of gases. Table 2 shows the concentration of each gas
during torrefaction of the selected biomass as the results from the TG-MS. To compare the evolution of gas
emissions during torrefaction, the gas composition was calculated as a function of temperature. The comparison is
made upon the gas concentration released by the biomass, which the yields of the gases were calculated for the
four range of torrefaction temperature (Table 2). From the table all the four stages of temperature, the degradation
of gases are mainly composed on O2, CO2, and CO, and then are followed by C2H4, NH3, CH2O and CH4. Carbolic
acid, acetic acid and phosphonic acid have the lowest degradation during the torrefaction. Switchgrass has
degraded carbon dioxide and carbon monoxide the highest compared to spruce, pine, and timothy. These results
are in agreement with the FTIR spectrum, which the functional group showed among the highest of transmittance
for switchgrass. The C-H functional group in the FTIR spectrum has depicted about the same transmittance
intensities for all the biomass. The mass spectrum analysis found for these same intensities was attributed due to
the methane (CH4) generated from switchgrass, timothy and pine, and ethylene (C2H4) generated from spruce.
Simple gas molecule O2 does not have infrared spectra, which cannot be scanned by FTIR. MS has this
advantage on O2 that has shown in this study the amount composed in biomass is as high as other main
composition such as H2O (depicted in FTIR spectrum). Volatile products released from biomass was identified
H2O, CO, CO2 as the main composition (Tihay and Gillard, 2010), which was found the same in this study
including O2. Pine has generated the oxygen at the highest concentration as compared to switchgrass, timothy
and spruce. Generally, the concentration of the selected gas degradation emitted during torrefaction is around 20
to 35% of the total concentration for all biomass, but vary in the propositions when compare between biomass and
stages of torrefaction temperatures. For the continuous torrefaction range of temperature, the degradation of
combustible gases (CH4, CO, C2H2, O2) released by switchgrass (1.662 mg/gswitchgrass) is more than timothy (1.562
mg/gtimothy), pine (1.532 mg/gpine) and spruce (1.135 mg/gspruce). The degradation of combustible gases during
torrefaction temperature 200 °C, 230 °C, and 260 °C by pine showed the highest (0.474 mg/gpine, 0.533 mg/gpine,
and 0.590 mg/gpine, respectively) as compared to switchgrass (0.441 mg/gswitchgrass, 0.512 mg/gswitchgrass and 0.584
mg/gswitchgrass, respectively), timothy (0.396 mg/gtimothy, 0.462 mg/gtimothy and 0.525 mg/gtimothy, respectively) and
spruce (0.311 mg/gspruce, 0.361 mg/gspruce, and 0.409 mg/gspruce, respectively). Whilst, for torrefaction temperature
290 °C, the degradation of combustible gases has similar trend as for continuous torrefaction range of temperature
that is switchgrass took place the highest (0.657 mg/gswitchgrass) as compared to pine (0.650 mg/gpine), timothy

391
(0.587 mg/gtimothy) and spruce (0.457 mg/gspruce). For all species, the percentages of combustible gases may
increases significantly after 300 °C, and thus, loose the energetic content of the biomass.

Table 2. Concentration of selected gaseous products evolved from torrefaction of spruce, pine, switchgrass and timothy.

Temp. Methane (mg CH4/ g biomass) Carbon dioxide (mg CO2 / g biomass)
(oC) Spruce Pine SW Timothy Spruce Pine SW Timothy
200 0.015 0.02 0.023 0.024 0.017 0.036 0.02 0.043
Residence 230 0.018 0.023 0.027 0.028 0.019 0.042 0.028 0.05
time 30
mins. 260 0.021 0.026 0.032 0.032 0.025 0.05 0.039 0.058
290 0.024 0.03 0.036 0.036 0.031 0.063 0.061 0.068
*Total of 81
30-800 0.07 0.088 0.110 0.105 0.171 0.284 0.339 0.341
mins.
Temp. Carbon monoxide (x 10-2 g CO / g biomass) Ethylene (mg C2H4 / g biomass)
(oC) Spruce Pine SW Timothy Spruce Pine SW Timothy
200 0.074 0.11 0.143 0.121 0.041 0.02 0.017 0.02
Residence 230 0.087 0.128 0.166 0.141 0.047 0.022 0.02 0.024
time 30
minutes 260 0.098 0.146 0.189 0.16 0.054 0.025 0.023 0.028
290 0.111 0.164 0.214 0.18 0.06 0.028 0.026 0.031
*Total of 81
30-800 0.309 0.454 0.575 0.511 0.167 0.074 0.081 0.092
mins.
Temp. Ammonia (mg NH3 / g biomass) Oxygen (mg O2 / g biomass)
(oC) Spruce Pine SW Timothy Spruce Pine SW Timothy
200 0.035 0.041 0.04 0.042 0.181 0.324 0.258 0.231
Residence 230 0.04 0.046 0.046 0.049 0.209 0.36 0.299 0.269
time 30
minutes 260 0.046 0.051 0.053 0.055 0.236 0.393 0.34 0.305
290 0.052 0.057 0.06 0.063 0.262 0.428 0.381 0.342
*Total of 81
30-800 0.152 0.16 0.18 0.183 0.589 0.916 0.896 0.854
mins.
Phenol (mg C6H6O / g biomass) Formaldehyde (mg CH2O / g biomass)
Temp.
Spruce Pine SW Timothy Spruce Pine SW Timothy
(oC)
200 0.006 0.007 0.007 0.009 0.034 0.026 0.027 0.027
Residence 230 0.007 0.008 0.008 0.011 0.044 0.03 0.032 0.032
time 30
minutes 260 0.008 0.009 0.009 0.013 0.046 0.034 0.037 0.036
290 0.009 0.01 0.01 0.014 0.047 0.039 0.043 0.042
*Total of 81
30-800 0.023 0.028 0.029 0.035 0.155 0.125 0.135 0.135
mins.
Acetic Acid (mg C2H4O2 / g biomass) Phosphonic Acid (mg CH5O3P / g biomass)
Temp.
Spruce Pine SW Timothy Spruce Pine SW Timothy
(oC)
200 0.006 0.008 0.009 0.009 0.005 0.007 0.007 0.009
Residence 230 0.007 0.01 0.01 0.011 0.006 0.008 0.008 0.01
time 30
minutes 260 0.008 0.011 0.012 0.013 0.007 0.009 0.009 0.012
290 0.009 0.013 0.014 0.014 0.008 0.01 0.011 0.013
*Total of 81
30-800 0.037 0.045 0.05 0.049 0.024 0.028 0.031 0.035
mins.
*Note: The continuous torrefaction from room temperature to 800 °C.

392
CONCLUSION
Four different types of Canadian biomasses were torrefied using TGA-FTIR analyser and TGA-Mass spectrometry.
Forestry and agriculture biomass showed single and two DTG peaks respectively during toreffaction. It was
observed that the species had different evolution patterns, which indicates the presence of different chemical
functional groups within the biomass samples. In general, TG-FTIR analysis showed that the torrefaction behavior
of forestry (spruce and pine) biomass were similar. Of all the biomass, switchgrass experienced considerable
weight loss at lower torrefaction temperature. Pine, timothy and spruce was found to have similar degradation
behaviour in terms of chemical functional groups when torrefied at lower temperature (200 and 230 °C).
Degradation of combustible gases: CH4, CO, C2H2 and O2, during torrefaction at lower temperature: 200 °C and
230 °C is around 20–30 %, whilst at temperature 260 °C and 290 °C is around 30-45%. Most of the chemical
compositions were decomposed at higher torrefaction temperature of 290 oC. Overall, the gaseous products
evolved during torrefaction of forestry and agricultural biomass showed similar FTIR spectra patterns and in
agreement with mass spectrometric, MS, with regard to the degradation of combustible gases. Torrefaction
temperature played a key role in formation of these gaseous products.

Acknowledgment
The authors appreciate the financial assistance from New Brunswick Soil and Crop Improvement Association, New
Brunswick Agricultural Council, Agriculture and Agri-Food Canada and Universiti Teknologi Petronas, Malaysia. .

REFERENCES
Acharya, B., I. Sule, and A. Dutta. 2012. A review on advances of torrefaction technologies for biomass
processing. Biomass Conversion and Biorefinery 2: 349-369.

Alen, R., R. Kotilainen and A. Zaman. 2002. Thermochemical behavior of Norway spruce (Piceaabies) at 180-
225oC. Wood Science Technology 36:163-171.

Antal, M. J. 1983. Biomass pyrolysis: a review of the literature. Part I – carbonydrate pyrolysis. Advances in Solar
Energy 11: 61–111.

Bassilakis, R., R. M. Carangelo, and M. A. Wójtowicz. 2001. TGFTIR analysis of biomass pyrolysis. Fuel 80: 1765-
1786.

Bridgeman, T. G., J. M. Jones, I. Shield, and P. T. Williams. 2008. Torrefaction of reed canary grass, wheat straw
and willow to enhance solid fuel qualities and combustion properties. Fuel 87: 844–856.

Chen, W-H. and P-C. Kuo. 2012. Isothermal torrefaction kinetics of hemicellulose, cellulose, lignin and xylan using
thermogravimetric analysis. Energy 36: 6451–6460.

Chen, W-H. and P-C. Kuo. 2010. A study on torrefaction of various biomass aaterials and its impact on
lignocellulosic structure simulated by a thermogravimetry. Energy 35: 2580–2586.

Demirbas, A. 2009. Pyrolysis Mechanisms of Biomass Materials. Energy Sources Part A: Recovery, Utilization,
and Environmental Effects 31(13): 1186–1193.

Hakkou, M., M. Pétrissans, P. Gérardin, and A. Zoulalian. 2006. Investigations of the reasons for fungal durability
of heat-treated Beech wood. Polymer Degradation Stability 91:393-397.

John, C. 2000. Interpretation of Infrared spectra, a practical approach. In Encyclopedia of Analytical Chemistry,
10815–10837, by R.A. Meyers (Ed.) Chichester: Ó John Wiley & Sons Ltd.

Mansaray, K.G. and A.E. Ghaly. 1998. Thermal degradation of rice husks in nitrogen atmosphere. Bioresource
Technology 65:13–20.

Melin, S. 2012. Development of the Canadian bulk pellet market. Wood Pellet Association of Canada REF to FII
Agreement #11/12-050.

Miranda, R., C. Sosa-Blanco, D. Bustos-Martinez, and C. Vasile. 2007. Pyrolysis of textile wastes. I. Kinetics and
yields. Journal of Analytical and Applied Pyrolysis 80: 489–495.

393
Mohan, D., C. U. Pittman, and P.H. Steele. 2006. Pyrolysis of Wood/Biomass for Bio‐oil: A Critical Review. Energy
& Fuels 20 (3): 848–889.

Nguila-Inari, G., M. Pétrissans and P. Gérardin. 2007. Chemical reactivity of heat treated wood. Wood Science
Technology 41:157-168.

Nguila-Inari, G., S. Mounguengui, S. Dumarcay, M. Pétrissans, and P. Gérardin. 2007. Evidence of char formation
during wood heat treatment by mild pyrolysis. Polymer Degradation and Stability 92:997-1002.

Nuopponen, M., T. Vuorinen, S. Jamsa, and P. Viitaniemi. 2004. Thermal modifications in softwood studied by
FT-IR and UV resonance Raman spectroscopies. Journal of Wood Chemical Technology 24(1):13-26.

Oladiran, F. and B. Littlefield. 2012. TG-FTIR analysis of Pecan shells thermal decomposition. Fuel Processing
Technology 102: 61–66.

Prins, M. J., K. J. Ptasinski and F. J. J. G. Janssen. 2006. Torrefaction of Wood Part 1: Weight loss kinetics.
Journal of Analytical and Applied Pyrolysis 77: 28–34

Raimie, H., H. Ibrahima, L. I. Darvella, M. Jenny, M. Jonesa, and A. William. 2012. Physicochemical
characterisation of torrefied biomass. Journal of Analytical and Applied Pyrolysis doi:10.1016/j.jaap.2012.10.004

Shang, L., J. Ahrenfeldt, J. K. Holm, S. Barsberg, R. Zhang, Y. Luo, H. Egsgaard, and U. B. Henriksen. 2013.
Intrinsic kinetics and devolatilization of wheat straw during torrefaction. Journal of Analytical and Applied Pyrolysis
100: 145–152.

Tihay, V. and P. Gillard. 2010. Pyrolysis gases released during the thermal decomposition of three Mediterranean
species. Journal of Analytical and Applied Pyrolysis 88: 168–174.

Tjeerdsma, B. and H. Militz. 2005. Chemical changes in hydrothermal treated wood: FTIR analysis of combined
hydrothermal and dry heat-treated wood. HolzalsRoh- und Werkstoff 63:102–111.

Tumuluru, J. S., S. Sokhansanj, C. T. Wright, and R, D. Boardman. 2010. Biomass Torrefaction Process Review
and Moving Bed Torrefaction System Model Development INL publication 08,
URL:http://www.inl.gov/technicalpublications/Documents/4737111.pdf [cited 18 March, 2013].

Welzbacher, C., C. Brischke, and A. Rapp. 2007. Influence of treatment temperature and duration on selected
biological, mechanical, physical and optical properties of thermally modified wood. Wood Material Science
Engineering 2:66-76.

Wood, S. M. and D. B. Layzell. 2003. A Canadian Biomass Inventory: Feedstocks for a Bio-based Economy.
BIOCAP Canada Foundation Final Report Industry Canada Contract # 5006125.

394
 A SOLAR ENERGY CALCULATION STUDY FOR THE BUILDINGDS IN BAYBURT IN ORDER
TO GET OPTIMUM BENEFIT FROM THE SUN DIRECTLY
1 2
U. Cakir , E. Sahin
1
Bayburt University, Engineering Faculty, Mechanical Engineering Dep. BAYBURT
2
Ordu University, Vocational School of Techical Sciences, ORDU
ucakir@bayburt.edu.tr, esahin1972@gmail.com

ABSTRACT
This study is a research made on evaluating solar energy utilization rates of the buildings and residences which
are under the climatic and meteorological conditions of Bayburt. In the scope of this study, with the aim of enabling
them to benefit from solar energy at the maximum level, a numerical modeling study in which we could determine
the most suitable directions, shapes and sizes of the buildings which would be built in Bayburt was carried out that
would. Although the modeling was applied for sample buildings that have specific sizes and shapes and some
results were obtained, when desired, this modeling can be used for other buildings which have different shapes
and the most suitable direction and location designation can be done. In the solar energy analysis made for the
buildings, a year time was divided into four different periods. Related evaluations were made by taking into
account that the building should be in shape and orientation which takes the maximum and minimum radiance in
the summer and winter season. Basic parameters were determined as azimuth angle, the ratio of the building
length to its width, and total periodical radiance that the building takes.

INTRODUCTUION
Life cycle analyses have demonstrated that the majority of energy in a building is consumed in operational energy
or during the post-occupancy phase of building’s life. A universally accepted method of reducing the energy
demands of active or mechanical means of heating or cooling buildings is through passive solar energy. The
operational energy demands of buildings can be reduced by incorporating passive solar design principles
appropriate to the local climate in preliminary stage [1, 2]. Passive heating or cooling by using solar energy can be
used for all of buildings on the world. Amount of the usable solar radiation that reaches to the earth should be
known firstly and then the shape, orientation and position of building must be determined according to get optimum
benefit from the sun. Design of collectors, determining the lightening loads of buildings, evaluating the heat energy
captured from the sun, designing the solar ovens and calculating the insolation rates of greenhouses must be
determined according to amount of solar radiation or solar energy that reaches to the relevant place.

As it is known solar incidence angle of the sun is different for every day, every hour or every second of the year
because of the motion of the sun and world. For that reason the solar radiation energy that is gained by any place
or building on the world changes continuously depend on the time. For that reason effects of solar radiation of the
on the energy losses and gaining of the buildings must be taken in account in the architectural design of buildings
especially in winter season and summer seasons. In summer season, solar energy gaining of any building is
wanted to be minimal and it is wanted to be maximal in winter season.

There are some studies made on improving the architectural design of buildings to get maximum energy from the
sun or use minimum energy if it uses conventional energy sources for heating or vice versa. The entire building
should be considered as it is just one system. While building is being sized and positioned, the most solar energy
gaining rate of the system should be put forward for cold seasons and least solar energy gaining rate of the
system should be put forward for hot seasons. It is wanted for comfort that let the building gains more energy in
winter while it gains little energy in summer.

In a paper it is aimed to give some results of a comparative analysis of two basic sky models, isotropic: Hottel–
Woertz–Liu–Jordan and anisotropic: the HDKR, Hay–Davies– Klucher–Reindl, to recommend one of these models
for determination of solar energy availability on a building envelope and to formulate the energy balance of a
building. Differences between results obtained from both models increase with the slope of exposed surfaces. The
biggest differences (12–15%) are evident for vertical south surfaces, especially in summer. The simplified isotropic
sky model is not recommended for evaluation of solar radiation availability on the building envelope.
Underestimation of solar gains can lead to the selection of an unsuitable concept and construction of a building
and result in poor indoor thermal comfort, i.e. overheating of rooms in summer [3]. Another paper presents the
results obtained from measurements and experiences gained from interviews on 12 advanced solar low energy
houses designed and built as part of the International Energy Agency (IEA) Solar Heating and Cooling
Programme—Task 13. Three years after the IEA Task 13 formally ended, the results were collected by means of
questionnaires sent to the former participants in Task 13. The paper gives a brief presentation of the houses and
the applied energy saving measures. Measured and expected energy consumptions and indoor climate conditions

395
are compared and differences explained. Special innovative installations and systems are described and
evaluated. In general the measured energy consumption was higher than the expected values due to user
influence and unforeseen technical problems but still an energy saving of 60% compared with typical houses was
achieved. Prevention of overheating requires special attention also at northern latitudes. Interviews with occupants
revealed the need to explain the building’s behaviour thoroughly to its users and elaboration of user manuals [4].
A building-integrated solar energy system is proposed, with the panels installed such that the overall morphology
resembles that of a traditional Chinese building, i.e., roofing (eaves) at each storey, in addition to that on top of the
building in another study. The panels include photovoltaic cells and solar thermal collectors, thus producing
electric power as well as heating. The particular morphology provides a number of advantages, in terms of solar
energy collection and shading, and their matching to temporal and locational variations in energy demand. These
are in addition to the advantages of solar energy generally. Solar heating and photovoltaic power generation were
calculated for a number of locations. These were compared with the space heating and air conditioning demands,
respectively. The requirement for supplementary energy was calculated. Equivalent calculations for similar
buildings without solar panels allowed the saving in non-solar energy to be estimated. Calculations were made for
Beijing in winter, as an example of high space heating demand, for Hong Kong in summer, as an example of high
air conditioning demand, and for Shanghai, as an intermediate example. These showed potential savings of up to
15% in space heating, and up to 55% in air conditioning energy demand [5]. The facade design is and should be
considered a central issue in the design of energy-efficient buildings. That is why dynamic facade components are
increasingly used to adapt to both internal and external impacts, and to cope with a reduction in energy consumption and
an increase in occupant comfort. To gain a complete picture of any facade’s performance and subsequently carry out a
reasonable benchmarking of various facade alternatives, the total energy consumption and indoor environment need to
be considered simultaneously. Nielsen et. Al. quantified the potential of dynamic solar shading facade components by
using integrated simulations that took energy demand, the indoor air quality, the amount of daylight available, and visual
comfort into consideration. Three types of facades were investigated (without solar shading, with fixed solar shading, and
with dynamic solar shading), and they simulated them with various window heights and orientations. Their performance
was evaluated on the basis of the building’s total energy demand, its energy demand for heating, cooling and lighting,
and also its daylight factors. Simulation results comparing the three facade alternatives show potential for significant
energy reduction, but greater differences and conflicting tendencies were revealed when the energy needed for heating,
cooling and artificial lighting were considered separately. Moreover, the use of dynamic solar shading dramatically
improved the amount of daylight available compared to fixed solar shading, which emphasizes the need for dynamic and
integrated simulations early in the design process to facilitate informed design decisions about the facade [6].

The purpose of the study of Peippo et.al. is to define a simple quantitative solar low energy building design
optimization procedure that will give the optimum design variables from a representative set of options. These
include building geometry, thermal insulation, windows, solar thermal collectors and PVs, whereas less readily
quantifiable technologies such as advanced passive solar designs are not discussed. First, a set of simplified
models is selected for the building energy analyses. A numerical optimization scheme is then integrated with these
models that finds the optimum building design trade-off path in terms of energy and cost, in order to solve the
design problems 1–3..To assess the applicability of the approach, a case study is presented for two generic
building types in three different locations in Europe. Finally, the computed designs are compared with a set a
realized low energy houses [7]. Further advance of glazed, healthy building’s energy efficiency and sustainability is
inextricable linked to the building’s envelopes/facades fundamental physics study related to the dynamic control of
sunlight and optimal control of solar heat gains. Relevant mathematical models and algorithms, as well as
infrastructure/ hardware and software integrated performance prediction and validation are studied. Reviewed is
the most recent analytical and experimental research, current state of science and art, as well as some of the on-
going R&D at the edge of new breakthroughs of the healthy buildings daylighting dynamic control’s performance
prediction and validation. It has been shown that, concerning the variability of the solar radiation spectra incident
on the building’s envelope, and also variability of outdoor and indoor air temperature differences, it is necessary
tuning control of glazing’s transmittance dependence on the solar radiation wavelength, with an aim to optimize
daylighting with the reference to people needs (their health and comfort), and energy (thermal and electrical loads
minimization). Finally, presented are elements of an analytical modeling approach, as initial results of study, aimed
to reach a challenging research goal – Tuning control of buildings Glazing’s transmittance dependence [8].

In this study a case study was made to determine the best orientation, type, length to width rate and position of
residential buildings for Bayburt city which is located in the northeastern part of Anatolia in Turkey. It is a city in the
Black Sea Region of Turkey. To achieve this aim most commonly used normal building types for residential
applications such were picked out. A mathematical model for computing the total solar energy that buildings
gained from the sun at every day and every month, for the latitude of Bayburt is developed for any orientation of
building by the help of MATLAB. Than the model was run and the solar energy gaining rates of building types are
calculated according to different floor areas for each orientation. By the way a comparison made among the
building types according to the total solar energy gained for monthly, annually and seasonally.

396
CASE STUDY
Geographic and Climatic Properties of Bayburt
This case study was made for geographical and climatic conditions of Bayburt which is being situated between 39o
52’ to 40o 37’ latitudes and 39o37’ to 40o45’ longitudes. It sits on 40.16 degree south and 40.15 degree north
latitude and has an altitude of 1556 meters [2, 9].

In the region renowned for its climatic extremes, the town of Bayburt experiences comparatively mild seasonality,
with an annual range of 15oC. In summer temperatures reach an average maximum of 27oC, tough figures in
excess of 36-38oC have been recorded in July or August. During the cold winter average daily maxima hover close
to freezing point. Air temperature levels reach an average minimum point of -6.08oC tough once again extreme
variation is seen in data with absolute figures plunging to -29oC in December and January. Generally January is
the coldest month of the year for Bayburt. Spring sees maximum temperatures jump dramatically from March to
April, whilst autumn temperatures drop steadily towards the cool winter months. Diurnal temperature variation
peaks during summer, when clear skies allow the escape of warm air in to the atmosphere at night. The Bayburt
plain itself receives high levels of solar radiation throughout the year and consequently is, on average, warmer
than the surrounding slopes. Briefly heating of surfaces during the day creates up-slope thermal currents, which
warms nearby slopes while at night, cool air drains into valleys from the surrounding mountains [10, 11].

In this study an annual period divided into two sections because of the reasons highlighted in the previous
paragraphs. First one of them are winter season (01 November to 01 May) and second one is summer season (02
May to 31) As understood from here related solar energy gaining calculations were made for two sections of the
year that they are summer and winter sections.

The buildings and some assumptions

In this study every analyzed building is thought as it is just one system. The solar energy gaining of any facade of
building is not important; all amount of the solar radiation that is taken from the sun is our main target. We
determined and selected seven types of building to make solar analysis. It is assumed that each apartment has
100m2 floor area and the height of every apartment is 3.5m. Table 1 shows the general specifications of selected
building types. Due to the Bayburt is near a cold climate region of Turkey, buildings are wanted to get more solar
radiation in winter.

Table 1. Selected building types

Type Number Number of Aparts for Total apart Building Building Basic floor
of Storey each floor (100m2) number Length (m) Width (m) area (m2)
1 2 2 4 20 10 200
2 4 2 8 20 10 200
3 6 2 12 20 10 200
4 8 2 16 20 10 200
5 4 4 16 20 20 400
6 6 4 24 20 20 400
7 8 4 32 20 20 400

In addition to assumption which has been mentioned over some other assumptions were made in order to make
the investigation easy and to make the comparison more realistic. In this study only direct solar radiation was
calculated, ground reflected and diffused radiation is neglected by assuming they will be equal for all shapes.
Additionally effects of the other factors like shading, type of building material are assumed equal for all situations.
These assumptions don’t change the main results of this study because of that a comparison is made in this
paper.

Approach of Model
Selected types of buildings were analyzed in 89 different orientations from east-west to north-south directions by
increasing the azimuth angle of the building one by one from 0o to 89o. Optimum orientation was determined to
obtain the wanted total radiation energy (heat energy) in winter and summer seasons for each type building
according to different sizes. Surface azimuth is the angle on a horizontal plane between the normal to a vertical
surface and the north-south direction line. A new definition was made here as building azimuth angle (BAA) and it
is explained in the Figure 1. Building azimuth angle (a2) is determined according to a reference position as
azimuth angle of one of the long wall is 0o as seen on Figure 1. In reference position one of the long walls of the
building is faced to the south. When green house is rotated from east to the west as seen in the figure BAA is
changes from 0o to 90o. For the last, when long wall of the building is faced to the west BAA becomes 90o.

397
In this study firstly each surface of the building was analyzed separately and then all system was discussed.
Surfaces of the building were encoded by using some abbreviations as seen in Figure 2. Figure 2 is presented as
a sample for encoding the surfaces of building.

South

Long wall faced to west and short

East West L

Top view of the building

wall faced to south at last

North a2=90o
a2=0o
BAA
L

a2 Top view of the building

W
Long walls are faced to south and W
Building
short wall faced to east at first

Fig. 1. Determining the building azimuth angle

SOUTH

The wall faced to south D2

EAST The wall

The roof faced to south C1 D3
faced
to east The wall
faced to
D1 The roof faced to north west
C2

The wall faced to north D4

NORTH
Figure 2 Encoding the surfaces of the building (first position)

Solar calculations
Solar radiation incident outside the earth's atmosphere is called extraterrestrial radiation. The intensity of solar
irradiation directly outside the earth’s atmosphere on a horizontal surface is almost constant at around 1.367
W/m2, so-called as solar constant. As mentioned above because of this study focuses on a comparison
extraterrestrial radiation calculations were used.

Instantly and hourly solar radiation incident on an inclined surface of a building depend upon some parameters,
like the time of the day, the day of the year (n), declination angle (d), solar altitude angle (z), surface azimuth angle

398
 (a) (in northern hemisphere, it is zero for south facing surfaces, 180o for north facing surfaces, -90o for east facing
surfaces and +90o for west facing surfaces) and latitude angle of the place (e) i.e.

In this study, the daily solar radiation coming on a surface in a day between sunrise and sunset is put forward.
Firstly, for a constant BAA the total daily solar radiation coming on any surface of the building was calculated for
each day of the year. This calculation was made for every surfaces of the building. Total daily solar radiation
coming on the building is calculated by collecting the amounts of daily total solar radiation coming on each surface
of it. The related and required values of these parameters have been computed by using 6, 7 and 8 numbered
references [12, 13, 14].

Zenith angle z is given with equation 1, where h is hour angle, d is declination angle and e is latitude angle.

z arccos(cos d .cos e.cosh + sin d .sin e) (1)

Sun rays become parallel with surface when the sun rises and sets. Solar incidence angle (H) of a surface at the
time that sun sets or rises can be calculated by using Eq. 2 and solar incidence duration (Tg) of any day can be
calculated in time unit by using Eq. 3.

sin d .sin e
cos H = − =− tan d .tan e (2)
cos d .cos e
2 2
=
Tg = .H arccos(− tan d .tan e) (3)
15 15
Solar incidence angle (g) of a surface at any time of the day is evaluated as;

cos ( g ) = [ cos d .cos e.cosh.cos s ] + [ cos a.cos d .sin e.cosh.sins ] + [sin a.cos d .sinh.sin s ] +
(4)
[sin d .sin e.cos s ] − [ cos a.sin d .cos e.sin s ]
Here, a is defined as azimuth angle and s is defined as inclination angle of the surface. Terms of H1p and H2p are
defined as the hour angles which solar incidence angle becomes zero (the sun's rays are parallel to the surface).
H1p and H2p are calculated as;
 C −D 
H1 p = 2 arctan  1  (5)
 C2 − C3 
 C +D 
H 2 p = 2 arctan  1  (6)
 C2 − C3 
Values of C1, C2, C3, and D can be calculated by using 7, 8, 9 and 10 numbered equations as seen below.

C1 =
 sin a.cos d .sins (7)

C2 cos d .[ cos e.cos s + cos a.sin e.sins ] (8)

C3 sin d .[sin e.cos s − cos a.cos e.sins)] (9)

D= C12 + C22 − C32 (10)

The times that solar radiation comes parallel to the surface may be before sunrise and after sunset. For that
reason if the absolute value of the hour angle at the time solar incidence comes parallel to the inclined surface
greater than the hour angle at the sunrise time, the hour angle at the time that solar rays first reach to the surface
is at the sunrise time. By using Equation 11 the cosine of solar incidence angle (g) can be calculated when value
of h is zero (solar noon).

cosg=
0  C2 + C3 (11)

399
After determining the g0 value, H1 and H2 values which are the hour angles at sunrise and sunset for any inclined
surface can be evaluated by using the algorithm showed on the Table 1.

Table 2. The algorithm which is used to evaluate the hour angles at sunrise and sunset for any inclined surface
H1 H2
2
cos (go) > 0 D >0 max (H1p, -H) min (H2p, H)
(go> 90o) D2 < 0 -H H

cos(go) < 0 D2 > 0 max (H2p, -H) min (H1p, H)

(go> 90o) D2 < 0
Solar radiation does not comes to the surface

It is a requirement to evaluate that how many hours the solar radiation comes on an inclined surface in a day. That
evaluation must be made for all days of a year. We can determine this by using the equations numbered as 12 and
13.

2
for d > 0 :           teg =
   arccos  − tan d .tan ( e − s )  (12)
15
2
for  d < 0 :           teg = arccos [ − tan e.tan d ] (13)
15
Generally extraterrestrial radiation is made use of in calculation of the solar radiation that comes to the earth.
Extraterrestrial radiation comes on unit area of a horizontal surface instantly can be calculated by using the Eq. 14.
Where Igs is the solar constant and on average it is 1367 W/m2. This value varies by ±3% as the earth orbits the
sun [15], and f is the solar constant correction coefficient and determined by using Eq. 15.
I 0 = I gs . f .cos z (14)
n
f = 1 + 0.033cos(360 ) (15)
365
Amount of solar radiation that comes on a horizontal surface in dt period (dQ0) is calculated by using Equation 16.
This equation can be improved to the Equation 17 by the help of hour angle and t=12/π. Then we get the Equation
18 by putting the Equation one which expresses the zenith angle into the Equation 17 and by integrating it from H1
to H2.

Equation 18 and 19 defines the amount of extraterrestrial solar radiation that comes on unit area of a horizontal
surface in one day from sunrise to sunset.

dQ0 = I gs . f .cos z.dt (16)

12
dQ0 = .I . f .cos z.dh (17)
π gs
H2
12
Qo .I gs . f . ∫ ( cos d .cos e.cosh + sind .sin e ) .dh (18)
π H1

24  π 
Qo * I gs * f * cos ( d ) cos ( e ) sin ( H ) + H sin ( d ) sin(e)  (19)
π  180 
Amount of extraterrestrial solar radiation (Qoe) that comes on the unit are of an inclined surface in one day can be
calculated bay using Equation 20 from sunrise to sunset. s is the slope angle of the surface.
 π
12  ( H 2 − H1 ) .sin d . ( sin e.cos s − cos e.sin s.cos a ) + ( sinH 2 − sinH1 ) .
Qoe = .I gs . f . 180
  (20)
π  
cosd . ( cose.coss + sine.sins.cosa ) − ( cosH 2 − cosH1 ) .cosd .sins.sina 

400
RESULTS
Figure 3 shows the change of solar energy gaining of type 1 building in winter and summer seasons according to
the building azimuth angle. As it is mentioned before type one has two stores and four apartments totally. Length
and width of it is 20m and 10m. If the building is positioned with 18o BAA the solar heat energy that is gained from
the sun becomes maximum for winter season and if it is positioned with BAA between 75-90o it will get minimum
solar radiation in both of the seasons. The optimum building azimuth angle is between 12-24o.

Figure 4 presents the solar energy utilization rates of building which is selected as Type 2. This type of building
has got four floors and 8 apartments totally. The BAA value for Type 2 building to get maximum solar energy in
winter season is 19o. But we should care the solar radiation coming on the building in summer season. We can
say that the optimum range for BAA of type 2 is between 12 to 23o.

Figure 5 presents the solar energy that comes from the sun to the surfaces of building- type 2 in winter and
summer seasons. This type of building has got six floors and 12 apartments totally. The best BAA value for Type 2
building to get maximum solar energy in winter season is 14o. But we should care the solar radiation coming on
the building in summer season. We can say that the optimum range for BAA of type 2 is between 11 to 34o.
6
x 10
1.75

1.7 WINTER SEASON (01 NOVEMBER - 01 MAY)

SUMMER SEASON (02 MAY - 31 OCTOBER)
INCOMING OF BUILDING (MJ / Season)

1.65
TOTAL SOLAR RADIATION

1.6

1.55

1.5

1.45

1.4

1.35
0 10 20 30 40 50 60 70 80 90
BAA (a2)

Fig. 3. Change of total solar energy gaining of Type 1 building according to BAA (2 floors, 4 apartments)
6
x 10
3

2.8 WINTER SEASON (01 NOVEMBER - 01 MAY)

INCOMING OF BUILDING (MJ / SEASON)

SUMMER SEASON (02 MAY - 31 OCTOBER )

TOTAL SOLAR RADIATION

2.6

2.4

2.2

1.8
0 10 20 30 40 50 60 70 80 90
BAA (a2)

Fig. 4. Change of total solar energy gaining of Type 2 building according to BAA (4 floors, 8 apartments)

401
 6
x 10
4.2

4 WINTER SEASON (01 NOVEMBER - 01 MAY)

SUMMER SEASON (02 MAY - 31 OCTOBER)
INCOMING OF BUILDING (MJ / SEASON)

3.8
TOTAL SOLAR RADIATION

3.6

3.4

3.2

2.8

2.6

2.4
0 10 20 30 40 50 60 70 80 90
BAA (a2)

Fig. 5. Change of total solar energy gaining of Type 3 building according to BAA (6 floors, 12 apartments)

Solar energy incoming rates change of type 4 is presented in Figure 5. Type 4 has 8 floors and 16 apartments on
each floor. It has got 16 apartments totally. As seen on the figure when BAA is higher than 78o solar energy
gaining of building becomes minimum in summer and winter seasons. If the building is positioned with 15o BAA,
the solar heat energy that is gained from the sun becomes maximum for winter season.

Figure 7 presents solar energy incoming rates change of type 5 building in two seasons depending on different
building azimuth angle values. Type 5 has four floors and four apartments on each floor. It has got 16 apartments
totally. As seen on the figure when BAA is higher than 70o solar energy gaining of building in summer season
becomes higher than winter. If the building is positioned with 27o BAA, the solar heat energy that is gained from
the sun becomes maximum for winter season and if it is positioned with BAA between 70-90o it will get minimum
solar radiation in winter season. The optimum building azimuth angle is between 13-33o

6
x 10
6

WINTER SEASON (01 NOVEMBER - 01 MAY)

5.5
INCOMING OF BUILDING (MJ / SEASON)

SUMMER SEASON (02 MAY - 31 OCTOBER)

TOTAL SOLAR RADIATION

4.5

3.5

3
0 10 20 30 40 50 60 70 80 90
BAA (a2)

Fig. 6. Change of total solar energy gaining of Type 4 building according to BAA (8 floors, 16 apartments)

402
 6
x 10
5.5

WINTER SEASON (01 NOVEMBER - 01MAY)

5.4
SUMMER SEASON ( 02 MAY - 31 OCTOBER)
INCOMING OF BUILDING (MJ / SEASON)

5.3
TOTAL SOLAR RADIATION

5.2

5.1

4.9

4.8

4.7
0 10 20 30 40 50 60 70 80 90
BAA (a2)

Fig. 7. Change of total solar energy gaining of Type 5 building according to BAA (4 floors, 16 apartments)

Change of solar radiation energy gaining of building-type 6 according to BAA is presented by Figure 8. As it is
understood from the Figure 8 that if the building is positioned with 24o BAA, the solar heat energy that is gained
from the sun becomes maximum for winter season and if it is positioned with BAA between 77-90o it will get
minimum solar radiation in winter season. The optimum building azimuth angle is between 20-40o

Figure 9 shows the change of solar energy gaining of type 1 building in winter and summer seasons according to
the building azimuth angle. As it is mentioned before type 7 has 8 stores and 4 apartments totally. Length and
width of it is 20m and 20m. If the building is positioned with 22o BAA, the solar heat energy that is gained from the
sun becomes maximum for winter season and if it is positioned with BAA between 73-90o it will get minimum solar
radiation in both of the seasons. The optimum building azimuth angle is between 16-32o.
6
x 10
7.2

7 WINTER SEASON (01 NOVEMBER - 01 MAY)

SUMMER SEASON (02 MAY - 31 OCTOBER)
INCOMING OF BUILDING (MJ / SEASON)

6.8
TOTAL SOLAR RADIATION

6.6

6.4

6.2

5.8

5.6
0 10 20 30 40 50 60 70 80 90
BAA (a2)

Fig. 8. Change of total solar energy gaining of Type 6 building according to BAA (6 floors, 24 apartments)

403
 6
x 10
9

WINTER SEASON (01 NOVEMBER - 01 MAY)

8.5 SUMMER SEASON (02 MAY - 31 OCTOBER)
INCOMING OF BUILDING (MJ / SEASON)
TOTAL SOLAR RADIATION

7.5

6.5

6
0 10 20 30 40 50 60 70 80 90
BAA (a2)

Fig. 9. Change of total solar energy gaining of Type 7 building according to BAA (8 floors, 32 apartments)

Table 3 shows us the optimum BAA values for buildings and the amount of heat energy gained by the sample
buildings from the sun when building is positioned with optimum BAA. Additionally Figure 10 presents the change
of amount of solar radiation gaining by the building when it is oriented by optimum BAA for seasonally and
annually. As understood from the table and figure number of apartments per storey is more important than height
of building to get more benefit from the sun due to that increase in total surface becomes higher when the number
of apartments per storey increases. Additionally according to the Figure 10 and Table 3 the solar radiation that is
gained by the building is always higher for winter season and that is a desired result for Bayburt.

Table 3.Maximum total solar energy gaining of buildings per season

Storey 2 4 6 8

Apartments 2 2 4 2 4 2 4
O
a2 ( ) 18 15 27 14 24 14 23
Qo Summer 1453115 2154185 4972490 2861943 5956462 4166146 6956460
Qo Winter 1728308 2930251 5463137 4134130 7082733 5725556 8709859

16
Gained heat energy from the sun (106

Summer
14
Winter
12
Annual
10
MJ/Season)

8
6
4
2
0
4/4 6/2 6/4 2/2 4/2 8/2 8/4
Storey / Apatments
Fig.10. Change maximum total solar energy gaining of buildings per season

404
CONCLUSION
This study was made to evaluate potential of direct solar radiation using of buildings for different types of
residential buildings and different orientations of them. To achieve this aim MATLAB program was used. This
modeling study and the algorithm can be used for any building on the world. Additionally this study can be used for
different solar energy calculations too, for example greenhouses of solar houses. Due to winter season is too cold
and summer season is cool and duration of spring/autumn seasons is very short in Bayburt, we concluded that
amount of solar radiation energy that is gained by buildings in Bayburt is more important. For that reason the
calculations and determination the optimum building azimuth angle is obtained according to winter conditions.

It is understood from the study that, solar energy calculations must be made for every building over the world and
sizes and positioning conditions of the buildings must be evaluated carefully for economic, efficient energy using
issues.

REFERENCES
[1] A briefing for Passive Solar Energy Design, Cement and concrete association of Australia
http://www.concrete.net.au/publications/pdf/briefing09.pdf (14.05.2013)
[2] Şahin E.. 2012. determining the optimum shapes and orientations of the buildingds and greenhouses in
Bayburt in order to get more benefit from the sun directly. Master Thesis, Bayburt University Graduate
School of Natural and Appl Sci Dep of Mechanical Eng. Turkey
[3] Chwieduk D.A., Recommendation on modelling of solar energy incident on a building envelope.
Renewable Energy 34 (2009) 736–741
[4] Thomsen K.E., Schultz J.M., Poel B. Measured performance of 12 demonstration projects—IEA Task 13
“advanced solar low energy buildings”. Energy and Buildings 37 (2005) 111–119
[5] Johnston D. Solar energy systems installed on Chinese-style buildings. Energy and Buildings 39 (2007)
385–392
[6] Nielsen M. V., Svendsen S., Jensen L.B. Quantifying the potential of automated dynamic solar shading in
office buildings through integrated simulations of energy and daylight Solar Energy 85 (2011) 757–768
[7] Peippo K., Lund P.D., Vartiainen E. Multivariate optimization of design trade-offs for solar low energy
buildings Energy and Buildings 29 _1999. 189–205
[8] Kima J.T., Todorovica M.S. Tuning control of buildings glazing’s transmittance dependence on the solar
radiation wavelength to optimize daylighting and building’s energy efficiency. Energy and Buildings
Volume 63, Pages 108–118
[9] Kemal Comaklı, Uğur Cakır, Mehmet Kaya, Kadir Bakirci. 2012. The relation of collector and storage tank
size in solar heating systems. Energy conversion and management, 63 (2012) 112-117
[10] Sagona A., Sagona C., 2004, An Historical geography and a field survey of the Bayburt province,
Archeology at the North east anatolian frontier, Peeters 2004 ISBN 90-429-1390-8
[11] Turkish State Meteorological Service (12.03.2013)
[12] Kılıç A., Öztürk A. 1983. Güneş Enerjisi (In Turkish), Kipaş Dağıtım ve Yayıncılık
[13] Duffie J.A., Beckman W. A., 1991 Solar Engineering of Thermal Processes, 2nd ed., John Wiley nd Sons,
New York, USA,
[14] DuffieJ. A.,BeckmanW. A., 1974 Solar Energy Thermal Processes. John Wiley and Sons.
[15] http://science.nasa.gov/science-news/science-at-nasa/2003/17jan_solcon/

405
 PREDICTION OF DAILY AVERAGE GLOBAL SOLAR RADIATION AND PARABOLIC MONTHLY
IRRADIATION MODEL PARAMETERS FOR TURKEY

1 1 2
Z. Oktay , C. Coskun , M. Ertürk
1
Energy System Engineering Department, Faculty of Engineering, Recep Tayyip Erdogan University, Rize, Turkey
2
Mechanical Engineering Department, Faculty of Engineering, Balikesir University, Balikesir, Turkey
Authors' e-mails: zuhal.oktay@gmail.com, dr.can.coskun@gmail.com, mustafaerturk65@gmail.com

ABSTRACT
The amount of solar radiation falling on the surface of the earth is very important data for engineers and scientists to
model the solar energy systems and to make designing and performance analysis. Solar global radiation values can
be estimated practically by employing parabolic monthly irradiation model. Following the establishing of the model
parameters, average hourly global solar radiation values per month can be obtained for any given place. In this
paper, parabolic monthly irradiation model parameters are calculated for 34 cities in Turkey. Also, daily average global
solar radiation values are determined by employing the measured data of 70 cities in Turkey and given in a parabolic
function. The predictions are compared with the solar data available in literature for Turkey. According to annual
average global solar radiation values, Turkey is divided in four regions, and a new comparison factor is introduced and
called as global radiation comparison rate. This model can be applied to any specific locations, where meteorological
data is measured in the world. Thus the calculation time will be reduced. It is expected that this study will be useful for
the designers and scientists working on solar energy systems.

INTRODUCTION

Long-term solar radiation values on the hourly and monthly basis are required in many applications of solar energy
designs, such as PV/PV-T, solar heating-cooling systems and passive solar design. The amount of the solar radiation
falling on the surface of earth is primarily important data for engineers and scientists involved in the design of solar
energy systems. Different solar energy models exist for predicting the solar radiation amount [1]. Solar-radiation
models range from simple empirical relations to complex models [2-4]. Many studies have been carried out for
different locations of Turkey to predict global diffuse and direct solar radiation [5-10].

In this paper, daily average global solar radiation values are determined for 70 cities in Turkey by employing the
measured data in meteorological stations first. Then the average solar radiation value is found and in accordance with
this value the country is divided into four parts. A new parameter, global radiation comparison ratio, is defined
according to the country average global solar radiation. As the first and second regions have the high solar radiation,
parabolic monthly irradiation model parameters are calculated for 34 cities. Thus, suitable places for solar energy
systems are determined and then daily global solar radiation distribution is presented.

GLOBAL SOLAR RADIATION DATA AND THEIR COMPARISON WITH THE STUDIES IN LITERATURE.

The data of global solar radiation values are taken from Turkish meteorological data stations for 70 cities in Turkey.
Data recording dates are given in Table 1 for each city. Monthly, daily and hourly average global solar radiation values
are calculated by employing actual data. Monthly and daily average solar radiation values are compared with the data
available in literature for Turkey.

Bulut et al.[11] has proposed a new model for the estimation of Turkey’s daily global solar radiation. Main function is
the given equation below,

1.5
 π
I = I 2 + (I1 − I 2 ) Sin  (d + 5) (1)
 365 

where d is the number of the day starting from January 1. For the January 1st, d = 1, and for December 31st d = 365.
by employing equation 1, daily and annually total global solar radiation of each city can be calculated. Then these
values are compared to verify our global solar radiation data. Our values are very similar for 58 cities but slightly
different for 12 cities. The difference reaches as high as 10% for Gumushane, Amasya, Yalova, Istanbul, Zonguldak,
Balikesir, Tekirdag, Bartin, Sinop, Kastamonu, Kocaeli and Edirne.

Sen et. al.[12] has proposed a new model for estimation of monthly based hourly solar radiation and performed it for a
site in the northwestern part of Turkey. Our monthly and hourly based solar radiation data are compared with that
study and similar values have been achieved. Also, it is seen that our average solar data are in compliance with many
studies in literature [13-16].

406
Table 1. Global solar radiation data recorded years for each city.
Data recorded years Cities
1975-2007 (33 years) Samsun, Eskisehir, Rize, Van, Urfa, Hakkari, Aksaray, Diyarbakir, Malatya, Erzurum,
Elazig, Kayseri, Kars, Siirt, Adana, Erzincan, Gaziantep, Adiyaman, Igdir
1975-2006 (32 years) Balikesir, Bursa, Kastamonu, Kocaeli, Isparta, Ankara, Mugla, Cankiri, Aydin, Konya,
Kirsehir, Sivas, Canakkale
1975-2005 (31 years) Trabzon
1975-2003 (29 years) Edirne
1977-2006 (27 years) Antalya
1980-2006 (27 years) Istanbul
1982-2006 (25 years) Izmir
1984-2007 (24 years) Karaman, Mersin, Nigde, Kilis, Ordu, Gümüshane, Tokat, Bitlis, Amasya, Antakya,
Artvin, K.Maras
1984-2006 (23 years) Afyon, Denizli, Tekirdag, Bartin, Sinop, Duzce, Corum, Bilecik, Yalova, Zonguldak,
Sakarya
1985-2007 (23 years) Batman, Mus
1986-2007 (22 years) Bingol, Agri, Yozgat
1985-2006 (22 years) Kutahya, Kirikkale
1975-1996 (22 years) Usak
1987-2007 (21 years) Tunceli
1996-2006 (11 years) Burdur

ESTIMATION OF DAILY AVERAGE GLOBAL SOLAR RADIATION

Daily average global radiation amount is calculated for 70 cities by employing solar data. Then daily average global
radiation values are adopted as fourth order parable by employing Matlab programming. The daily average global
solar radiation on a horizontal surface ( I D ) in kW/m2 can be estimated using the following equation;

ID = k + l ⋅ d + m ⋅ d 2 + n ⋅ d 3 + t ⋅ d 4 (2)
where d is the number of the day starting from January 1. For the January 1st, d = 1, and for December 31st, d = 365.
In Eq. (2), k, l, m, n and t are the model parameters. Annually and daily average global solar radiation amounts are
given in Table 2. In this study, Turkey is classified into four groups in accordance with the annually average global
solar radiation.

First region: Annually average global solar radiation exceeds 0.45 kW/m2. There are nine cities in the first region,
namely Nigde, Van, Antalya, Karaman, Mersin, Burdur, Diyarbakir, Kilis, Tunceli (see Fig. 1 )

Fig. 1. First region in Turkey

Second region: Annually average global solar radiation ranges between 0.40 and 0.45 kW/m2. There are 25 cities in
the second region, namely Izmir, Aydin, K.Maras, Konya, Kirsehir, Afyon, Urfa, Hakkari, Aksaray, Bingol, Malatya,
Erzurum, Siirt, Gumushane, Mus, Elazıg, Kayseri, Kars, Bitlis, Adana, Cankırı, Kirikkale, Kutahya, Tokat, Ankara (see
Fig. 2 )

Third region: Annually average global solar radiation ranges between 0.35 and 0.40 kW/m2. There are 26 cities in the
third region, namely Mugla Yozgat, Gaziantep, Corum, Erzincan, Canakkale, Usak, Agri, Amasya, Bilecik, Isparta,

407
Adiyaman, Igdir, Yalova, Sivas, Istanbul, Antakya, Artvin, Zonguldak, Batman, Sakarya, Balikesir, Samsun, Denizli,
Tekirdag, Bursa (see Fig. 3 )

Fig. 2. Second region in Turkey

Fig. 3. Third region in Turkey

Fourth region: Annually average global solar radiation is lower than 0.35 kW/m2. There are ten cities in the fourth
region, namely Bartin, Eskisehir, Sinop, Duzce, Kastamonu, Ordu, Kocaeli, Trabzon, Rize, Edirne (see Fig. 4 )

Fig. 4. Third region in Turkey

408
Table 2. Daily and annually average global solar radiation

Annually Daily average global solar radiation (kW/m2)

Cities
Longitude Latitude GRCR average global ID = k + l ⋅d + m⋅d 2 + n⋅d 3 + t ⋅d 4
(°E) (°N) (-) solar radiation
(kW/m2) k l m n t
-4 -5 -7
Nigde 34.40 37.59 1.30 0.532 0.231 3.56·10 6.63·10 -3.79·10 5.34·10-10
Van 43.41 38.28 1.22 0.496 0.214 8.60·10-4 5.56·10-5 -3.32·10-7 4.75·10-10
Antalya 30.42 36.53 1.16 0.473 0.201 6.57·10-4 5.64·10-5 -3.32·10-7 4.73·10-10
Karaman 33.14 37.11 1.16 0.473 0.188 7.30·10-4 5.80·10-5 -3.41·10-7 4.84·10-10
Mersin 34.36 36.49 1.16 0.472 0.202 1.30·10-3 4.61·10-5 -2.86·10-7 4.11·10-10
Burdur 30.17 37.43 1.15 0.468 0.173 4.43·10-4 6.20·10-5 -3.55·10-7 4.98·10-10
Diyarbakir 40.12 37.55 1.14 0.466 0.171 1.83·10-4 6.67·10-5 -3.76·10-7 5.26·10-10
Kilis 37.05 36.44 1.13 0.461 0.185 5.52·10-4 5.74·10-5 -3.32·10-7 4.67·10-10
Tunceli 39.32 39.06 1.11 0.452 0.179 -7.88·10- 7.65·10-5 -4.12·10-7 5.71·10-10
4
Izmir 27.10 38.24 1.09 0.447 0.180 -1.31·10- 6.67·10-5 -3.73·10-7 5.24·10-10
4
Aydin 27.50 37.51 1.08 0.442 0.175 -1.02·10- 6.49·10-5 -3.62·10-7 5.07·10-10
4
K.Maras 36.56 37.36 1.08 0.440 0.148 -2.39·10- 7.27·10-5 -4.02·10-7 5.61·10-10
4
Konya 32.30 37.52 1.08 0.439 0.177 6.39·10-4 5.32·10-5 -3.11·10-7 4.40·10-10
Kirsehir 34.10 39.08 1.07 0.438 0.183 5.53·10-4 5.38·10-5 -3.14·10-7 4.45·10-10
Afyon 30.32 38.45 1.07 0.437 0.159 7.39·10-4 5.57·10-5 -3.28·10-7 4.65·10-10
Urfa 38.46 37.08 1.07 0.436 0.166 9.71·10-5 6.25·10-5 -3.52·10-7 4.93·10-10
Hakkari 43.46 37.34 1.07 0.436 0.214 6.07·10-4 4.69·10-5 -2.79·10-7 4.00·10-10
Aksaray 34.03 38.23 1.06 0.435 0.184 7.66·10-4 4.86·10-5 -2.88·10-7 4.09·10-10
Bingol 40.30 38.52 1.06 0.434 0.177 -8.91·10- 7.68·10-5 -4.16·10-7 5.82·10-10
4
Malatya 38.18 38.21 1.06 0.434 0.153 -1.63·10- 6.87·10-5 -3.80·10-7 5.29·10-10
4
Erzurum 41.16 39.55 1.04 0.426 0.192 1.00·10-3 4.25·10-5 -2.61·10-7 3.75·10-10
Gumushane 39.27 40.27 1.04 0.426 0.128 1.18·10-3 5.62·10-5 -3.42·10-7 4.91·10-10
Siirt 41.56 37.56 1.04 0.425 0.169 1.57·10-4 5.93·10-5 -3.37·10-7 4.75·10-10
Mus 41.31 38.44 1.01 0.412 0.149 2.27·10-4 5.89·10-5 -3.34·10-7 4.68·10-10
Elazıg 39.13 38.40 1.00 0.409 0.141 -4.02·10- 6.38·10-5 -3.55·10-7 4.95·10-10
5
Kayseri 35.29 38.43 1.00 0.407 0.159 -1.03·10- 5.96·10-5 -3.31·10-7 4.61·10-10
4
Kars 43.05 40.36 0.99 0.406 0.198 4.25·10-4 4.73·10-5 -2.79·10-7 4.01·10-10
Bitlis 42.06 38.22 0.99 0.404 0.161 -1.82·10- 6.28·10-5 -3.50·10-7 4.91·10-10
4
Adana 35.18 36.59 0.99 0.404 0.175 6.16·10-4 4.67·10-5 -2.76·10-7 3.93·10-10
Cankırı 33.37 40.36 0.99 0.404 0.134 5.50·10-4 5.60·10-5 -3.24·10-7 4.56·10-10
Kirikkale 33.30 39.50 0.98 0.402 0.138 1.36·10-4 6.10·10-5 -3.45·10-7 4.84·10-10
Kutahya 29.58 39.24 0.98 0.401 0.127 2.60·10-4 6.35·10-5 -3.65·10-7 5.18·10-10
Tokat 36.54 40.18 0.98 0.401 0.148 1.08·10-4 5.94·10-5 -3.36·10-7 4.73·10-10
Ankara 32.53 39.57 0.98 0.400 0.124 2.90·10-4 5.99·10-5 -3.39·10-7 4.73·10-10
Mugla 28.21 37.12 0.97 0.398 0.152 6.04·10-4 4.89·10-5 -2.86·10-7 4.04·10-10
Yozgat 34.49 39.50 0.97 0.397 0.145 3.60·10-4 5.51·10-5 -3.17·10-7 4.47·10-10
Gaziantep 37.22 37.05 0.97 0.397 0.133 -1.10·10- 6.04·10-5 -3.36·10-7 4.67·10-10
5
Corum 34.58 40.33 0.97 0.397 0.134 4.10·10-4 5.72·10-5 -3.30·10-7 4.66·10-10
Erzincan 39.30 39.44 0.96 0.395 0.153 6.82·10-4 4.84·10-5 -2.87·10-7 4.09·10-10
Canakkale 26.24 40.08 0.95 0.387 0.122 1.90·10-4 6.11·10-5 -3.49·10-7 4.94·10-10
Usak 29.29 38.40 0.95 0.386 0.164 -2.34·10- 5.56·10-5 -3.06·10-7 4.26·10-10
4
Agri 43.08 39.31 0.94 0.385 0.146 5.52·10-5 5.69·10-5 -3.22·10-7 4.54·10-10
Amasya 35.51 40.39 0.94 0.383 0.124 -2.94·10- 6.84·10-5 -3.82·10-7 5.39·10-10
4
Bilecik 29.58 40.09 0.93 0.379 0.134 -3.28·10- 6.40·10-5 -3.55·10-7 4.99·10-10
4
Isparta 30.33 37.45 0.92 0.375 0.152 8.44·10-4 3.96·10-5 -2.37·10-7 3.35·10-10
Adiyaman 38.17 37.45 0.91 0.373 0.130 6.97·10-4 4.66·10-5 -2.75·10-7 3.89·10-10
Igdir 44.02 39.56 0.91 0.371 0.142 6.23·10-4 4.69·10-5 -2.79·10-7 3.99·10-10
Yalova 29.16 40.39 0.91 0.370 0.098 -8.66·10- 8.04·10-5 -4.39·10-7 6.17·10-10
4
Sivas 37.01 39.49 0.90 0.369 0.106 3.59·10-4 5.41·10-5 -3.06·10-7 4.24·10-10
Istanbul 29.05 40.58 0.89 0.367 0.102 -4.36·10- 7.17·10-5 -3.98·10-7 5.62·10-10
4
Antakya 36.07 36.15 0.89 0.366 0.113 5.27·10-4 5.15·10-5 -2.99·10-7 4.22·10-10
Artvin 41.49 41.10 0.89 0.366 0.136 5.40·10-4 5.09·10-5 -3.03·10-7 4.37·10-10
Zonguldak 31.48 41.27 0.89 0.362 0.119 -1.80·10- 8.78·10-5 -4.62·10-7 6.43·10-10
3

409
Table 2 (continued)

Annually Daily average global solar radiation (kW/m2)

Cities
Longitude Latitude GRCR average global ID = k + l ⋅d + m⋅d 2 + n⋅d 3 + t ⋅d 4
(°E) (°N) (-) solar radiation
(kW/m2) k l m n t
-4 -5 -7
Batman 41.10 37.52 0.88 0.357 0.121 6.29·10 4.65·10 -2.73·10 3.86·10-10
Sakarya 30.24 40.46 0.87 0.356 0.126 -1.19·10-3 7.45·10-5 -3.98·10-7 5.56·10-10
Balikesir 27.52 39.39 0.87 0.354 0.105 -6.99·10-4 6.93·10-5 -3.75·10-7 5.22·10-10
Samsun 36.20 41.17 0.86 0.353 0.095 -2.10·10-4 6.40·10-5 -3.57·10-7 5.02·10-10
Denizli 29.05 37.47 0.86 0.352 0.112 3.05·10-4 5.28·10-5 -3.04·10-7 4.30·10-10
Tekirdag 27.29 40.59 0.86 0.351 0.083 2.41·10-4 6.02·10-5 -3.44·10-7 4.86·10-10
Bursa 29.04 40.11 0.86 0.350 0.117 -3.55·10-4 6.08·10-5 -3.36·10-7 4.71·10-10
Bartin 32.21 41.38 0.86 0.349 0.087 -4.11·10-5 6.28·10-5 -3.53·10-7 4.96·10-10
Eskisehir 30.32 39.46 0.85 0.347 0.090 6.63·10-4 4.85·10-5 -2.81·10-7 3.92·10-10
Sinop 35.10 42.02 0.84 0.344 0.087 -9.13·10-5 6.42·10-5 -3.62·10-7 5.12·10-10
Duzce 0.80 0.326 0.088 -8.77·10-5 5.81·10-5 -3.25·10-7 4.56·10-10
Kastamonu 33.46 41.22 0.80 0.326 0.098 8.34·10-4 4.26·10-5 -2.57·10-7 3.67·10-10
Ordu 37.52 40.59 0.79 0.324 0.093 5.49·10-4 4.86·10-5 -2.88·10-7 4.13·10-10
Kocaeli 29.54 40.46 0.75 0.305 0.096 -2.49·10-4 5.59·10-5 -3.10·10-7 4.35·10-10
Trabzon 39.43 41.00 0.75 0.304 0.108 9.82·10-4 3.47·10-5 -2.20·10-7 3.22·10-10
Rize 40.30 41.02 0.70 0.287 0.094 1.72·10-3 2.42·10-5 -1.77·10-7 2.68·10-10
Edirne 26.34 41.40 0.70 0.286 0.076 9.28·10-6 5.12·10-5 -2.91·10-7 4.12·10-10

GLOBAL RADIATION COMPARISON RATIO:

Country’s annual average global solar radiation is taken as a reference point for that parameter. Global radiation
comparison ratio (GRCR) represents the annual average solar radiation situation of the city. If the GRCR is higher
than 1, city receives more annual solar radiation than country. Each city can be compared easily by employing GRCR
value. GRCR value can be found as follows (using the equations below);

Annually average global solar radiation of the city I

GRCR = = Ave,City (3)
Anually average global solar radiation of county I Ave,Country
Annually average global solar radiation amount of any city (kW/m2) in Turkey is given in Table 1. Annually average
global solar radiation of country can be found employing the equations below

i=n I cty1 ⋅ area1 + I cty 2 ⋅ area2 + ... + I cty n ⋅ arean

I Ave,County = ∑ (4)
i =1 area1 + area2 + .... + arean

where I cty indicates annually average solar radiation of each city. ‘Area’ represents surface area of city. When 70
countries are considered, annually average global solar radiation of Turkey is found as 0.408 kW/m2. Global radiation
of each city in Turkey is given in Table 1.

PARABOLIC MONTHLY IRRADIATION MODEL

Sen and Tan [12] noticed that after an evaluation carried on a monthly basis, hourly solar radiation values fluctuated
around a given trend. They adopted daily general trend as the second order parabola and provided a parabolic
monthly irradiation model (PMIM) to evaluate the monthly based average hourly global solar radiation values of any
place. Model is given in equation below;

I = a ⋅t2 + b ⋅t + c (5)

where I represents either monthly average hourly global or diffuse radiation amounts, and t indicates time in hours
within one day. In Eq. (1), a, b and c are the model parameters. In this paper, daily general trend is adopted as the
third order parabola to get more precise values.

The third order parabola is found to give better results for the model parameters and parabolic function. Our main
parabolic equation is given in the equation below

I = a + b ⋅ t + c ⋅ t2 + d ⋅ t3 (6)

410
In Eq. (2), a, b, c and d indicate the model parameters. Four model parameters and parabolic function are calculated
for each month in a year for 34 cities in first and second region. Model parameters are given in Table 3. Versus of
global solar radiation amount for Antalya is shown in Fig. 5 as an example.

Fig. 5. Versus of global solar radiation for Antalya

CONCLUSIONS

In many applications such as in solar energy technologies, an accurate solar database is required to estimate the real
trend of solar radiation in the region. In this regard, in this paper, long-term hourly, daily, monthly average global solar
radiations are utilized to find the accurate model parameters. In this paper, parabolic monthly irradiation model
parameters are calculated for 34 cities in Turkey. Also, daily average global solar radiation values are determined for
70 cities in Turkey by employing the data measured and given in a parabolic function. According to annually average
global solar radiation values, Turkey is divided into four regions and new comparison factor is introduced and called
as global radiation comparison rate.

NOMENCLATURE
t hour in a day
ID daily average hourly global radiation, kW/m2

REFERENCES

[1] Sen, Z. 2008. Solar energy fundamentals and modeling techniques, Springer-Verlag London Limited
[2] Oturanc, G., Hepbasli, A. and Genc, A. 2003. Statistical Analysis of Solar Radiation Data. Energy Sources
25(11):1089-1097
[3] Ulgen, K. and Hepbasli, A. 2004. Solar Radiation Models. Part 1: A Review. Energy Sources 26(5):507-520
[4] Ulgen, K. 2006. Optimum Tilt Angle for Solar Collectors. Energy Sources, Part A: Recovery, Utilization, and
Environmental Effects 28(13):1171-1180
[5] Şenkal, O., Şahin, M. and Peştemalci, V. 2010. The Estimation of Solar Radiation for Different Time Periods.
Energy Sources, Part A: Recovery, Utilization, and Environmental Effects 32(13):1176-1184
[6] Bezir, N. C., Akkurt, I. and Özek, N. 2010. The Development of a Computer Program for Estimating Solar
Radiation. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects 32(11):995-1003
[7] Yilmaz, E. 2010. Zonal Comparison in Measured and Predicted Global Solar Radiation in Turkey and Chile. Energy
Sources, Part A: Recovery, Utilization, and Environmental Effects 32(15):1454-1459

411
[8] Bezir, N. C., Akkurt, I. and Ozek, N. 2010. Estimation of Horizontal Solar Radiation in Isparta (Turkey). Energy
Sources, Part A: Recovery, Utilization, and Environmental Effects 32(6):512-517
[9] Oğulata R. T. and Oğulata S. N. 2002. Solar Energy Potential in Turkey. Energy Sources 24(12):1055-1064
[10] Ulgen, K. and Hepbasli, A. 2002. Prediction of Solar Radiation Parameters Through Clearness Index for Izmir,
Turkey. Energy Sources 24(8) 773-785
[11] Bulut, H. and Buyukalaca, O. 2007. Simple model for the generation of daily global solar-radiation data in Turkey.
Applied Energy 84:477–491
[12] Sen, Z. and Tan, E. 2001. Simple models of solar radiation data for northwestern part of Turkey. Energy
Conversion and Management 42:587-598
[13] Kaygusuz, K. and Ayhan, T. 1999. Analysis of solar radiation data for Trabzon, Turkey. Energy Convers Manage.
40(5):545–56.
[14] Kaygusuz, K. 1999. The comparison of measured and calculated solar radiations in Trabzon, Turkey. Energy
Source 21(4):347–53
[15] Togrul, I. T., Onat, E. 2002. Global solar-radiation over Turkey: comparison of predicted and measured data.
Renew Energy 25(1):55–67
[16] Ulgen, K. and Hepbaslı, A. 2002. Comparison of solar-radiation correlations for Izmir, Turkey. Int J. Energy Res.
26(5):413–30.

412
 COMPARISON OF ALTERNATING-CURRENT LOSSES IN TWO-LAYER
SUPERCONDUCTING CABLES CONSTRUCTED BY SHELL-TYPE AND SOLID-CORE
CYLINDRICAL WIRES
1,* 2
F. Inanir and A. Cicek
Department of Physics, Faculty of Arts and Sciences, Recep Tayyip Erdoğan University, 53100, Rize/Turkey.
1
2
Department of Physics, Faculty of Arts and Sciences, Mehmet Akif Ersoy University, Campus 15100,
Burdur/Turkey.
*
Corresponding author. E-mail: inanir@ktu.edu.tr

Abstract: Alternating-current losses in two-layer power transmission cables of type-II superconducting

wires with cylindrical geometry are numerically investigated with regard to wire cross section. Losses
in shell-type and solid-core superconducting wires are calculated through the Finite Element Method
for an applied alternating current with 50 Hz frequency. Each cable layer is composed of 20 wires
which have 1.0 mm radii, while the thickness of shell-type wires is 0.1 mm. The two wire layers are
wound over a copper core such that the inner and outer layer radii are 20.0 mm and 25.0 mm,
respectively. Alternating-current losses at small applied current amplitudes in shell-type wires are
three-times smaller than in solid-core wires, where the discrepancy diminishes for high current
amplitudes above 90% of the critical current. Besides, losses in both configurations are considerably
higher in outer-layer wires for current amplitudes less than half the critical current, while they converge
at higher amplitudes. The reason for smaller losses in shell-type wires at low applied current
amplitudes is associated to the fact that current distribution is more homogeneous, whereas magnetic
field lines penetrate into the hollow core of these wires.

1. INTRODUCTION
A great deal of effort is concentrated on reduction of alternating-current (AC) transport losses in
superconducting power transmission systems. In this respect, especially geometries of the high-
temperature superconducting (HTS) wires play an important role [1-3]. Existing round-type HTS cable
technologies rely on winding multiple layers of wires with rectangular cross-sections around a
cylindrical conductor, such as copper (Cu) [4-6]. However, AC current losses in such cables are still
considerably high. Thus, utilization of cylindrical wires, rather than rectangular ones, draws notable
attention in HTS cable technologies [7-9]. In this respect, HTS cables composed of superconducting
coatings over a cylindrical metallic substrate with a radius on the order of a few millimeters are devised
[10, 11]. Alternatively, composite cylindrical cables obtained by winding wires around a flexible core in
a helical manner are also demonstrated [12]. Besides, Odier et al [13] achieved manufacturing shell-
type (annular) cylindrical HTS wires.

The magnetic response of either solid-core or shell-type HTS wires with circular cross-sections is
substantially different than the response of rectangular wires if they are exposed to a transport AC
current or an external magnetic field. A theoretical investigation of the magnetic behavior and AC
transport losses of circular HTS wires under a magnetic field is carried out by Mawatari et al [14].
Besides, Gömöry and Inanir numerically investigated AC losses in coils composed of solid-core and
shell-type HTS wires [15]. Pi Wei et al conducted a numerical study of AC losses in an HTS shell
under an external magnetic field [16]. The above approaches can be adopted to investigation of AC
losses in multi-layer HTS cables with circular wires.

This work focuses on a numerical study of the AC transport losses in two-layer superconducting
cables with either solid-core or shell-type circular HTS wires under an applied alternating current. A
comparison of induced losses in the two configurations is carried out. Furthermore, the loss
mechanisms are discussed with regard to distribution of current density and magnetic field in an
around the wires, respectively.

2. CABLE GEOMETRIES AND COMPUTATIONAL METHODS

Cross-sections of the HTS power transmission cables on the xy plane are depicted in Figure 1. Wires
are composed of second-generation high-temperature superconductors. Inner sections of the shell-
type wires are vacuum with relative permeability r=1.0. Inner core of both cables in Figure 1 are
®
circular Cu wires with a radius of Ri=20.0 mm. The space between the layers is filled with Kapton
dielectric with a thickness of 5.0 mm, such that the outer cable radius is Ro=25.0 mm. The total

413
number of wires in each layer is N=20 so that the system is symmetric with respect to discrete
rotations by an angle of e=18º (2 /N), as depicted in Figure 1(a). The radius of all wires is rsc=1.0
mm. The thickness of shell-type wires is t=0.1 mm, as depicted in the inset in Figure 1. Twisting of
wires around the core is ignored so that the system is assumed to extend indefinitely along the z
direction and the computational problem is reduced to two dimensions (2D).

Figure 1. (Color online) Design of two-layer HTS cables with solid-core (a) and shell-type (b) cylindrical wires. The dash-
dotted lines represent computational domain boundaries, where discrete rotational symmetry is exploited. The inset in the
middle is a close-up view of a shell-type wire.

Numerical calculations are carried out through the use of the AC/DC module of COMSOL
®
MultiPhysics package via the Finite Element Method (FEM). Computations are based on solution of
Ampere’s Law in differential form:

1 
∇× ∇× A = j (1)
µ 
where the magnetic flux density (B) is written as the rotational of the magnetic vector potential (A) as
B = ∇ × A . Thus, Equation (1) is solved via FEM for A. In equation (1),  is permeability and j is the
applied current density. In 2D, assuming permeabilities of all elements in the problem are equal to 0,
Equation (1) can be written as

1  ∂ 2 Az ( x, y ) ∂ 2 Az ( x, y ) 
−  + =j z ( x, y ) (2)
µ0  ∂x 2 ∂y 2 

where jz(x,y) corresponds to the current density distribution over the wires applied along the z
direction.

According to the Critical State Model, the magnitude of the current density, j, in type-II
superconductors is either zero or equal to the critical current density Jc. The current density in domains
where the electric field (E) vanishes within the superconductor is zero. In addition, the direction of the
current density is determined by the local electric field within the superconductor with non-vanishing
amplitude.

Non-zero components of both E and j are those along the z axis in Figure 1. The reduced
dimensionality of the computational problem requires assumption of a non-linear relation between E
and j in the form of [17]

 E ( x, y ) 
jz ( x, y=
) J c ⋅ tanh  z  (3)
 Ec 
-4
where Ec is a scaling parameter for the electric field, whose value is taken as 10 V/m. In obtaining
Ez(x,y) at small frequencies through the sum of the time-derivative of A and the gradient of the scalar
potential () as E = −∂A / ∂t − ∇ϕ , a point collocation method (Brandt-type formulation) is utilized [18,

414
19]. In this case Equation (3) becomes

 ∂Az ( x, y ) 
 − ∇ϕ ( x , y ) 
jz ( x, y=
) J c ⋅ tanh  ∂t
 (4)
 Ec 
 

where the only non-vanishing component of A is Az. By obtaining jz(x,y) and Ez(x,y) simultaneously,
the AC losses can be calculated as will be explained below.

Due to the discrete rotational symmetry stated above, the computational domain in FEM simulations is
a pie of a circle, centered on the HTS cable’s center, with a radius of RCD=100.0 mm and a base angle
of e=18º. Furthermore, the influence of the core and the inter-spacing dielectric is neglected for
simplicity. Thus, the computational domain comprises only two wires in each case.

The adopted boundary condition is such that the normal component of the magnetic field across the
boundary of the domain along the azimuthal direction is continuous. Besides, a constraint requiring
that each wire in the HTS cable carries the same current I(t) is adopted. This can be accomplished by
adjusting ∇ϕ , for which an iterative technique was devised. Constant current in each wire requires

∫
wire
jz dAwire = I , while ∇ϕ is constant over the cross-sectional area of a wire.

The FEM elements in simulations are triangular with a total number of 75252, for which the minimum
element quality is 0.6. The critical current for each HTS wire is set as Ic=100 A. The applied current is
sinusoidal in the form of I(t)=Imax.cos(2 ft), where Imax is current amplitude and the frequency is f=50
Hz.Moreover, magnetic field, temperature and position dependence of current is ignored. Imax is set to
9 different vales ranging between 5 A (0.05Ic) and 90 A (0.9Ic).

3. RESULTS AND DISCUSSION

Variation of total AC transport loss, Qe, which is the sum of the individual losses occurring in the inner
(Qi) and outer (Qo) wires with the applied current amplitude is presented in Figure 2. When Imax<Ic/2,
both loss curves vary almost linearly in the log-log scale of Figure 1. The loss for the shell-type wires is
-6
significantly smaller than that of the solid-core wires in this current range. In fact, Qe equals 1.8×10
-6 -4 -3
J/m and 5.8×10 J/m at Imax=5.0 A (0.05Ic), respectively, while it becomes 8.5×10 J/m and 1.6×10
J/m at Imax=50.0 A (0.5Ic). Thus, Qe in the solid-core wires is approximately three-times and twice as
high as that in the shell-type wires at 0.05Ic and 0.5Ic, respectively. However, the increase rate of Qe at
higher current amplitudes is higher for the shell-type wires, where Qe for the solid-core and shell-type
-2 -2
wires becomes 1.0×10 J/m and 1.2×10 J/m at Imax=90.0 A (0.9Ic). That, is Qe of shell-type wires is
16.7% higher at this current amplitude.

Figure 2. (Color online) Variation of total AC transmission losses in cylindrical cables with respect to the applied current
amplitude.

AC losses in both cable types are not equally distributed between the inner and outer layers, as

415
depicted in Figure 3. Losses are considerably higher in the outer cable layer for Imax< Ic/2, while they
are higher in the inner layer for large amplitudes. At Imax=5.0 A (0.05Ic), Qi and Qo for the solid-core
-6 -6 -7
wires are calculated as 1.2×10 J/m and 1.7×10 J/m, respectively, whereas they are 3.7×10 J/m
-7
and 5.5×10 J/m for the shell-type wires. Discrepancy between losses in different layers at that
amplitude is 46.6% and 50.2% for the solid-core and shell-type wires, respectively. At Imax=5.0 A
(0.05Ic), however, Qi is 2.8% and 17.5% higher than Qo for the solid-core and shell-type wires,
respectively.

Figure 3. (Color online) Comparison of individual losses in the inner (Qi) and outer (Qo) layers of solid-core and shell-type
wires.

To understand the variation of AC loss curves in Figure 2 and Figure 3, distribution of equi-potential
lines of A and current density around and in the wires, respectively, for a number of instantaneous
values of the applied current with Imax=90.0 A. Figure 4(a) demonstrates that current flow is almost
homogeneous, except for a small circular cross-section over which current flows in the opposite
direction, when the value of the instantaneous current is close to ±Imax (i.e. at the beginning or the end
of a full-cycle of the applied sinusoidal signal). The opposite-current-flow region expands when the
instantaneous current is small (or the half of the period of the applied sinusoidal current is reached), as
the integral of current density over a wire’s cross-sectional area approaches zero. Moreover, current
flow in this region is not homogeneous and it does not appear at the center of neither the inner nor the
outer wires. Figure 4(a) also indicates that the magnetic field lines are concentrated mainly around the
outer wire at any moment of the applied current, whereas penetration into this wire is more significant.
Besides, the outer wire is subject to the normal component of the magnetic field to a larger extent.
Thus, the fact that the outer wire suffers more AC losses at moderate or small current amplitudes can
be explained in terms of the responses of the HTS wires to their self-fields.

The situation is significantly different for the shell-type wires, in which the current distribution is more
homogeneous when the instantaneous current is close to Imax, Figure 4(b). Towards the half of the
cycle, however, almost equal amounts of current flows in opposite directions over the halves of the
shells. Although magnetic field pattern of the shell-type wires is similar to that of the solid-core wires,
normal components of the field is primarily trapped in the interiors of the shell (vacuum region),
especially on the outer wire, at any instant of the applied current, Figure 4(b).

416
Figure 4. (Color online) Distribution of the equi-potential contours of the vector potential (lines) and the current distribution
(color shades) in solid-core (a) and shell-type (b) wires at several instantaneous current values over a cycle of an applied
current with Imax=90.0 A.

4. CONCLUSION

Alternating-current transport losses in two-layer high-temperature superconducting cables with circular

wires is numerically investigated with respect to wire cross-section. Smaller AC losses can be
obtained if the wires are in the form of annuli, rather than filled circles. Calculated losses for the former
are several times smaller than those in the latter if the amplitude of the applied current with 50 Hz
frequency is smaller than half the critical current in each wire. AC losses are generally higher in the
outer-layer wires for such current amplitudes, while losses in the inner wires become higher at
amplitudes close to the critical current. Current distribution over a cycle is more homogeneous in both
wires of shell-type conductors. In addition, the self-field of the wires is concentrated more around the
outer wire in both configurations, while penetration into the conductor is more pronounced in the outer
layer of solid-core wires. Moreover, the boundaries of these wires are subject to the normal
component of the field to a larger extent, whereas normal component is trapped in the interior space of
the shell-type conductors.

ACKNOWLEDGMENTS
This study is supported by The Scientific and Technological Research Council of Turkey (TÜBİTAK)
under the grant number 110T876.

REFERENCES
[1] M. Majoros, M. D. Sumption, M. A. Susner, C. Kovacs, E. W. Collings, X. Peng, D. Doll, M. Tomsic
and D. Lyons, 2012 A model superconducting helical undulator fabricated using a small filament, tube-
type, multifilamentary Nb3Sn wire, 25, 115006.

417
[2] J. Jiang, W. L. Starch, M. Hannion, F. Kametani, U. P. Trociewitz, E. E. Hellstrom and D. C.
Larbalestier 2011 Doubled critical current density in Bi-2212 round wires by reduction of the residual
bubble density 24, 0822001.

[3] N. Dechoux, C. Jiménez, P. Chaudouët, L. Rapenne, E. Sarigiannidou, F. Robaut, S. Petit, S.

Garaudée, L. Porcar, J. L. Soubeyroux, P. Odier, C. E. Bruzek and M. Decroux, 2012 Textured YBCO
films grown on wires: application to superconducting cables, 25 125008.

[4] L. Rostila, J. Lehtonen, M. Masti, N. Lallouet, J-M. Saugrain, A. Allais, K. Schippl, F. Schmidt, G.
Balog, G. Marot, A. Ravex, A. Usoskin, F. Gömöry, B. Klinčok, J. Šouc and H. C. Freyhardt 2006
Design of a 30 m long 1 kA 10 kV YBCO cable Supercond. Sci. Technol. 19, 418.

[5] O. Maruyama, T. Ohkuma, T. Masuda, Y. Ashibe, S. Mukoyama, M. Yagi, T. Saitoh, T. Hasegawa,

N. Amemiya, A. Ishiyama, N. Hayakawa, 2012 Development of 66 kV and 275 kV Class REBCO HTS
Power Cables, 23 (3), 5401405.

[6] SK. Kim, SK.. Ha, JG. Kim, S. Kim, M. Park, IK. Yu, S. Lee, K. Sim, K 2013 Design and AC Loss
Analysis of a 22.9 kV/50 MVA Class Triaxial HTS Power Cable, J. Supercond. And Novel Magn. 26(4)
755-758.

[7] A. Allais, D. Isfort, C. F. Theune, K. Porscher 2008 Patent Number: US2008/0119365.A

[8] A. Allais, D. Isfort, C. F. Theune, K. Porscher, Patent Number: EP1916720

[9] C. E. Bruzek, A. Allais, S. Morice, C. F. Theune, S. Petit, M. Mikolajczyk, N. Dechoux, C. Jimenez,

E. Sarigiannidou, L. Porcar, J.L. Soubeyroux, P. Odier, T. Waeckerle, 2012 New HTS 2G Round
Wires, IEEE Transc. Appl. Supercond. 22(3) 5800204

[10] B. Ma and U. Balachandran, 2006 Prospects for the fabrication of low aspect ratio coated
conductors by inclined substrate deposition Supercond. Sci. Technol., 19 (6), 497.

[11] V. Ferrando, P. Orgiani, A. V. Pogrebnyakov, J. Chen, Qi Li, J. M. Redwing, X. X. Xi, J. E.

Giencke, Chang-Beom Eom, Qing-Rong Feng, J. B. Betts, and C. H. Mielke, 2005 High upper critical
field and irreversibility field in MgB2 coated-conductor fibers 87, 252509.

[12] J. Šouc, M. Vojenčiak and F. Gömöry, Experimentally determined transport and magnetization ac
losses of small cable models constructed from YBCO coated conductors 2010 Supercond. Sci.
Technol. 23 045029.

[13] P. Odier, A. Allais, C. Millon, S. Morlens, L. Ortega, C. Jiménez, L. Porcar, X. Chaud, P.

Chaudouët, S. Pairis, P. Tixador and J. L. Soubeyroux 2009 New YBCO superconducting wires
obtained from narrow textured tubes 22 125024.

[14] Y. Mawatari 2011 Superconducting tabular wires in transverse magnetic fields Phys Rev. B 83,
(2011) 134512.

[15] F. Gömöry and F. Inanir IEEE Applied. Supercond. 2012 AC losses in coil wound from round wire
coated by a superconducting layer 22(3) 4701704.

[16] Pi Wei, Wang Yin-Shun, Dong Jin and Chen Lei 2010 AC Alternating-Current Loss Analyses of a
Thin High-Temperature Superconducting Tube Carrying AC Transport Current in AC External
Magnetic Field Chin. Phys. Lett. 27, 37401.

[17] F. Gömöry, M. Vojenciak, E. Pardo, J. Šouc 2009 Magnetic flux penetration and AC loss in a
composite superconducting wire with ferromagnetic parts. Supercond. Sci. Technol. 22, 034017.
[18] E. H. Brandt 1996 Superconductors of finite thickness in a perpendicular magnetic field: Strips
and slabs Phys. Rev. B 54 4246-4264.

418
[19] F. Sirois and F. Roy 2007 Computation of 2-D Current Distribution in Superconductors of Arbitrary
Shapes Using a New Semi-Analytical Method IEEE Trans. Appl. Supercond. 17 3836.

419
 MULTI MODAL STRUCTURE FOR THE MANAGEMENT OF ENERGIES
IN A RESIDENTIAL HOME

1 1 2 3
H. Bouguessa , N. Saadia , N. Touileb and A. Makhlouf
1’2’3
Laboratoire de Robotique Parallélisme et Electroénergétique, Faculté d’Electronique & Informatique, U.S.T.H.B., BP
32 El Alia Bab Ezzouar, 16111 Alger, Algérie.
hadia.bouguessa@gmail.com , saadia_nadia@hotmail.com

ABSTRACT

Buildings consume a significant fraction of the energy produced in the world (39% in Algeria), so they are
responsible for important emissions of greenhouse gases; since climate change has become an international
concern, we attempt to adopt energy policies to mitigate global warming and its side effects. This is the reason
for which the migration to positive energy buildings is more and more obvious for better environment protection.
The aim of this paper is a conception of a home automation platform controlling energy. It consists of designing
an adaptive software architecture using multiple modalities for managing two important requirements: the switch
between the energy sources (wind, solar and the network) and the control of different households in a smart
building to improve energy efficiency. This multimodal software architecture is modeled by colored, timed and
stochastic Petri nets (CTSRP) simulated in CPN Tools.

Keywords: Multi-agent Multimodal Platform: Energy Efficiency: Home Automation: Renewable Energies

1. INTRODUCTION

Global demand for energy keeps increasing with economic development and population growth, raising
concerns about the depletion of energy resources and heavy environmental impacts (depletion of the ozone
layer, global warming, etc.) during the period (1984 - 2004), the emission of CO2 increased by 43%, with an
average annual rate of 1.8% as shown in figure 1 (Lombard et Or., 2008). Given that the building sector
(residential and commercial) is the largest consumer of energy produced in Algeria and around the world with
the rate of 39% according to (APRUE., 2010), exceeding the consumption of large sectors (industrial and
transportation); it represents a potential research axis to respond these energy and environmental challenges as
it provides important opportunities to reduce energy consumption and CO2 emissions by maximizing the thermal
performance of the building envelope, minimizing energy demand whilst maintaining energy services levels (De
Silva et ch., 2012) and implementing renewable energy sources.

In residential sector, several computer technologies and methodologies has been used in building energy
management systems (BEMS), such as genetic algorithms, neural networks and weighted linguistic fuzzy rules
(Sierra et ho., 2008), to reach a better energy efficiency and make ordinary building smarter.

In this paper, we focused on multimodal multi-agent systems (SMA) that have evolved significantly over the
past twenty years; they constitute a preferred approach to deal with complex systems in many fields such as
image processing, robotics, eldercare, home automation …, due to their decentralized processing component
entities and their interactions with the environment.

The objective is to develop a Home Automation System Controlling Energy (HASCE) based on multi-agent
approach for managing several modalities such as temperature, light …, in a smart building. To evaluate and
validate the proposed architecture, we modeled the HASCE using CPN Tools. The paper is organized as follows:
Section 2 presents the evolution of smart buildings. Section 3 presents an overview of multi-agent systems in
smart buildings. Section 4 describes the overall multimodal system architecture. Case studies, results and
discussion are presented in Section 5.

420
 Fig. 1. Primary energy consumption, CO2 emissions and world population.
Reference year 1984. Source: International Energy Agency (IEA).

2. INTELLIGENT BUILDINGS

The word intelligent appeared, in the beginning of the 70’s in the United States under the name Building
Energy Management System (BEMS). That’s a center of aggregation and data processing (dump outstation)
for buildings, with the development of Information and Communication Technologies (MISSAOUI., 2012).
This concept is reinforced in 80’s by the increasingly sophisticated demand for “comfort living environment
and requirement for increased occupant control of their local environments” (Wong et li., 2005).
Since then, the intelligent building (IB) concept converged on several meanings, according to the research
conducted by (Wigginton et ha., 2002), there is over 30 different definitions related to intelligent building with
the development of other technologies and communication protocols, the introduction of renewable energy
sources and sustainable development; other terms have emerged such as low- energy building, passive
building, net zero energy building, positive energy building etc.
In (*), we find a definition of IB that we adopt in this article: "The building is more efficient, with integrated
management equipment: equipment manufacturers and equipment storage of electricity, energy efficiency
also depends on the technical construction of the building, insulation, for example, intelligent building concept
is the integration of energy management solutions in the home, particularly to ensure in buildings with positive
energy”. From this point of view, IB should have these main characteristics:
• Exploiting and managing Renewable Energy on the site;
• Reduce Energy Consumption and Costs;
• Improve the safety in buildings;
• Improve the comfort and building.

3. OVERVIEW OF MULTI-AGENT SYSTEMS

Researches around the smart building, home automation and energy management reporting use of SMA
appear for many years in the literature (Clarke et sh., 1999) (Dujardin et ro., 2011). The SMA approach meets
the requirements of distributed systems, it allows to understand, model and simulate complex systems, which
consist of several entities in dynamic interaction with each other and the outside world, to collectively solve a
global problem.

In this paper, we are interested in the SMA developed for energy management in buildings, Davidsson
and Boman developed a platform for energy management for tertiary building with the establishment of three
agents controlling lighting, heating and ventilation (Abras et pl., 2008). In the approach proposed by (Dujardin
et ro., 2011), four active and three passive agents take charge of control and energy management in a room
with 4 offices (Fig. 2), the active agents are defined as follows: Avatar agent representing humans, the Lamp
agent for the control of office lamps, the Room agent to control temperature and finally the Energy agent to
turn On/Off the halogen lamp if desk lamps consume too much energy, the passive agents are charged with
the control of radiator, air conditioner and the halogen lamp. The SMA Interactions on the environment are
the ignition and extinction of electrical appliances as well as increase and decrease the level of heating
appliances. The system is articulated around a software bus named WSE (Web Server Event), all messages
exchanged by agents of the system, such as interactions users or actions required of devices, transit through
the bus (Fig. 3).

421
 Fig. 2. The real environment of the proposed SMA. Fig. 3. Communication protocol between the SMA, peripherals and
control interfaces through WSE bus.

The proposed layered control system by (Wang et ya., 2010) is made up of multiple agents, which are
classified into two levels. The first level interacts with the power grid or micro source termed as central
coordinator-agent. The second level termed local controller-agents in charge of users' comfort in the smart
and energy-efficient buildings mainly decided by three factors, which are environment temperature, indoor air
quality (CO2 concentration), and illumination. A similar approach is proposed by (Wang et wa., 2011), it
consists of a hierarchical multi-agent control system with an intelligent optimizer (PSO) to optimize the overall
system and enhance the intelligence of the integrated building and micro grid system. (Duan et li., 2008) use
a pure distributed architecture to achieve a number of important environmental control parameters inside the
IB, temperature control and lighting experiments are performed on a Windows CE embedded platform. Abras
(Abras., 2008) proposes a system composed of software agents performing each a specific task, interacting
and communicating to find dynamically a production and energy consumption policy, satisfying various
constraints. In this work, agents are associated with equipment and sources of energy production. The Multi-
Agent System (SMA) is called MAHAS "Multi-Agent Home Automation System." The following diagram (Fig. 3)
shows the overall structure of the MAHAS.

Fig. 4. The MAHAS structure for energy management.

422
 4. HOME AUTOMATION SYSTEM CONTROLLING ENERGY

4.1. System architecture

The objective of this work is to design a Home Automation System Controlling Energy (HASCE) based on
multi-agent approach for managing multiple modalities such as temperature, light ... in a smart building.
Indeed, the platform is constituted of an IB supplied with various sources of energy (solar, wind, and the
network), light sensors placed inside and outdoors of the building, presence and temperature sensors and a
switch for selecting the mode of operation (automatic or manual). The proposed interface will allow according
to the input modalities to toggle between renewable energy sources and the network switch on or off the light,
start or stop the air conditioner and heater and raise or lower the roller shutters.
The figure 5 illustrates the general aspect of our multimodal multi-agent architecture. Modalities ensure
the communication between the environment and the IB, while SMA's role is to merge data from the sensors
of the input modalities to control the actuators. This merger will ensure occupant comfort and better control of
the energy produced and consumed.
TEMPERATURE SENSOR

INDOOR LIGHT SENSOR

SMA OUTDOOR LIGHT SENSOR

INPUT SWITCH OPERATING MODE

MODALITIES
DATA PRESENCE SENSOR

FUSION Renewable Energy Produced

Network
Fig. 5. The multimodal software architecture.

4.2. Modeling architecture by Timed Petri Nets:

Introduced by Carl Adam Petri in his thesis, timed petri nets (TPN) are very suitable models for
representing parallel processes. They are at once a mathematical and a graphical language for modeling
complex systems (Boussier., 2007). The most appropriate formalism tool for modeling multimodal
network is timed colored Petri nets (TCPN) (Touileb., 2011) because it can:
a. Let assume the temporal parameters;
b. Take into account the dynamic aspects of a dialogue;
c. Represent the data exchanged in a dialogue as well as media and data structures;
d. Have a distributed representation, graphical and hierarchical.
e. Allow agents to communicate easily with the same semantics.
The combination of SMA and TCPN will allow us to model a multimodal system through the interaction
between agents. Figure 6 illustrates the architecture of the proposed system, it consists of four agents:
• Switch energy sources agent;
• Light comfort agent;
• Temperature comfort agent;
• Inhabitants agent.
Input Modalities

Switching energy agent

Inhabitants agent

Light comfort agent Thermal comfort agent

Fig. 6. Architecture of the proposed SMA.

423
 In order to validate the proposed multi-agent based network petri multimodal platform, we build a
model using CPN Tools software, each agent is modeled by a TCPN.

4.3. CPN Tools software

CPN Tools is a graph editor and simulator timed colored Petri net stochastic. Developed at the University
of Aarhus (Denmark), it is based on direct manipulation of menus and dialog boxes. It can produce networks
with more than a thousand structured places, transitions and arrows into a hundred modules or more. Also,
CPN Tools tool allows multiple subnets in the same window and have a link between them, with the possibility
of spatially distributed places. Therefore, this feature can be used to produce structures of multi-agent
systems by representing each agent separately and make a connection between them. The CPN Tools
license can be obtained free of charge, also for commercial use (Touileb et sa., 2012).

A. Switching energy agent

The switching energy agent (Fig. 7) represents the binding between the network and the IB.
Considering consumer demand and the amount of energy supplied by renewable sources; if
production exceeds the estimated consumer demand, switching agent disconnects IB from the
network to power it by renewable energies. Otherwise, if production is less than consumption
estimated demand or there are disruptions in the production of renewable energies, the switching
agent request relief to the network.

Fig. 7. Switching energy agent.

B. Inhabitants agent
Inhabitants agent has two functions: people detection in the IB and the choice of the operating
mode of the platform. Indeed, this function gives them the hand to switch to automatic mode if they
want that HASCE pilots the IB, or they put the switch in the Manual position if they prefer control the
actuators themselves. These two functions of the inhabitants agent operate in all other agents except
the switching energy agent. The inhabitants agent modeling is shown in figure 8.

C. Light comfort agent

In the case where the presence detector sends information that there's a person in the IB, taking
into account the choice of the operating mode transmitted by agent population, Light comfort agent
modeled (Fig. 9) ensures a graduated ignition and raising or lowering shutters depending on the time

424
and intensity of the light detected by the light sensors placed respectively inside and outside the
building.

Fig. 8. Inhabitants agent.

Fig. 9. Light comfort agent.

D. Thermal comfort agent

The actuation of air-conditioner and heater is ensured by the thermal comfort agent (Fig. 10).
The agent acts according to the temperature value sent by the sensor. Considering that the ambient

425
 temperature is between 18 and 22 degrees, the agent regulates temperature to 19 degrees by
actuating either heater or air-conditioner if temperature is below or above ambient temperature. If the
temperature is within the ambient temperature interval, the agent does nothing.

Fig. 10. Thermal comfort agent.

5. RESULTS AND DISCUSSION

In this section, we present first an illustrative simulation example for HASCE on CPN Tools which allows
us to generate random values for the simulation, indicate the different states of each agent and display the
results in comments (in green rectangles) as shown in figures 7 to 10. Then, to validate our results, we
plotted system outputs by recovering 50 samples of random values modalities to visualize and simulate the
behavior of the proposed SMA.

5.1. Simulation results on CPN Tools

Indeed, as shown in figure 7, the switching energy agent perceives that energy demand is greater than the
amount of renewable energy produced, so it requests relief to the network (see comment in figure).
In figure 8, the values generated randomly give that the selected mode is automatic and the presence
sensor has detected the presence of a person in the intelligent building (IB). So, it sends the information to
the thermal and light agents. If inhabitant switches to the manual operating mode of HASCE, the switching
control agent still operational.
Light comfort agent acts according to the random values provided by the modeled outdoor and indoor light
sensors in CPN Tools. In our example (Fig. 9), the shutters are down with 40% while the graduated lamp
lights by 20%.
Thermal comfort agent is able to maintain ambient temperature in the IB. In figure 10, it detects that the
temperature is equal to16 degrees; then it turns on the heater for regulating temperature by adding 3 degrees
to achieve ambient temperature.

5.2. Results validation on Matlab

The figure 11 shows us 50 samples of input modalities and their respective system responses. We note
for example sample 1 (indicated by a dash-dotted line in figure 11), manual operating mode is selected,
thermal and light comfort agents are off service. In the second example at the sample 11, automatic
operating mode is selected. The temperature is 22 °, the thermal comfort agent actuates neither heater nor
air-conditioner since temperature is ambient; while the lighting comfort agent downs the shutters and the
graduated lamp lights by 100% at 1pm.
After checking all the samples, we can conclude that the system responds correctly based on criteria
established in the beginning of our work. The proposed architecture displays the appropriate command
considering the input modalities for better energy management in IB.

426
 AUTO/MANU
1

0.5

0
0 5 10 15 20 25 30 35 40 45 50
time modality

40

20

0
0 5 10 15 20 25 30 35 40 45 50
light control shutters control thermal comfort temp modality

40

20

0
0 5 10 15 20 25 30 35 40 45 50
20

-20
0 5 10 15 20 25 30 35 40 45 50
100

50

0
0 5 10 15 20 25 30 35 40 45 50
100

50

0
0 5 10 15 20 25 30 35 40 45 50
50 samples generated modalities

Fig. 10. Graphical representation of results.

6. CONCLUSION

This paper has presented the principles of a proposed HASCE, which allows the agents to cooperate
and coordinate their actions in order to find an acceptable near-optimal solution for power management.
We showed why autonomous and cooperative agents are particularly well suited for power management
in buildings. In our structure we used four agents: switch energy sources agent, light comfort agent,
temperature comfort agent and inhabitant agent. A cooperation mechanism that reduces the problem
complexity has been detailed.
The validation of the proposed structure is done using CPN Tools and Matlab. We used the TCPN for
modeling the proposed agents and simulating the proposed structure. The obtained results show the
good correlation between inputs and outputs of the system. Indeed, the proposed multimodal multi-agent
architecture model acts with the expected behavior. Therefore, we can say that the results validate the
proposed approach.

Subscripts

IB Intelligent Building
TPN Timed petri nets
TCPN Timed Colored Petri Nets
HASCE Home Automation System Controlling Energy

REFERENCES

Luis Pérez-Lombard J.O and C.P. Elsevier 2007. A review on buildings energy consumption information. Energy and
Buildings, Volume 40, Issue 3, 2008, Pages 394-398.

APRUE Bilan de la consommation énergie finale en Algérie, 2010. Agence de promotion et de rationalisation de
l’utilisation de l’énergie, affiliée au ministère de l’énergie et des mines.

Liyanage C. De Silva, C.M. and M.P. Elsevier 2012. State of the art of smart homes. Engineering Applications of
Artificial Intelligence. 25 (2012) 1313–1321.

427
E. Sierra, A. H. ,D. R., M. G-M. and P. B. Springer 2008. Optimizing Building’s Environments Performance Using
Intelligent Systems . New Frontiers in Applied Artificial Intelligence Lecture Notes in Computer Science Volume 5027,
pp 486-491.

Shadi Abras. 2009. Système domotique Multi-Agents pour la gestion de l’énergie dans l’habitat. Thèse de doctorat.
Laboratoire d’Informatique de Grenoble – Imag. Ecole doctorale Mathématiques, Sciences et Technologies de
l’Information, Informatique (MSTII) - Grenoble INP.

J.K.W. Wonga, H. L., S.W. Wang. Elsevier 2005. Intelligent building research a review. Automation in
Construction 14 143– 159.

Michael Wigginton, J. H. Taylor & Francis, Elsevier 2002. Intelligent Skins. Intelligent building research. 52-175.

(*) http://www.smartgrids-cre.fr/index.php?rubrique=dossiers&srub=batiment&action=imprimer

T.Dujardin, J.R. J-C. R. and J-C. T. UNDEF 2011. Gestion intelligente d’un contexte domotique par un Système
Multi-Agents. Actes Journées Francophones sur les Systèmes Multi-Agents, France.

Zhu Wang, R.Y and L.W. IECON 2010. Multi-agent Control System with Intelligent Optimization for Smart and
Energy-efficient Buildings. 36th Annual Conference on IEEE Industrial Electronics Society.

Lingfeng Wang, Z.W. and R.Y. 2012. Intelligent Multiagent Control System for Energy and Comfort Management
in Smart and Sustainable Buildings. IEEE TRANSACTIONS ON SMART GRID, VOL. 3, NO. 2.

Rim MISSAOUI BADREDDINE. 2012. Gestion Énergétique optimisée pour un bâtiment intelligent multi-sources
multi-charges : différents principes de validations. Thèse de Doctorat - Laboratoire de Génie Électrique de
Grenoble (G2eLab) - Ecole Doctorale ‘Electronique, Electrotechnique, Automatique et Traitement de Signal’
EEATS.

G. Clarke, S.S and V. C. 1999. A multi-agent architecture for intelligent building sensing and control. International
Sensor Review Journal, 19(2) :135–140.

Jianmin Duan and F.L. INDIN 2008. Research of Intelligent Building Control Using an Agent-based Approach. The
IEEE International Conference on Industrial Informatics DCC, Daejeon, Korea.

Nadia Touileb Djaid, S.N. and A.R. 2012. Multimodal architecture to strengthen the interaction of the robot in
ambient intelligence environments, Multi-disciplinary workshop on Smart Gadgets Meet Ubiquitous and Social
Robots on the Web (UbiRobs), 14th ACM International Conference on Ubiquitous, computing, Pittsburgh,
Pennsylvania, United States.

Jean-Marie. Boussier. 2007. Modélisation de comportements dans les systèmes dynamiques : Application à
l’organisation et à la régulation de stationnement et de déplacement dans les Systèmes de Trafic Urbain. Thèse
de doctorat – Laboratoire Informatique, image et Interaction de LA ROCHELLE – Ecole Doctorale ED 0374.

428
 QUANTUM DOTS CONJUGATED E. COLI LIVING CELLS AS FLUORESCENT REPORTERS
TO DETECT CYTOTOXICITY OF CHEMICALS

R.S. Chouhan, J.H. Niazi, and A. Qureshi*

Sabanci University Nanotechnology Research and Application Center, Orta Mah., Tuzla, Istanbul, TURKEY
e-mail: anjum@sabanciuniv.edu, javed@sabanciuniv.edu

ABSTRACT
Quantum dots (QDs) have attracted a much of interest in recent years for imaging, diagnostics and therapy due to
their unique optical properties, such as broad excitation spectra and long fluorescence lifetime. In this study, a
controlled bioconjugation using CdTe QDs with gram negative E. coli cells was performed to develop QD-E. coli
bioconjugates. These bioconjugates were used as baits to determine cytotoxicity of model toxic chemicals, such as
oxidative stress inducer (H2O2) and a pesticide (methyl viologen or paraquat). These chemicals at concentrations
ranging 0.5-5 mM were exposed to QD-E.coli bioconjugates that interacted with cells and the fluorescent responses
with QD-E.coli bioconjugates was analyzed. The results showed that the fluorescent ability of QD-E.coli
bioconjugates tend to diminish with increasing concentration of toxic chemicals. This stress is attributed to the
damages occurred as a result of interaction of toxic chemicals to the cell-wall or membrane of cells that resulted in
loss of fluorescence signal. This loss in fluorescence (signal off phenomena) of QD-E.coli bioconjugates can be used
as probes to develop a variety of fluorescence based detection kits for the rapid determination of toxic drugs or food
sample testing.

INTRODUCTION

The properties and applications of fluorescent semiconductor nanocrystals stand among the most exciting
research fields in chemistry, physics and biology (Stinaff et al., 2006). QDs have great advantages over traditional
fluorescent dyes such as stronger luminescence, better photo-stability against bleaching and physical environments,
such as pH, temperature and optical tunability. These properties have been used in immunoassays, molecular
imaging and in vivo biological labels (So et al., 2006). Optically quenched QDs have recently gained tremendous
considerable interest in environmental applications (Donald et al., 2005). Simple assays such as by direct
visualization of luminescence from the living cells decorated with quantum dots (QDs) as labeling probes can be
used as baits. The water soluble quantum dots (QDs) are non-toxic because the polymer coatings are better
alternatives as labeling probes in biological and biomedical applications. Until now, QDs as labeling probes have
been successfully used in fluorescent resonance energy transfer (FRET) (Li et al., 2010), in vitro and in vivo
imaging (Bae et al., 2009; He et al., 2009), immunoassay(Chen et al., 2010; Jokerst et al., 2009) and DNA
hybridization (Li et al., 2011). Hence, bioconjugation of QDs is of great importance in developing whole-cell based
biosensors.
The carboxylated QDs can be conjugated to the amino groups of biomolecules, such as proteins, enzymes or
antibodies. This linking approach is simple, rapid and it is widely used in certain biosystems, and the conjugation
efficiency can be improved by chemically modifying surface charge states of proteins and QDs as well as their
surface modifications(Shakya, Sami, Srivastava, & Kumar, 2010),(Gerhards, Schulz-Drost, Sgobba, & Guldi,
2008). Repulsion of QDs and proteins/whole-cells may prevent their conjugation process. In this case, a high ionic
strength or higher concentration of proteins/cells can be used for conjugation, which may be accompanied by
partly quenching of QDs’ luminescence intensity.(Ren, Jeung, Han, & Choi, 2007)
In the present work, we covalently attached QDs on the surface of living E. coli cells to give rise to QD-E. coli
bioconjugates. These bioconjugates were used as a new molecular tool for probing cellular damages induced by
model chemicals. It required labeling of cells in order to probe their responses against the model chemical in the form
of detectable signals. Labeling of cells was done by non-toxic chromogenic water-soluble, capped QDs that emit light
or attain fluorescent abilities while cells are active and viable. This enables monitoring the cellular interaction and
behaviour of cells when exposed to external perturbations, such as nanomaterials, environmental contaminants, food
toxins that also find application in medical diagnostics.
Our results demonstrated that the fluorescence signal was diminished in a concentration dependent manner with
H2O2 and paraquat, which is an oxidative stressor and a pesticide, respectively. The developed method demonstrated
the cellular integrity can be evaluated in terms of fluorescent signal generated by the QDs localized on the cell-
surfaces, after exposure of different toxic chemicals. Thus, it is possible to probe the cellular damages occurred by
the interaction of toxic chemicals using QD-E. coli bioconjugates developed in this study.

MATERIALS AND METHODS

Chemical and reagents

Wild-type E.coli DH5α strain was used as a model living bacterial cells in this study. Luria-Bertani broth (LB-broth)
and Luria-Bertani agar (LB-agar) were obtained from Difco (MI, USA). N-hydroxysuccinimide (NHS) and N-ethyl-N′-
(3-(dimethylamino) propyl) carbodiimide (EDC), Paraquat and H2O2 were purchased from Sigma-Aldrich. Triton-X 100
was procured from Merk, Germany. A 10 mM Phosphate Buffer Saline (PBS), pH-7.4 used in this study was prepared
from a 10X stock solution accordingly and prepared with deionized water (Milli Q, Millipore, Barnstead, CA, USA).
Qdot® 585 carboxyl quantum dots (Invitrogen) were used as labeling probes having emission maxima at ~585 nm.

429
All glassware’s and solutions used in these studies prior to experiment were autoclaved with HMC, Hirayama
autoclave, Che scientific co., Hong Kong. All bioconjugation studies and related work were carried out in Thermo
scientific HERAsafe® KS microbiological safety cabinet. The spectral studies of QD-E. coli bioconjugate with
Paraquat and H2O2 were done in a range 500 nm to 750 nm using advanced NanoDrop 3300 Fluorospectrometer
(Thermo Scientific NanoDrop Products). All other reagents used were of analytical grade.

Preparation of E.coli culture

Lyophilized cells of E.coli DH5α were precultured in LB-broth at 37 °C for 15 h and then the cells were harvested by
centrifugation at 3000 rpm for 5 min at 4 °C. The cells were subsequently washed thrice with sterile PBS followed by
centrifugation for 5 min at 3000 rpm at 4 °C. The cell pellet was suspended in PBS and colony forming units (CFU)
were measured by serial dilution followed by plating on LB-agar plates. The cell-suspension was divided into several
aliquots carrying same number of cells (~109 CFU/mL) for test and control experiments.

Bioconjugation of E. coli with QDs

Bioconjugation of E. coli with QD was successfully conjugated and detailed studies were done in our laboratory.
Briefly, 50 mM EDC and 5 mM NHS and incubated for 5 min at 25 °C. This mixture was quickly added with 2 μL of 8
mM QD solution and the entire reaction mixture was again incubated at 25 °C for 15 min. under constant shaking at
110 rpm. The QD-conjugated cell-suspension (QD-E. coli bioconjugates) was centrifuged at 1500 rpm for 3 min and
the supernatant was removed and washed the pellet thrice after resuspending in PBS, pH 7.4 followed by
centrifugation with same buffer and stored at 4 °C until further use.

Viability tests of cells before and after treatment

About 25 µL of cell suspension from all the sample tubes (control and chemical treated samples) were spread on
Luria-Bertani (LB) agar plates. Colony forming units (CFU) originating from a single living cell was counted. Images of
plates with visible colonies were captured for comparison.

RESULTS AND DISCUSSION

Response of bioconjugates against H2O2

Wild-type E. coli (DH5α) cells were covalently conjugated with carboxyl-QDs by EDC/NHS coupling. The QD-E.coli
bioconjugates generated exhibited a characteristic fluorescent emission peak at 585 nm, which was derived from the
cell-bound QDs (Fig. 1). The covalent coupling strategy was employed to immobilize QDs on the surface of the cells
without affecting the cellular integrity or viability.

Fig.1. Fluorescence emission spectra of QD-E. coli bioconjugates with or without H2O2 incubation showing complete
disappearance of fluorescence emission. The fluorescent spectra shown in black was obtained against PBS as the blank.

H2O2 generates hydroxyl radical (OH) that induces oxidative toxicity in all living cells. The cellular damage or toxicity
that occurred to the QD-E.coli bioconjugates was studied by exposing H2O2 and paraquat as model toxic chemicals
that induce generation of OH and O2−, respectively in cells. The cytotoxic effects of these model chemicals on QD-
E.coli bioconjugate were determined using change in fluorescence emission spectra. The effect of various
concentrations (0.5–5mM) of these chemicals on cells was studied through changes in fluorescent characteristics of
bioconjugates. Treatment of bioconjugates with initial 0.5 mM H2O2 concentration resulted in diminishing response of
their characteristic fluorescent peak at 585 nm (Fig.1). At above 0.5 mM concentrations of H2O2, the fluorescent peak

430
was completely disappeared due to severe cellular toxicity. These results showed that higher concentrations of H2O2
severely affected the cellular integrity of E. coli cells as a result of change in cellular integrity of QD-E. coli cells and
therefore, the fluorescence signal was diminished.

Response of bioconjugates against methyl viologen (Paraquat)

Paraquat is a pesticide chemical that generates superoxide radicals (O2−) that are toxic to living cells. The cellular
damage that occurs to cells with paraquat was studied by exposing with three different concentrations (0.5- 5 mM) of
paraquat (Fig.2).

Fig.2. Fluorescence emission spectra of QD- E. coli bioconjugates with or without paraquat incubation showing diminishing
of fluorescence emission. The fluorescent spectra shown in black was obtained against PBS as the blank.

It was observed that the fluorescence emission at 585 nm gradually decreased with increasing concentration of
paraquat (Fig. 2). Exposure of above 1 mM concentration of paraquat, the fluorescence signal dramatically dropped
suggesting that the cells severely affected with paraquat above 1 mM concentration. Paraquat, however did not have
detrimental effects on cellular structure at 2 h initial exposure with lower concentrations. Elevated levels of PQ
however showed destructive effects because of the significant loss of fluorescence from bioconjugates with
extended incubation (3 h). The above results was also supported by the results of viability tests where H2O2
severely damaged the cells and therefore cells tend to lose their viability right at the initial stages compared to PQ
(Fig. 3). The colonies of living cells in control and test samples that formed visible colonies suggest that the
bioconjugated cells tend to lose their viability with increasing time of chemical exposure (from left to right in Fig. 3)
consistent to the results of fluorescence emission measurements (Figs. 1 and 2).

Fig. 3. Control and test QD-E. coli bioconjugate cells exposed to paraquat and H2O2 at different time intervals were plated
on LB-agar plates. The figure shows colonies of living cells in control and test samples that formed visible colonies. Those
cells that severely affected by the chemical fail to grow or form colonies.

CONCLUSIONS

431
 In this work, QD-E. coli bioconjugate was developed to probe the cellular damages caused by model toxic
chemicals, such as H2O2 and paraquat. To probe the cellular damages or toxic effects of model chemicals, optical
properties of QD was utilized through labeling the living E. coli cells, which provided the cellular damage information
in the form of fluorescence emission responses. The results demonstrated that the fluorescent ability of these
bioconjugates diminishes if the toxic chemicals interact with QD- E. coli conjugates. The strategy reported in this
study may be useful for creating a novel methodology for investigating cellular interactions with toxic chemicals to
determine their cytotoxicity. Our results demonstrated that QD-conjugated cells served as an important tool to
monitoring the impact of chemicals has on cells. The method reported in this paper is robust, rapid and easy to use
that find applicable in environmental and biomedical applications.

Acknowledgements

This work was supported by the Scientific and Technological Research Council of Turkey (TUBITAK), project grant
no. 112E051.

REFERENCES
Bae, P. K., Kim, K. N., Lee, S. J., Chang, H. J., Lee, C. K., & Park, J. K. 2009. The modification of quantum dot
probes used for the targeted imaging of his-tagged fusion proteins. Biomaterials, 30(5): 836-842.
Chen, L. P., Sheng, Z. H., Zhang, A. D., Guo, X. B., Li, J. K., Han, H. Y., & Jin, M. L. 2010. Quantum-dots-based
fluoroimmunoassay for the rapid and sensitive detection of avian influenza virus subtype H5N1. Luminescence,
25(6): 419-423.
Donald S., Marcus J., Garry R., Pingrong Y., Jovan N. and Sean S. 2005, Quenching of Semiconductor Quantum Dot
Photoluminescence by a π-Conjugated Polymer. J. Phys. Chem. B 109: 15927-15932.
Gerhards, C., Schulz-Drost, C., Sgobba, V., & Guldi, D. M. 2008. Conjugating Luminescent CdTe Quantum Dots
with Biomolecules. J. Phys. Chem. B 112(46): 14482-14491.
He, Y., Su, Y. Y., Yang, X. B., Kang, Z. H., Xu, T. T., Zhang, R. Q., Fan, C. H., & Lee, S. T. 2009. Photo and pH
Stable, Highly-Luminescent Silicon Nanospheres and Their Bioconjugates for Immunofluorescent Cell Imaging. J.
Am. Chem. Soc. 131(12): 4434-4438.
Jokerst, J. V., Raamanathan, A., Christodoulides, N., Floriano, P. N., Pollard, A. A., Simmons, G. W., Wong, J.,
Gage, C., Furmaga, W. B., Redding, S. W., & McDevitt, J. T. 2009. Nano-bio-chips for high performance
multiplexed protein detection: Determinations of cancer biomarkers in serum and saliva using quantum dot
bioconjugate labels. Biosens. Bioelectron. 24(12): 3622-3629.
Li, Y. Q., Guan, L. Y., Wang, J. H., Zhang, H. L., Chen, J., Lin, S., Chen, W., & Zhao, Y. D. 2011. Simultaneous
detection of dual single-base mutations by capillary electrophoresis using quantum dot-molecular beacon probe.
Biosens. Bioelectron. 26(5): 2317-2322.
Li, Y. Q., Wang, J. H., Zhang, H. L., Yang, J., Guan, L. Y., Chen, H., Luo, Q. M., & Zhao, Y. D. 2010. High-
sensitivity quantum dot-based fluorescence resonance energy transfer bioanalysis by capillary electrophoresis.
Biosens. Bioelectron. 25(6): 1283-1289.
Ren, L., Jeung, W. Y., Han, H. C., & Choi, H. J. 2007. Optical activation of Si and Ge nanowires codoping with Er:
Yb rare earth by sol-gel methods. J. Nanoelectron. Optoelectron. 2(2): 191-196.
Shakya, A. K., Sami, H., Srivastava, A., & Kumar, A. 2010. Stability of responsive polymer-protein bioconjugates.
Progress in Polymer Science, 35(4): 459-486.
So, M. K., Xu, C., Loening, A. M., Gambhir, S. S. and Rao, J. 2006. Self-illuminating quantum dot conjugates for in
vivo imaging. Nat. Biotechnol 24:339-343.
Stinaff, E. A., Scheibner, M., Bracker, A. S., Ponomarev, I. V., Korenev, V. L., Ware, M. E., Doty, M. Reinecke, F. T. L.
and Gammon, D. 2006. Optical signature of coupled Quantum Dots. Science 311:636-639.

432
 THE EVALUATION OF TURKISH ELECTRICITY DEMAND PROJECTIONS

Betül Özer
Istanbul Technical University Eurasia Institute of Earth Sciences, Maslak, İstanbul, Turkey
betlozer@hotmail.com

ABSTRACT
Electricity demand and energy models are required for a variety of utility activities. According to the results of the
large number of studies in the literature about electricity consumption forecasting and modeling, variables affecting
demand and energy consumption may vary from one region to another. A model developed for one region may not
be appropriate for another region. Economic indicators such as population, Gross Domestic Product (GDP), GDP
per capita, imports and exports, prices, and the other socioeconomic factors such as continuous pressures for
better living standards, access to the electricity are the variables those were taken into consideration for the
forecast studies. There are also several studies in the literature on forecasting the electricity consumption of
Turkey used various models with different variables accordingly presenting different results. The aim of this study
is to overview of these studies and to evaluate the different results of them.

Key words: electricity demand, forecasting, energy modeling, Turkey.

Abbreviations:
ARIMA: Autoregressive Integrated Moving Average, ACOEEE: ant colony optimization electricity energy
estimation, MAED: Model for Analysis of Energy Demand, ANN: Artificial Neural Networks, ACO: Ant Colony
Optimization, DPT: State Planning Organisation, EFOM: Energy Flow Optimisation Model, GA: Genetic Algorithm,
GDP: Gross Domestic Product, GHG: Greenhouse Gas, GNP: Gross National Product, GP: Genetic
Programming, GPRM: Grey Prediction with Rolling Mechanism, GNP: Gross National Product, GW: gigawatt,
kWh:kilowatt-hour, MENR: Ministry of Energy and Natural Resources, MLP: Multi Layer Perceptron, SPO: State
Planning Organization, SIS: State Institute of Statistics, STSM: Structural Time Series Model, SVR: Support Vector
Regression, TEIAS: Turkish Electricity Transmission Company, TWh: terawatt-hour, UEDT: Underlying Energy
Demand Trend.

1. INTRODUCTION

Energy is one of the most significant components in the economic development of countries. It is also certain that
energy is the most important necessity of human life and, hence, there is an increasing relation between the level
of development and the amount of energy consumption. Worldwide energy consumption is rising fast because of
the increase in human population, continuous pressures for better living standards, emphasis on large-scale
industrialization in developing countries and the need to sustain positive economic growth rates. Given this fact, a
sound forecasting technique is essential for accurate investment planning of energy production, generation and
distribution (Bianco et al., 2009).

The estimation of electricity consumption is of great importance for effective implementation of the electricity
policies, such as conservation programmes, the planning of capacity expansion and the construction of nationwide
interconnection of power network (Öztürk and Ceylan, 2005). Accurate energy planning in order to avoid an
electricity shortage and guarantee adequate infrastructures studies focusing on the forecasting of electricity
demand using different techniques are important (Bianco et al., 2009).

The most common difficulty for a good prediction of the electricity demand is the determination of sufficient and
necessary information and selecting the most related variables. Electricity is not demanded for its own sake; it is a
derived demand that comes from the demand for lighting, heating, cooling, it is the demand for the services it
produces with the capital stock in place at a certain time. The amount of energy consumed is therefore connected
to the technological level of the energy appliances to assure the demanded level of services. Consequently there
are a number of exogenous factors that might influence electricity demand behavior, such as improved technical
efficiency of the capital and appliance stock, and changes in consumer tastes, preferences, demographic
parameters, social structures, environmental regulations, economic structure like GDP per capita or economic
growth, business cycles etc.

Various studies in the literature indicate that the relationships between electricity consumption and several
potentially relevant variables, like prices, consumer income, key weather parameters, number of customers and
tourists, sectoral demand trend, change from a country to another (Egelioğlu et al., 2001). These variables can all
or partially have an influence on electricity demand according to the country specific characteristics. The predictive
ability of the models those are applied to energy demand forecasts depends on the variables, estimated value of
them and the assumption of the influence factor of them to the electricity demand (Bianco et al., 2009).

433
It is obvious that consumption is related to population as the population increases, more electricity will be
consumed. Besides the increase ratio of the population is usually less than the increase ratio of the energy
consumption. This proves there are some other factors affecting the electricity increase. Gross national product
(GNP) is a measure of all economic activities and increasing GNP means improved living standards that cause
increase of energy usage. Imports and exports are related to manufacturing processes and therefore strongly
affect electricity consumption. (Kavaklıoğlu et al., 2009). These factors are generally difficult to observe directly but
arguably play an important role in determining electricity demand behavior (Dilaver and Hunt 2011a).
Consequently the goal of accurate modeling of consumption requires attention to a few extremely important points.
The first point is to identify all the necessary variables and parameters that contribute to electricity consumption in
a given country. The second point is to choose a modeling methodology that can handle the difficulties of the
consumption modeling task. One difficulty in this area, similar to many other modeling studies, is the fact that the
relationship between the input variables and the output variable is nonlinear and the nature of the nonlinearity is
not known very well. Therefore it is not easy to postulate the form and/or order of a mathematical function whose
parameters could be estimated through regression type computations (Lippmann 1987; Kavaklıoğlu 2011).
Although being the most important aspect, accuracy of the model is never the only concern in any modeling
analysis. Models should also be able to handle imperfections in the data such as noise, errors, missing data
points, disturbances and short term effects. Another area of concern in modeling is so called local minima. Most
modeling techniques compute model parameters by minimizing a cost function which may not have a single local
minimum that is also the global minimum. Then getting trapped in one of the local minima avoids reaching the real
optimal parameters. Therefore a good modeling method should either be able to guarantee reaching global
minimum or should have a single global minimum for its cost function. In summary, the models should exhibit good
global optimization and robustness characteristics as well as accuracy in prediction. Furthermore, the
methodology used should allow future prediction by letting the model topology cover one-step or multi-step ahead
predictions. This is critical since the eventual goal here is to predict the future values of electricity production that
leads to dynamic model structures. It is well-known that time series formalism using nonlinear models is very
suitable for these types of problems. In addition, the methodology should allow rangeability as often the data
available for modeling does not cover the ranges of data for the times the predictions need to be done
(Kavaklıoğlu, 2011).

Since energy security is the one of the most important issues of the 21st century. Arguably, the policies and
strategies cannot be neither assessed nor constructed without sound demand forecasts (Dilaver and Hunt 2011b).
Therefore, demand forecast results of different kind of studies have great importance when implementing future
energy policies especially for energy supply and security and when demand grows faster, as in the case of Turkey.

Turkish Parliament ratified the Kyoto protocol in February 2009, which is leading to the introduction of legally
compulsory commitments for the reduction of greenhouse gases (GHGs). It is therefore commonly expected there
will be a change in Turkish energy policy that might include CO2 taxes. That has already regulations including the
incentives of the renewable energy uses. Regarding this new environment in which markets will operate,
determining the total and sectoral electricity demand (including industrial, residential and services etc.) function
might have a significant importance on evaluating new policy implications. Sensible and reliable energy demand
forecasts assist in financing and developing the necessary measures for the sustainable economic growth of
Turkey (Dilaver and Hunt 2011b). As the various studies have been done, the results differ from each other as
they used different models, variables with different future assumptions. The effects of the variables to the demand
forecast mainly depend on the model setup.

From the point of view the aim of this paper is to present an overview of the studies in literature on electricity
demand projections of Turkey. The results were evaluated by putting forward to what extend the projections can
be done with various assumptions of the variables used in different methodologies. The results can also indicate to
the officials the importance of using alternative forecasting methodologies.

2. BACKGROUND: DEVELOPMENT OF TURKISH ELECTRICITY SECTOR

Turkey’s electricity demand has been growing very rapidly as a result of the rate of population growth, a rapid
industrialization, and economic development. It has increased from 47 TWh in 1990 to 186 TWh in 2011, with an
annual average growth rate of 7%. Only in the economic crisis years (2001 and 2009) the demand decreased. The
installed capacity has increased from 16.3 GW in 1990 to about 53 GW in 2011 by 5.8%growth rate (Fig. 1,
TEIAS; 2012a).

The electricity industry is a large, high-growth sector in the Turkish economy. The industry contributes significantly
to the country’s GDP and is a US$12 billion industry at current end-user prices (Balat 2009).
On the other hand an electricity consumption per capita in Turkey is below the world average. Despite increasing
demand, Turkey’s per capita gross consumption is still very low at about 2,900 kWh compared to the EU average
of about 5,700 kWh.

434
 200 60
180
Electricity consumption, TWh 50
160

Installed capacity, GW
140 40
120
100 30
80
60 20
40 10
20
0 0
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
INSTALLED Capacity Net Electricity consumption

Fig. 1: Turkish electricity system development (1990-2011) (TEIAS, 2012a)

In Turkey the increase in electricity demand for the period of 1990 to 2006 was faster than the GDP increase as
shown in the Fig. 2. GDP grew by 4% on an average growth rate where the electricity consumption growth rate
was higher by approximately 7.3% in the same period (DPT, 2010, TEIAS, 2012a).

14
12
10
8
Growth rate, %

7,29
6
4 4,00
2
0
-2
-4
-6
-8
-10
1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

GDP growth rate Gross demand increase

Average GDP increase Average gross demand increase
Fig.2: Turkish electricity demand and GDP growth (1990-2006) (DPT, 2010, TEIAS, 2012a).

3. REVIEW of TURKISH ELECTRICITY DEMAND STUDIES

In the planning of investment strategies of production, transmission and distribution systems, mid- and long-term
electricity forecasting is very important.

Studies before the 1970s that directly focused on analyzing Turkish Electricity demand are very limited, being
generally carried out by governmental institutions with their own approaches, namely the State Planning
Organization (SPO), the State Institute of Statistics (SIS) and the Ministry of Energy and Natural Resources
(MENR). Although some research in those institutions tried to apply mathematical modelling techniques to analyze
energy demand in the late 1970s, these methods were not used in official energy planning until 1984. Before 1984,
national energy policy was shaped by the forecasts of the SPO in which they employed various simple best-fit
curves (Ediger and Tatlidil, 2002, Dilaver and Hunt 2011b). The MENR utilized ‘balance’ and ‘impact’ models in
order to determine energy demand functions and to make future projections and also used the simulation models:
Model for Analysis of Energy Demand (MAED), Wien Automatic System Planning WASP III) and Energy Flow
Optimisation Model (EFOM-12 C Mark). The SPO also developed its own models based on sectoral energy
demand for different consumer groups and subgroups whereas the SIS explored the relationship between
demographic factors and economic parameters with energy demand in its models (Dilaver and Hunt 2011c).

435
Since 1984, the MENR is preparing Turkey’s future electricity projections yearly and has carried out projections for
Turkey’s energy demands by using MAED simulation technique. In the model, energy demand forecasting is
realized by considering past situations of the country’s economic, demographic and social structure. They also
have some deviations due to the deviations of the variables.

There was a large increase in the number of Turkish energy modelling studies after the late 1990s and they can be
divided into three main groups. In the first group, the focus is on investigating the causality between energy
consumption and economic variables termed Causality Studies. In the second group, the focus is on identifying the
relationship and the magnitude of the key relationships mainly the elasticities between economic variables and
energy consumption termed Relationship Studies. And in the third group, the aim is to forecast future energy
demand using a number of different approaches termed Forecast Studies (Dilaver and Hunt 2011b).

Most of the studies for Turkish electricity demand projections in the literature are overviewed in this section.

Yumurtacı and Asmaz (2004) proposed an approach to calculate future energy demand of Turkey for the period of
1980–2050, based on the population and energy consumption increase rates per capita. The multiplication of
population value and the consumption per capita gives the necessary energy amount. According to the
calculations for the year 2050, if an individual uses energy of 10,197 kWh in a year, then 115 million individuals will
need an electric energy of 1,173 billion kWh.

Ozturk et al., (2005) used the genetic algorithm (GA) approach to investigate the relationship between total
electricity consumption, GNP, population, imports and exports for the period 1980-2001 in Turkey with annual
data. Total electricity demand of Turkey was estimated as 220 TWh and 330 TWh in 2020 with exponential and
quadratic forms of the genetic algorithm electricity demand models respectively. They mentioned that the
estimation of electric energy demand, based on the socioeconomic indicators, may be estimated with various
forms of mathematical expressions. These equations may be linear or non-linear forms. The non-linear forms of
the equations can better estimate the future electricity demand of Turkey due to the fluctuations of socioeconomic
indicators.

However in the study of Ozturk and Ceylan (2005) total electricity consumption is estimated based on GNP,
population, import and export figures of Turkey. Industrial sector electricity is calculated based on the GNP, import
and export figures. Three forms of the genetic algorithm electricity demand (GAED) models for the total and two
forms for the industrial electricity consumption are developed. The best-fit GAED model in terms of total minimum
relative average errors between observed and estimated values which is GAEDquadratic (GAEDquad) is selected
for future demand estimation. High- and low-growth scenarios are proposed for predicting the future electricity
consumption. The electricity consumption in 2020 with the GAEDquad for the low-growth scenario will be about
462TWh. The GAEDquad estimation for the high growth scenario is about 500TWh. The following conclusions can
be drawn from the study as the GA notion can be applied as an alternative way of electricity consumption
estimation for the future by appropriately fitting the mathematical expressions to the existing data. The reason for
applying the GA is that it is flexible in nature to provide many optimal, or near optimal solutions to estimate the
future trends of the electricity consumption. This advantage comes from the GA approach itself, because the GA
starts the solution of the problem from a large population base. The model is actually a multi-parameter solution,
which has many feasible solution points. The inclusion of the import and export figures in the electricity demand
estimation improves the GAED model performance.

In the study of Tunç et al., (2006) Turkey’s electric energy consumption rates are predicted by regression analysis
for the years of 2010 and 2020 and finally linear mathematical optimization model is developed to predict the
distribution of future electrical power supply investments in Turkey. In 2020 it was predicted that Turkey’s
electricity consumption will be about 322 TWh.

Erdoğdu (2007) has used cointegration analysis and autoregressive integrated moving average (ARIMA)
modelling, for both providing an electricity demand estimation and forecast, and comparing the results with official
projections. The study concludes, first, that consumers’ response to price and income changes is quite limited and
therefore there is a need for economic regulation in the Turkish electricity market; and second, that the current
official electricity demand projections highly overestimate the electricity demand, which may endanger the
development of both a coherent energy policy in general and a healthy electricity market in particular. The data
used in the estimation process is quarterly time series data on real electricity prices, real GDP per capita and net
electricity consumption per capita for the period 1984-2004, a total of 84 observations. The electricity demand of
Turkey is predicted to be 160 TWh by the year 2014.

Hamzaçebi (2007) used the Artificial Neural Networks (ANN) model with a time series structure for the period 1970
and 2004 to analyze sectoral electricity consumption in order to forecast sectoral and aggregate electricity
demand. Residential, industrial, agricultural and transportation sectors were considered. The results indicate that
aggregate electricity demand would be just below 500 TWh in 2020. Although this network structure is well suited
for prediction beyond existing data set; the confidence in predictions decreases as time moves forward
(Kavaklıoğlu et al., 2009). ANN, imitating the functioning of the human brain, is a tool of great importance in

436
sample classification, pattern recognition, and forecasting. Most widely used ANN type for forecasting is a multi
layer perceptron (MLP). In an MLP designed for time series forecasting, determining the variables, such as
number of input, hidden and output neurons, is highly important. However, these parameters are subject to change
with respect to the problem at issue (Hamzaçebi, 2007).

Grey prediction with rolling mechanism (GPRM) approach is proposed to predict the Turkey’s total and industrial
electricity consumption in the study of Akay and Atak (2007). GPRM approach is used because of high prediction
accuracy, applicability in the case of limited data situations and requirement of little computational effort. Results
show that proposed approach estimates more accurate results than the results of MAED, and have explicit
advantages over extant studies. Grey prediction is thought to be an alternative forecasting tool for those systems
whose structure is complex, uncertain and chaotic as in the case of Turkey’s energy demand structure. Grey
prediction has been widely used in forecasting studies due to its advantages, requiring low data items to build
forecasting models and higher forecasting accuracy when compared with other forecasting techniques. The rolling
mechanism is an efficient technique to increase forecasting accuracy of grey prediction in case of having
exponential and chaotic data. The electricity consumption data for Turkey from 1970 to 2004 is modelled and it
showed an exponential trend for the period of 1970-2004. The model is applied separately to forecast total and
industrial sector electricity consumption. It is expected that industrial and total consumption will be 140.37 and
265.7 TWh in 2015, respectively. The reason for the difference with the MAED model results are mentioned as
MAED uses too many indicators, and in case of the existence of a high level of variability between them, this
simply brings about error effect on the results. On the other hand, GPRM uses only the consumption data of the
last four years. In this respect, it has both simplicity and much better prediction accuracy, and can be implemented
reliably (Akay and Atak 2007).

Yüksek (2008) has evaluated some previous demand projection studies. He has evaluated nine projection results
of MAED model which is used for Turkey’s long-term electricity energy demand by Turkish Electricity Transmission
Company (TEIAS). According to the comparison of the results with the observed electricity demand values for the
years 2005, 2006 and 2007, due to very well prediction values only three of those were taken into consideration for
future demand values. As the results of comparison values for the years 2005, 2006 and 2007 Yumurtacı and
Asmaz (2004), Tunç et al., (2006) and Hamzaçebi (2007) have average 11.3%, 3.5% and 19.6% relative errors
respectively, these studies weren’t taken into consideration in Yüksek (2008)’s study. Six new scenarios were
applied in the study according to the different assumptions of the increase ratios for population, electric energy
consumption per capita and total electric energy consumptions values. Depending on the applied scenarios and
MAED model results, Turkey’s annual electric energy demand in 2010, 2015 and 2020 varies between 222 and
242 TWh; 302 and 356 TWh; and 440 and 514 TWh; with average values of 233, 334 and 476 TWh, respectively.

Kavaklıoğlu et al., (2009) used ANN for predicting electricity consumption of Turkey. They mentioned that is better
to use models that can handle nonlinearities among variables as the expected nature of the energy consumption
data is nonlinear. It is a well-known fact that ANN can model any nonlinear relationship to an arbitrary degree of
accuracy by adjusting the network parameters. Electricity consumption was modeled in their study as a function of
economic indicators such as population, GNP, imports and exports. Imports and exports to Turkey are related to
manufacturing processes and therefore strongly affect electricity consumption. Turkey’s electricity consumption
was predicted to reach 240 TWh in 2020 and 279 TWh by the year 2027. From year 2010 to 2027, a slower pace
growth is predicted with an average 5.6 billion kWh increase year over year. In essence, the neural network model
indicates that the fast growth is not sustainable. They mentioned Murat and Ceylan(2006) obtained that modeling
the energy consumption may be carried out with ANN with a lack of future prediction since ANN models they built
are good at solving existing data, but are not good for prediction since they do not use any mathematical models.

Toksarı (2009) has studied models obtained by using the ant colony optimization (ACO) based algorithm to
forecast net electrical energy generation and demand. ACO is a multi - agent system in which the behavior of each
unit is inspired by the foraging behavior of real ants to solve optimization problem. The estimation of electrical
energy generation and demand based on economic indicators was to be modeled by using various forms of
equations. These forms are linear and quadratic. The economic indicators that were used during the model
development are GDP, population, import and export data. Ant colony optimization electricity energy generation
and demand (ACOEEGE and ACOEEDE) were developed inspiring from ACOEDE algorithm used to estimate
energy demand (Toksarı, 2007). Both models are developed by using the ACO-based algorithm and observed data
between 1979 and 2006. The models are developed by using proposed ant colony optimization electricity energy
estimation (ACOEEE) algorithm. These are linear_ACOEEGE and linear_ACOEEDE, which is the linear form of
the ACOEEE, quadratic_ACOEEGE and quadratic_ACOEEDE which is a quadratic form of the ACOEEE.
Quadratic models for both generation and demand provided the better fit solution due to the fluctuations of the
economic indicators. The ACOEEGE and ACOEEDE models were used under three scenarios with different
assumptions of the variables for estimating Turkey’s electricity generation and demand in the years 2007–2025.
Electricity generation and demand estimations were thought to be separate and varied according to the forms of
the equations. Consequently the electricity demand estimates varied from 125 TWh to 2,500 TWh in 2025 for three
scenarios.

437
In Küçükali and Barış’s study (2010) fuzzy logic electricity demand model was developed in Turkey. Unlike most of
the other forecast models about Turkey’s electricity demand, which usually uses more than one parameter, GDP
based on purchasing power parity was the only parameter used in the model. In most of the previous studies
(Hamzacebi, 2007; Erdogdu, 2007) the net electricity consumption of Turkey was predicted. MENR, Küçükali and
Barış’s (2010) study have tried to forecast gross electricity consumption of Turkey. According to the study fuzzy
model follows the general trend in the scatter more closely than the regression solutions. This is because internal
uncertainties and system behavior at different time periods are not taken into account in the classical regression
approach. The advantage of the fuzzy logic procedure lies in the ability to mimic the human thinking and
reasoning. The results indicate that the electricity demand growth mainly reflects the changing expectations of
GDP. So, forecasting the Turkey’s electricity consumption with the country’s economic performance in the short-
term will be more meaningful and it would provide more reliable data for the policy makers and investors. Turkey’s
electricity demand was predicted to be 230 TWh in 2014.

Çunkaş and Taşkiran (2011) have carried out Turkey’s electric consumption forecasting by using Genetic
Programming (GP). Unlike GA, GP is a symbolic regression tool that performs well on assigned tasks. GP does
not need any functional relationship between dependent and independent variables (Lee et al., 1997, Çunkaş and
Taşkiran 2011). In general, Turkey’s electricity consumption can be characterized as both unstable and fluctuating.
Therefore, the GP method is a good forecasting tool for systems which are complex, uncertain, and unstable. The
proposed model effectively predicts the future electricity demand through GP using annual data for the previous
years. By using the available data, finding the function which represents the data well is called curve fitting. There
are many different curve fitting techniques like interpolation, regression, and symbolic regression. Conventional
regression techniques are used to find the function and the coefficients best representing the data. Most of the
time, it is very hard to find the best fit function. On the other hand, symbolic regression does not look for function
coefficients; instead, it looks for the best fitting structure that represents the model. GP is known as the best
symbolic regression tool (Koza, 1992; Çunkaş and Taşkiran 2011). According to GP results, the electricity
consumption is predicted to be 315 TWh in 2020.

In the study of Kavaklioglu (2011) Support Vector Regression (SVR) methodology is used to model and predict the
Turkey’s electricity consumption. Electricity consumption is modeled as a function of time, socioeconomic
indicators such as population, GNP, imports and exports. An analysis of historical data for these four variables has
shown that except for the population variance, there are no steady trends within these indicators. In order to
facilitate future predictions of electricity consumption, a separate SVR model was created for each of the input
variables using their current and past values; and these models were combined to yield consumption prediction
values. This is probably due to the fact that Turkey is a developing country and it is very vulnerable to global
influences as well as it has a highly dynamic internal structure. In addition, these four variables do not constitute a
full list of all the variables that affect electricity consumption although they are responsible for the majority of the
input–output relationship. Based on the optimal consumption model by SVR method, future electricity consumption
of Turkey is predicted until 2026 using data from 1975 to 2006. The model predicts that the consumption will reach
285 TWh in the year 2026.

Dilaver and Hunt (2011a) investigated the relationship between Turkish aggregate electricity consumption, GDP,
average real electricity prices, and an Underlying Energy Demand Trend (UEDT) in order to forecast future Turkish
aggregate electricity demand. Structural Time Series Model (STSM) is the adopted methodology which allows the
estimation of a stochastic UEDT. The STSM decomposes a time series into explanatory variables, a stochastic
trend and an irregular component. As mentioned in Dilaver and Hunt (2011a)’s study arguably, some of the
previous studies failed since the relative big deviations of the actual demand values. A key reason for the failure of
them is to capture adequately the impact of the main economic drivers and the UEDT for Turkish electricity
demand. They argued that their results attempts to rectify these perceived defects by using the STSM to estimate
a Turkish aggregate electricity demand function and using it to produce future forecast scenarios. Turkish
aggregate electricity demand is predicted to be 259, 310 and 368 TWh in 2020 in the low, reference and high case
scenarios respectively according to the forecast scenarios based upon GDP, electricity price, STSM coupled with
the UEDT.

In another study of Dilaver and Hunt (2011b) it is mentioned that technological progress of the capital stock is an
important factor that influences industrial energy demand. Overall based upon the different forecast assumptions,
Turkish industrial electricity demand is predicted to be between 90 and 106 TWh in 2015 and between 97 and 148
TWh in 2020. This is somewhat less than the previous forecasts for Turkish industrial electricity demand; Akay and
Atak (2007) suggested that demand would be 140.4 TWh in 2015 and Hamzacebi (2007) suggested demand
would be 219.2 TWh in 2020 — both of which are somewhat higher than the high case scenario of this study. The
differences in forecasts, it is argued, being primarily due to these other studies neglecting the relationship between
economic variables, underlying trend and electricity consumption.

In the study of Dilaver and Hunt (2011c) Turkish residential electricity demand is predicted to be 48, 64 and 80
TWh in low, reference and high case scenarios, respectively by 2020. They mentioned that although the UEDT
cannot anticipate how future unknown structural changes will affect the evolution of electricity demand—but this is
an issue with all forecasts, whatever methodology is utilized.

438
TEIAS is preparing Turkish electricity demand projection report as an affiliated foundation of MENR. In the last
report it is predicted that the electricity demand will be 398 and 434 TWh in 2020 in low and high scenario
respectively according to MAED model results (TEIAS, 2012b).

In the study of Özer (2012) electricity demand forecast is mainly based on the population, historical consumption,
GDP, value added per activity sector, the energy intensity of the different sectors with sectoral growth rates and
the ratio of each sector in total electricity demand. The gross electricity demand was predicted to be 416 TWh in
2020 and 848 TWh in 2030, whereas net electricity demand was predicted to be 341 TWh in 2020 and 696 TWh in
2030. Electricity demand per capita is calculated as 5.3 MWh in 2024 which is about the average value of EU-27
as 5.7 MWh in 2008 (Eurostat, 2010). Annual average total electricity demand growth rate is predicted 6.8% (from
2006 to 2030) which is also convenient with the average growth rate of 7.3% for the period of 1990-2006. As the
predicted Turkish net electricity demand for the year 2020 341 TWh, it is the range of the results of Dilaver and
Hunt (2011a).
The study results indicate the growth rates for total, industrial and residential electricity demand are different from
each other which are convenient to the historical data. This is also mentioned in the study of Dilaver and Hunt’s
studies (2011a,b,c).

RESULTS

The Turkish electricity demand forecast results of the researches overviewed above are given in Table 1. Most of
the studies have results for the year 2020. The minimum electricity demand prediction is 220 TWh in the study of
Öztürk et al. (2005) as the result of the GA model and the maximum prediction is 514 TWh from the study of Yüksek
(2008) that used MAED model which is convenient to the decision of the most of the studies overviewed here that
MAED model results are higher than the other model results and the actual demand values.

Besides for the year 2020 the results of the studies such as Özer (2012), Dilaver and Hunt (2011), Cunkaş and
Taşkıran (2011), Tunç et al., (2006), and Öztürk et al., (2005) are in the same range as they are 341 TWh, 310 TWh,
315 TWh, 322 TWh and 300 TWh respectively. In most of them the previous data and the energy demand trend were
taken into consideration.

On the other hand Turkish net actual electricity consumption in 2010 and 2011 are 172 TWh and 186 TWh
respectively (TEIAS, 2012a). Kavaklıoğlu (2011) has the best forecast for the year 2010 with 1.163% relative error,
whereas for 2011 Dilaver and Hunt (2011) predicted 185 TWh and Özer (2012) predicted 184 TWh with 0.5% and
1.075% relative errors respectively. Besides Turkish actual gross electricity demand values are 211 TWh and 229
TWh in 2010 and 2011 respectively. Cunkaş and Taşkiran (2011) have forecast values as 216 TWh in 2010 and 227
TWh in 2011 (with 2.37% and 0.87% relative errors).

Turkey’s electricity consumption can be characterized as both unstable and fluctuating. According to the results of
the studies in the literature overviewed here the non-linear forms of the equations can better estimate the future
electricity demand of Turkey due to the fluctuations of socioeconomic indicators. In the studies overviewed here
population, GDP, import and export data are the most used variables whereas per capita consumption, increase
ratios, income elasticities, prices, previous data, and sectoral demand trend are the less used variables for the
forecast results.

DISCUSSION

Turkey has an annual electricity growth rate about 7%. This will continue until somewhat as the consumption per
capita reaches a convenient level for Turkey. Then it will be more stable and can be expected to have a more
similar growth rate of population and GDP. Besides sectoral electricity demand projections are also important in
the case of aggregate electricity projections.

According to the results of the several numbers of studies in the literature about electricity consumption forecasting
and modeling for Turkey, the electricity demand projections differ from each other with varying methodologies
based on the different assumptions of the variables. All the researchers mentioned that the results of their own
studies give better results than the results of MAED used by MENR, as the absolute relative errors of the other
model results for the actual values were less than that of the MAED results. All the studies overviewed in this study
indicate that electricity demand growth mainly reflects the changing expectations of population and GDP. However
Turkish electricity demand growth rate was higher than GDP growth rate according to historical data. This indicates
there are some other factors affecting electricity demand. These can be prices, GDP per capita, imports and
exports, sectoral demand trends etc. The results of different model studies used various methodologies and
variables are important for policy makers to plan future energy investments.

As still being a developing country in Turkey, aforementioned indicators are not stable. Thus it is rather difficult to
handle the uncertain factors such as socioeconomic indicators for the electricity demand projections. Prediction of
these indicators is another significant issue which has to be studied with them for electricity demand projections.

439
Table 1: Summary of Turkish electricity demand studies
TWh Projection year TWh Variables/Assumptions Methodology

Yumurtacı and Asmaz 2010 ~195 Population and energy Increase rates
(2004) 2011 ~198 consumption per capita
2050 1,173
Özturk et al., (2005) 2010 150-180 Population, GNP, import GA
2011 170-190 and export
2020 220- 300
Ozturk and Ceylan 2010 270-290 Population, GNP, import GAEDquad
(2005) 2011 290-310 and export
2020 462-500
Tunç et al., (2006) 2010 225 Regression
2011 270 analysis, linear
2020 322 mathematical
optimization
model
Erdoğdu (2007) 2010 155.6 Income elasticities, Cointegration
2011 156 population, analysis and
2014 160 real electricity prices, real ARIMA
GDP per capita and net
electricity consumption per
capita
Hamzaçebi (2007) 2010 207 ANN, time
2011 226 series structure
2020 500
Akay and Atak (2007) 2010 180 Previous consumption data GPRM
2011 200
2015 266
Yüksek (2008) 2010 222-242 Population, electric energy MAED and
(233) consumption per capita, increase ratios
2015 302-356 total electric energy
(334) consumptions
2020 440-514
(476)
Kavaklıoğlu et al., 2010 183 Population, GNP, imports ANN
(2009) 2011 189 and exports
2015 212
2020 240
2027 279
Toksarı (2009) 2010 175-190 GDP, population,import ACO
2011 160-200 and export
2025 125-2500
Küçükali and Barış 2010 193 GDP Fuzzy logic
(2010) 2011 206 method
2014 230
Cunkaş and Taşkıran 2010 216 Population, annual data of GP
(2011) 2011 227 the previous years.
2020 315
Kavaklıoğlu (2011) 2010 170 Population, GNP, imports SVR
2011 175 and exports
2026 285
Dilaver and Hunt 2010 175 GDP, average real STSM
(2011a) 2011 185 electricity prices, and an
2020 259-310-368 UEDT

TEIAS (2012b) 2010 - Macroeconomic targets, MAED

2011 - demand trend forecast
2020 398 434
Özer (2012) 2010 - Population, GDP, historical Time series
2011 184 data, value added per approach
2020 341 sector, the energy intensity
2030 696 of sectors, sectoral demand
trend, the ratios of the
sectors in total demand

440
REFERENCES

Akay D. and Atak M. 2007. Grey prediction with rolling mechanism for electricity demand forecasting of Turkey.
Energy 32:1670–1675.
Balat, M. 2009. Electricity consumption and economic growth in Turkey: A case study, Energy Sources Part B,
4:155-165.
Bianco V., Manca O., Nardini S. 2009. Electricity forecasting regression models. Energy 34:1413-1421.
Çunkaş M. and Taşkiran U. 2011. Turkey's Electricity Consumption Forecasting Using Genetic Programming.
Energy Sources, Part B: Economics, Planning and Policy 6:4, 406-416.
Dilaver, Z. and Hunt, L.C. 2011a. Turkish aggregate electricity demand: An outlook to 2020. Energy 36: 6686-
6696.
Dilaver, Z. and Hunt, L.C. 2011b. Industrial electricity demand for Turkey: A structural time series analysis. Energy
Economics 33: 426-436.
Dilaver, Z. and Hunt, L.C. 2011c. Modelling and forecasting Turkish residential electricity demand. Energy Policy
doi:10.1016/j.enpol.2011.02.059.
DPT 2010. Economic and Social Indicators (1950-2010), National income and production Table1-24: By 1987
Prices by Sectors GDP (1968-2006), State Planning Organisation, Ankara
Ediger, V.S. and Tatlidil, H., 2002. Forecasting the primary energy demand in Turkey and analysis of cyclic
patterns. Energy Conversion and Management 43, 473–487.
Egelioğlu F., Mohamad A.A., Guven B. 2001. Economic variables and electricity consumption in Northern Cyprus.
Energy 26:355–362.
Erdoğdu E. 2007. Electricity demand analysis using cointegration and ARIMA modelling: A case study of Turkey.
Energy Policy 35:1129-1146.
Eurostat. Energy, yearly statistics 2008; 2010. Luxembourg.
Hamzaçebi, C. 2007. Forecasting of Turkey’s net electricity energy consumption on sectoral basis. Energy Policy
35:2009–16.
Kavaklıoğlu K., Ceylan H., Ozturk H.K., Canyurt O.E., 2009. Modeling and prediction of Turkey’s electricity
consumption using Artificial Neural Networks.Energy Conversion and Management 50: 2719–2727.
Kavaklıoğlu K. 2011. Modeling and prediction of Turkey’s electricity consumption using Support Vector
Regression. Applied Energy 88: 368–375.
Koza, J. R. 1992. Genetic Programming: On the Programming of Computers by Means of Natural Selection.
Cambridge, MA: MIT Press.
Küçükali, S. and Barış, K. 2010. Turkey’s short-term gross annual electricity demand forecast by fuzzy logic
approach. Energy Policy 38: 2438-2445.
Lee, D.G., Lee, B. W., and Chang, S. H. 1997. Genetic programming forecasting of electric power demand. Electr.
Pow. Syst. Res. 40:7–22.
Lippmann, R.P.1987. An introduction to computing with neural nets. IEEE ASSP Mag 4(4):4–22.
Murat Y.S. and Ceylan H. 2006. Use of artificial neural networks for transport energy demand modeling. Energy
Policy 34:3165–72. )
Özer, B. 2012. The scenario analysis of CO2 emission mitigation potential in the Turkish electricity sector. PhD.
Dissertation, Istanbul Technical University, Graduate School of Science Engineering and Technology,
Istanbul, Turkey.
Öztürk, H.K., and Ceylan H. 2005. Forecasting total and industrial sector electricity demand based on genetic
algorithm approach: Turkey case study. International Journal of Energy Research 29:829-40.
Öztürk, H.K., Ceylan, H., Canyurt O.E., Hepbaşlı, A. 2005. Electricity estimation using genetic algorithm approach:
a case study of Turkey. Energy 30: 1003-1012.
TEIAS Turkish Electricity Generation-Transmission Statistics, 2012a. <http://www.teias.gov.tr/Türkiye Elektrik
İstatistikleri/ (accessed 25.12.2012).
TEIAS (Turkish Electricity Transmission Company). 2012b. Turkish electrical energy 10 year generation capacity
projection 2012-2021.
Toksarı, M.D. 2007. Ant colony optimization to estimate energy demand of Turkey. Energy Policy 35: 3984–3990
Toksarı, M.D. 2009. Estimating the net electricity energy generation and demand using the ant colony optimization
approach: Case of Turkey, Energy Policy 37: 1181-1187.
Tunç, M., Çamdalı Ü., Parmaksızoğlu, C. 2006. Comparison of Turkey’s electrical energy consumption and
production with some European countries and optimization of future electrical power supply investments in
Turkey. Energy Policy 34: 50-59.
Yumurtacı, Z. and Asmaz, E. 2004. Electric Energy Demand of Turkey for the year 2050. Energy Sources
26:1157-1164.
Yüksek, Ö. 2008. Reevaluation of Turkey’s hydropower potential and electric energy demand.Energy Policy
36:3374-3382.

441
 EXPERIMENTAL STUDY OF THIN LAYER DRYING BEHAVIOR OF A FISH ( MYKISS)

M. Erdem , Y. Varol , H. F. Öztop and F. Kamışlı

1* 2 3 4

1
Fırat University Vocational School of Technical Sciences, Elazig, Turkey
Fırat University Technology Faculty Department of Automotive Engineering, Elazig, Turkey
2

Fırat University Technology Faculty Department of Mechanical Engineering, Elazig, Turkey

3

Fırat University Engineering Faculty Department of Chemical Engineering, Elazig, Turkey

4

e-mail: m.erdem82@gmail.com, yvarol@gmail.com, hfoztop1@gmail.com and fkamisli@firat.edu.tr

ABSTRACT
This study covers the drying behavior of rainbow trout (Oncorhynchus mykiss) in a square cross sectioned
tray dryer. For this purposes, approximately 200g fish was used to dry at different air inlet velocities as 0.5,
1.5 and 2.5 m/s for the constant air inlet temperature at 46 °C in ranges of 28-43 percent of relative humidity.
In the experiments, interior temperature values and moisture loses of fish were measured as a function of
the drying time. It is seen that the moisture content of the fish has decreased with time and drying procedure
for inlet air velocity to be effective. And it is suggested that the samples should be dried as single layer at
o -1
46 C±0.1 and velocity of 2.5ms for optimal boundary conductions.

Keywords: Drying, Fish drying, Food drying, Tray dryer

1. INRODUCTION
The drying of nutrients in agriculture and animal science is a method, which has been known since ancient
times. The method can be applied many kinds of food. The process of fish drying is one of the most reliable
methods that is used to storage fish as others foods for a long time. One of the drying techniques applied in
the world is the open sun techniques. But nutrients dried with this technique don’t protect from harmful
matters as dust and rain, insects. Therefore, drying of fish and other nutritional products in closed places is
healthier than open sun. In this way, control of drying process brings benefit on behalf of healthier for human
life. In this regard, works of some researches can be listed as follows: Walnut, pine, beech and poplar which
are commonly used in wood material industry in Turkey by Kamisli (2011) were dried at constant temperature
but different velocities of drying air in the tray dryer. It was observed that the drying rates and heat transfer
coefficients decrease with drying time. Bala et al. (2003) examined the drying of pine apple by using a solar
tunnel dryer. They indicated that the pine apple drying in the solar dryer is a good quality dried product for
human consumption. Bala and Mondol (2001) investigated experimentally drying of fish using solar tunnel
dryer. They found that moisture content decreases with drying time and the fish dried in the solar tunnel
dryer was completely protected from rain, insect, dust and that the dried fish was a high quality product.
Bellagha et al., (2007) studied on salting and drying of sardine (sardinella aurita). Erdem, (2011) investigated
experimentally behavior the drying of fish (Oncorhynchus mykiss) in a tray dryer under different flow velocity
and air inlet temperature at different dimension of the fish, and drying of salted fish at different temperature
and at different air flow velocity. It observed that moisture content of either case (both salted fish and
unsalted fish) dwindle with drying time. Jain and Pathare (2007) investigated the drying kinetics of open sun
drying of fish. The fishes used in the experiment are prawn and chelwa fish (Indian minor carp). They
observed that the drying rate curves contained no constant and showed linear falling rate throughout the
drying process. Kituu et al. (2010) developed a drying model in Visual Basic 6 for simulating the drying of
Tilapia fish (Oreochromis niloticus) in a solar tunnel dryer. They showed that the model can be used to
predict the drying of Tilapia fish in a solar tunnel dryer. Oztop and Akpinar (2008) examined the moisture
transfer during the convective drying of the some products experimentally and numerically. Both
experimental and numerically results showed that there is a good agreement between numeric and
experimentally measured data and predicted values. Erdem (2013) investigated experimentally the behavior
of drying of rainbow trout at different inlet temperatures but constant inlet velocity in a tray square cross
cross-sectioned tray dryer. It observed that the curves decreased with drying time during the drying process.

442
Arason and Arnason (1992) carried out a study on drying of fish products using geothermal energy. They
suggested that the equipment designed for drying of fish can be also used to dry the other industrial
products. Koyuncu (2006) actualized the performance of improvement of greenhouse-type agricultural
dryers. It observed that systems improved are more efficient than open air dryer.

The main purpose of this work is to examine the thin layer drying behavior of fish (Oncorhynchus mykiss) in
a tray dryer under different air flow velocities and at a constant inlet temperature. Because there is no
sufficient data related to drying behaviour of the rainbow trout in literature, the results obtained from the
experiments are very important for the future studies.

2. MATERIALS AND METHODS

2.1. Experimental setup and method
Figure 1 shows the dryer set-up used in this experimental work. It mainly consists of an electric fan with 375
W, heater resistance, drying chamber, air duct and measurement instruments.
The air flow rate of the experimental setup was adjusted by the fan speed control unit. The heating system
consists of an electric heater (maximum power of 800 W), which is placed into the duct. The drying chamber
temperature was adjusted by the heater power control. The drying duct was constructed from st-52 steel
sheet as a square section in 2000 mm length, 300 mm width, 300 mm height and length of drying chamber
also as 590 mm. In the measurements of temperature, three numbers of K type thermocouples were used
with a manually controlled two-channel digital thermometer (CHY, 806AW, accuracy of ±0.1 °C). The velocity
of air passing through the system was measured by a 0.4-3 m/s range anemometer (RAM DT-619, m/s -
range: 0.4 ~ 3.0 accuracy: 3%). Moisture loss was recorded by a digital balance (Avery Berkel, Model
CC061) at time interval of 10 min. during drying period to determine a removal rate of the moisture. The
digital balance has the measurement range between 0 and 6000g with the accuracy of ±0.1 g. For thin layer
drying experiments, the fish samples were selected carefully cleaned, washed and rested for a while to
remove the water on the samples.

443
 a)

b)
Fig.1. Experimental set up; a) the picture of experimental set-up, b) the schematic picture of experimental set-up.

2.2. Experimental uncertainty

During experiments some errors take place because of measuring instruments, unknown causes and
electronic oscillations. These errors are determined with following equation (Holman, 1994);

2 2 2 1/2
W x = [(x1) + (x2) + ……………… + (x∞) ] (1)

Total error values of measuring instruments are illustrated in Table 1:

Table1. Uncertainties of the measuring parameters.

Parameter Unit Total error

Temperature measurement °C ±0.07

Air inlet velocity m/s ±0.025

Mass measurement g ±0.1

Moisture loss values min ±0.17

Dimensionless moisture content: min ±0.2

Drying rate g/s ±0.14

444
3. MATHEMATICAL FORMULATION

Dimensionless moisture content is determined by following equation (Afzal and Abe, 2000);

Mt − Me
M = (2)
M −M
i e

In this equation, M, explains dimensionless weight losses of the fish during drying, Mi, weight of the fish at
beginning, M t , weight of the fish at time of t, M e , weight of the fish in equilibrium (final weight). Moisture
content wet basic (w.b.) is calculated by following equation (Gokoglu, 2002);

Mw
M wb = (3)
M
i

Where, Mwb refers to the moisture content of fish and Mw, the amount of water at time of t (g) for fish
samples. The following equation is used to calculate drying rate (Akpinar and Bicer, 2003);

dM M t + dt − M t
= (4)
dt dt
In above equation, dM/dt is the drying rate and Mt, the weight of fish at time of t, Mt+dt, the weight of fish at
time of t+dt.

4. RESULTS AND DISCUSSION

In this presentation, the effect of inlet air velocity on drying of a fish (Oncorhynchus mykiss) was determined
in the tray dryer. In the experiments, inlet air temperature and air inlet velocities was selected as 46 °C and
0.5, 1.5, 2.5 m/s, respectively. Because there is no sufficient data related to drying behaviour of the rainbow
trout in literature, the results obtained from the experiments are very important for the future studies.

Fig. 2 refers to the variation of the interior temperature with drying time of fish at different velocities but
constant temperature of T = 46 °C in the tray dryer. As shown in the figure the interior temperature of fish as
a function time increases at all velocities after first 15 hours. The curve of interior temperature of the fish
sample at air velocity of 0.5 m/s and 1.5 m/s was broken due to adverse working conditions. Next day (after
7 hours) is continued to working where it left off. As can be seen in the figure the constant air velocity of 2.5
m/s the curve of sample is more irregular according to others curve. Figure 3 displays the variation of
moisture content with time of drying fish at three different velocities but constant temperature (T = 46 °C).
After first 15 hours the curve of interior temperature of the fish sample at air velocity of 0.5 m/s and 1.5 m/s is
broken, since drying is broken for seven hours (night time), and then is continued to working. As expected
the moisture content of the sample decreases with time at each air velocity used in the experiments. As seen
in the figure increasing velocity of drying air has an important positive effect on drying time. In other words,
when air inlet velocity increases, drying time decreases. At all studied parameters are seen to be the best
model the sample dried at V = 2.5 m/s and T = 46 °C.

445
 45

40
Temperature (°C)

35

30

25 V = 0.5 m/s
V = 1.5 m/s
V = 2.5 m/s

20
0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time (m)

Fig.2. Variation of interior temperature of fish with drying time at T = 46 °C and different inlet air velocities.

0.8
V = 0.5 m/s
V = 1.5 m/s
0.7
V = 2.5 m/s

0.6
Moisture content (w.b.)

0.5

0.4

0.3

0.2

0.1
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Time (m)

Fig.3. Variation of moisture content of fish with drying time at T = 46 °C and different inlet air velocities.

Variation of the dimensionless moisture content with drying time for the air temperature of 46 °C and at
different air velocities is given in Figure 4. A case as Fig.2 is seen. Where, how to be effective of velocity on
drying time is showed clearly. In the figure, it is observed that moisture loss decreases with time. At study
except median was worked with the time interval 7 hours, and then was continued to working. Variation of
drying rate with drying time at the air temperature of 46 °C is illustrated in Figure 5 at different inlet velocities.
As seen in the figure up to first 400 minutes the drying rate decreases very rapidly, but then this rapid
decreasing varies less, and then almost constants. It is observed that this case be valid for all velocities. No

446
change appeared on the surface of the samples in these velocities. And Among the drying experiments of
the rainbow trout, the best results were obtained in 2.5m/s.

1.2
V = 0.5 m/s
V = 1.5 m/s
1.0 V = 2.5 m/s
Dimensionless Moisture Content

0.8

0.6

0.4

0.2

0.0
0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time (m)

Fig.4. Variation of dimensionless moisture content of fish with time at T = 46 °C and different inlet air velocities.

0.008

V = 0.5 m/s
V = 1.5 m/s
V = 2.5 m/s
0.006
Drying rate (g/s)

0.004

0.002

0.000
0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time (m)

Fig.5. Variation of drying rate of fish with drying time at T = 46 °C and different inlet air velocities.

5. CONCLUSION

In this study, behaviors of drying of the rainbow trout samples were experimentally investigated in a square
cross-sectioned type tray dryer. In the study, curves of interior temperature, moisture content, dimension
moisture content and drying rate were analyzed. It saw that while increasing with drying time the interior

447
temperature of fish, moisture content and drying rate curves decrease with time. It is concluded that drying
procedure for inlet air velocity to be effective.

NOMENCLATURE

dM/dt Drying rate, g/s

t Time, min
T Air inlet temperature, °C
M Dimensionless moisture weight

M Final weight (equilibrium weight), g

e
M First weight, g
i
Mt The weight at t, g

Mt+dt The weight at t+dt, g

Mw The amount of water, g
Mwb Moisture content, g/g
V The air inlet velocity, m/s
Wx Total uncertainty in measurement

REFERENCES

Afzal, T.M. and Abe, T. 2000. Simulation of moisture changes in barley during far infrared
radiation drying. Computers and Electronics in Agriculture 26:137-145.

Akpinar, E.K.; Bicer, Y. 2003. Investigation of drying behaviour of pumpkin in a cyclone type dryer. Gazi
University Journal of Science 16 1: 159-169.

Arason, S. and Arnason, H. 1992. Utilization of geothermal energy for drying fish products. Geothermics
21(5/6): 745-57.

Bala, B.K. and Mondol, M.R.A. Experimental investigation on solar drying of fish using solar tunnel dryer.
Drying Technology 19: 427-436.

Bala, B.K., Mondol, M.R.A., Biswas B.K., Das Chowdury, B.L. and Janjai, S. 2003. Solar drying of pineapple
using solar tunnel drier. Renewable Energy 28:183-190.

Bellagha, S., Sahli, A., Farhat, A., Kechaou, N. and Glenza, A. 2007. Studies on salting and drying of sardine
(Sardinella aurita): Experimental kinetics and modelling. Journal of Food Engineering 78: 947-952.

Erdem, M. 2011. Experimental investigation of fish drying in a tray drier. MSc Thesis, Firat University,
Graduate School of Natural and Applied Sciences, Elazig.

Erdem, M. 2013. the effect of inlet air temperature on drying of rainbow trout dried in a tray drier. 2.Anadolu
Enerji Sempozyumu, Diayrbakır, Turkey.
Gokoglu, N. 2002. Su ürünleri işleme teknolojisi. Su ürünleri temel yayınları, Faculty of Agricultural,
University of Akdeniz, Istanbul.

448
Holman, J. 1994. Experimental methods for engineers, second edition New York: McGraw-Hill.

Jain, D. and Pathare P.B. 2007. Study the drying kinetics of open sun drying of fish. Journal of Food
Engineering 78: 1315-1319.

Kamisli, F. Dynamics of wood drying and determination of heat transfer coefficients. 2011. Turkish Journal of
Science & Technology 6 (1): 35-45.

Kituu, G.M., Shitanda, M., Kanali, C.L., Mailutha, J.T., Nijoroge, C.G., Wainaina, J.K. and Silayo, V.K..
2010.Thin layer drying model for simulating the drying of Tilapia fish (Oreochromis niloticus) in a
solar tunnel dryer. Journal of Food Engineering 98: 325-331.

Koyuncu T. 2006. An investigation on the performance improvement of greenhouse-type agricultural dryers.

Renewable Energy 31:1055-71.

Oztop, H.F. and Akpinar, E.K. 2008. Numerical and experimental analysis of moisture transfer for convective
drying of some products. International Communications in Heat and Mass Transfer 35: 169-177.

449
 THE USE OF TRANSPARENT INSULATION MATERIALS AS ONE OF THE MEASURES OF
IMPROVING THE ENERGY OF THE STRUCTURES

M. Miletic
Student of Phd Studies, Architectural Faculty, Belgrade University
p_mirjana@yahoo.com

ABSTRACT
Structures can be repaired in several ways: physically, economically or in terms of their energy consumption.
When talking about energy efficiency, buildings are certainly repaired in terms of their energy saving
capacity, provided that there is a need for it. The energy saving aspect of reparation is reflected in the
rationalization of their energy consumption, preservation of their engineering substance, improving their
thermal comfort, as well as in the use of renewable energy sources. In terms of thermal comfort, one of the
traditional building reparation methods is improving the thermal properties of sheathing by using transparent
insulation materials. Depending on the purpose of structure, it is often necessary to use transparent
insulation. Heat flows through these materials that are part of the structure's thermal mass are calculated
specifically according to regulations on energy efficiency of structures applying in our country. These thermal
insulations are belonging to the group of newer materials. This paper provides an overview of transparent
insulation materials, along with the opportunities they are offering regarding the improvements of energy
efficiency of heating, i.e. improvements in thermal comfort of buildings, with the minimum use of primary
energy.

Keywords: energy efficiency, heating, transparent insulation materials

INTRODUCTION
Buildings, as derived based on the consumption of other resources of the earth are the only continuously
growing resources. Previously, in the developed EU countries, buildings accounted for 50% of energy
consumption, while 25% of energy was consumed by the traffic; today, 40% of the energy is spent on
mechanical energy, 2% on lighting, and 33% on room heating. When talking about energy efficiency,
buildings are certainly being repaired in terms of their energy saving capacity, provided that there is a need
for it. The energy saving aspect of reparation is reflected in the rationalization of their energy consumption,
preservation of their engineering substance, improving their thermal comfort, as well as in the use of
renewable energy sources. The energy aspect of reparation is closely related to the spatial and economic
aspects. In addition to the need for improving the buildings in terms of their energy-efficiency, it is necessary
to expand the current space floor areas in homes, or at least to keep the existing ones. Thermal comfort is
being achieved by using proper thermal insulation materials. In Serbia, conventional materials are commonly
used, i.e. polystyrene and Styrodur. However, their thickness to which they are traditionally applied neither
can meet the required criteria nor the conditions for achieving the specific levels of energy classes. By
adopting the Regulation on Energy Efficiency ("Official Gazette of the Republic of Serbia" no. 061/2011 of
19.08.2011.) and the forthcoming Law on Energy Efficiency in Buildings, the situation in the near future will
certainly be changed also regarding the use of materials, particularly in the facade assembly as part of the
building's thermal envelope.

This paper provides a description of a typical residential building (GF+0), along with the materials applied
and thickness of the thermal envelope layers commonly used in Serbia. The aim is to present a building
where the standard thicknesses of thermal insulation fails to meet the minimum criteria to be achieved as
prescribed by the Regulation on Energy Efficiency in Buildings in our country (class C), which would mean
that the annual energy use for heating purposes should not exceed QH,nd ≤ 65 kWh/m a.
2

Regarding the layers of thermal envelope, as applied based on the calculations, the present building falls
into class D. By increasing the thickness of thermal insulation in order to achieve the desired criteria, the
layers' thickness will grow to 25 cm, reducing thereby the residential space and making the possibility of
repair questionable. Therefore, innovative approaches are being proposed, i.e. using new insulating
materials such as lightweight construction materials and innovative insulation materials: Neopor, aerogels
and coatings. The use of transparent insulations, belonging to the group of new materials, is also
recommended. When it comes to transparent thermal insulation, according to the Regulation on Energy

450
Efficiency, heat flows through these materials are calculated separately (Appendix 6, "Official Gazette of the
Republic of Serbia" no. 061/2011), which is not discussed here.

IMPROVING THE THERMAL PROPERTIES OF ELEMENTS OF THE THERMAL ENVELOPE

Improving the thermal properties of elements of the thermal envelope requires treating the thermal layers of
facade walls, basement structures, ceilings above the open passages, roof structures, windows and window
protection elements, i.e. the transparent and non-transparent positions of the external envelope.

The underlying building is a family house of a surface area of A = 103.82 m² in Belgrade (Fig. 1).

Fig 1. Ground floor and south facade of residential house: Elaborate for energy efficiency was done based on Methodology on heat protection and
heat losses. Appendix 6. Article 3.1., "Official Gazette of the Republic of Serbia" no. 061/2011. Author: Mirjana Miletic, November 2012.

Architectural requirements for achieving energy efficiency in buildings

Buildings can be rehabilitated and enhanced with respect of their energy efficiency. One of the traditional
methods of doing this is improving the thermal properties of the thermal sheathing elements. The term of
"thermal sheathing" refers to structural elements by which the heated part of the building is separated from
its unheated part, i.e. parts of the building with different conditions of comfort (Regulations on energy
efficiency in buildings, "Official Gazette of the Republic of Serbia", no. 61/2012, Art. 2). The sheathing
elements consist of facade walls, basement constructions, ceilings and roof structures. One of the measures
of reducing the energy consumption in buildings is to insulate the external walls and eliminate thermal
bridges.

THERMAL SHEATHING

Achieving the appropriate heat transfer coefficients

The structure's thermal sheathing refers to its elements in contact with the outside environment – these are
the elements by which the heated and the unheated spaces are separated. When speaking about the façade
of the mentioned building, we need to refer to its transparent and non-transparent positions as parts of the
thermal sheathing. Transparent positions are the windows, patio doors, front doors and storefronts, while the
exterior walls are considered to be the non-transparent part of the facade assembly. The transparent
positions account for the highest percentage of heat energy being lost, while 40% of energy is being lost
through standard facade doors and windows; therefore, they need to receive a special attention.

Transparent positions of the facade assembly

In the context of materials that are important for the thermal insulation of buildings in addition to insulation
materials, glass is treated as a specific class, because its nature makes it a poor thermal insulator. Thus, in
order to achieve adequate thermal protection and energy efficiency in terms of heating, glass has to be
improved.
For heat and light energy transfer, glass needs to meet the following requirements (Table 1):
1. to minimize heat loss (outward heat transfer)

451
 2. to minimize heat gains (inward heat transfer)
3. to provide the optimum amount of light

Table 1. Requirements for glass for heat and light energy transfer
Type of glass Heat transfer coefficient U (W/m2 K) Total sun energy trasfer g Light permeability Т
Duplex thermo insulation glass ≥ 1.1 0.55-0.65 0.8
Triplex thermo insulation glass ≥ 0.5 0.5 0.4-0.7
Sun radiation controlling glass ≥ 1.1 0.5-0.65 0.7-0.8

Innovative glass

For the purpose of reducing the heat transfer coefficient, thermal insulating glasses are being used. These
glasses are made of two or three layers; the space between them is completely filled with dry air or some
gas.

The heat transfer coefficient can be further reduced by mounting a heat-reflecting layer on the outward face
of the inner glass or the inward face of the outer glass. The disadvantage of these glasses is that their use
results in a somewhat reduced light permeability, which is not suitable for working spaces, where this is one
of the essential prerequisites for achieving comfort. (Radivojevic, 2012)

Solar radiation is controlled by adequate sheathings. The use of inorganic substances as a protection
against UV rays is an innovation. Three components are used: titanium dioxide (TiO2 – absorbing the energy
along with the UV rays), zinc oxide (ZnO – protecting against UVB and UVA rays) and ceroxide (CeO –
absorbing the UV rays, in addition to a small amount of light), (Laydecker, 2009)

Non-transparent positions of the façade assembly

The non-transparent positions of thermal sheathing which are parts of the elements in contact with the
external environment are walls, roofs and floor trusses. The external walls are the non-transparent facade
positions.
Consider the layered external wall of an object which is made of gitter block and mortar. Its thickness is 25
2
cm and under the new regulations Umax = 0.4 W/m K. Along with the other measures of improvement, the
thickness of insulation is recommended to be about 15 cm, although usually it is 5 cm. (Radivojevic, Rajičić,
2012)

The hygrothermal properties of external wall of the following facade assembly shоwn in the table below are:
(Table 2)

Table 2. Hygrothermal properties of external wall of the residential house

δ λ ρ
Construction layer
[cm] [W/mK] [kg/m3]
1. composition lime mortar 3 0.85 1700
2. gitter block 25 0.61 1400
3. styrodur 15 0.038 33
4. baumit silicate facade 1 0.70 1600

Then we have the following values for heat transfer resistance (Calculation from the report on a building's
energy efficiency made based on Methodology on heat protection and heat losses. Appendix 6. Article 3.1.,
"Official Gazette of the Republic of Serbia" no. 061/2011. (Table 3.)

Table 3. Heat transfer resistance and temperature values for the layers listed below
λ R ∆θ θ
Description
[W/mK] [(m2⋅K)/W] [oC] [oC]

452
 Inside max20
Transfer 0.13 0.97 19.3
1. composition lime mortar 0,85 0.035 0.26 19.04
2. gitter block 0,61 0.15 1.12 17.92
3. styrodur 0,038 3.95 29.39 -11.47
4. baumit silicate facade 0,70 0.01 0.07 -11.54
Heat transfer 0.04 0.3 -11.84
Outside min-12.1
Total resistance 4.315

Where the surface heat transfer coefficient is as follows:

U =1 / R =1 / 4.315= 0,23

U < Umax

0,23 W/m²K < 0,3 W/m²K (1)

The above equitation Eq. (1) shows that assembly meets the requirements ("Official Gazette of the Republic
of Serbia", Appendix 3, Art. 3.2)

Higrothermal properties for the materials in the layers of the ceiling that is part of the thermal sheating of the
above mentioned house (Table 4.)

Table 4. Hygrothermal properties of ceiling of the residential house

d λ ρ
Construction layer
[cm] [W/mK] [kg/m3]
1.gauged mortar 2 0,87 1800
2. ceiling 16 0,61 1400
3. concrete 4 2,04 2400
4. vapour permeable foil 0,04 0,19 215
5.mineral wool 25 0,041 120
6. vapour permeable foil 0,04 0,19 215

Then we have the following values for heat transfer resistance (Calculation from the report on a building's
energy efficiency made based on Methodology on heat protection and heat losses, Appendix 6. Article 3.1.,
"Official Gazette of the Republic of Serbia" no. 061/2011. (Table 5.)

Table 5. Heat transfer resistance and temperature values for the ceiling layers listed below
R ∆θ θ
Description  [W/mK]
[(m2⋅K)/W] [oC] [oC]
inside 0,100 20
transfer 0,7276 19,272
1.gauged mortar 0,87 0,023 0,167 19,105
2. ceiling 0,61 0,262 1,906 17,199
3. concrete 2,04 0,0196 0,1426 17,056
4. vapour permeable foil 0,19 0,0021 0,01527 17,041
5.mineral wool 0,041 6.098 21,296 -4,254
6. vapour permeable foil 0,19 0,0021 0,01527 -4,269
Heat transfer 0,7276 -4,997
outside 0,100 -5
Ru 6,5

U=1 / Ru = 0,15 W/m²K

453
U < Umax

0,15 W/m²K < 0,3 W/m²K (2)

This calculation Eq. (2) shows that in order to meet the requirements ( to have U less than Umax) we need to
have thermal insulation layer, in this case mineral wool, of 25 cm of thickness.

Another part of thermal sheating of residential house is floor. The layers are listed below.( Table 6.)

Table 6. Hygrothermal properties of floor of the residential house

d λ ρ
Construction layer
[cm] [W/mK] [kg/m3]
inside
transfer
1. ceramic tiles 1 1,28 2300
2. cement mortar 4 1,40 2100
3. vapour permeable foils 0,04 0,19 215
4. mineral wool 14 0,039 180
5.hydroinsulation 1 0,17 1100
6. concrete 10 2,04 2400
7. gravel 10 0,81 1700
transfer
outside
total

Then we have the following values for heat transfer resistance (Calculation from the report on a building's
energy efficiency made based on Methodology on heat protection and heat losses, Appendix 6. Article 3.1.,
"Official Gazette of the Republic of Serbia" no. 061/2011.) (Table 7.)

Table 7. Heat transfer resistance and temperature values for the floor layers listed below
λ R ∆θ θ
Construction layer
[W/mK] [(m2⋅K)/W] [oC] [oC]
inside 0,17 20
transfer 0,590 19,41
1. ceramic tiles 1,28 0,00781 0,027 19,383
2. cement mortar 1,40 0,0285 0,099 19,284
3. vapour permeable foils 0,19 0,0021 0,0073 19,2767
4. mineral wool 0,039 3,5897 12,474 6.803
5.hydroinsulation 0,17 0,0588 0,204 6.5987
6. concrete 2,04 0,049 0,170 6.4287
7. gravel 0,81 0,1234 0,4288 6
transfer 0 6
outside 0,00 6
total Ru=4.029

U=1 / Ru = 0,248

U < Umax

0,248 W/m²K < 0,3 W/m²K (3)

The above equitation Eq.(3) shows that assembly meets the requirements ("Official Gazette of the Republic
of Serbia", Appendix 3, Art. 3.2)

With increased insulation thickness for whole thermal sheating of the treated house- criteria for achieving the
energy efficiency of the building is fulfilled. (Table 8.)

454
Table 8 . The list of the heat transfer coeficient of the theramal sheating of the facade assembly
Fulfilled
U Umax
Part of thermal sheating criteria for
[W/(m2K)] [W/(m2K)]
C class
External walls 0,231 0,3 Yes
Ceiling 0,15 0,3 Yes
Floor 0,248 0,3 Yes

The annual energy required for heating purposes according to SRPS EN ISO 13790 is

QH,nd = 6142.66 kWh/a

corresponding class C.

The insulation thickness of the presented external wall is 15 cm, which is seldomly encountered in today's
objects. With the insulation on the inward face of the wall being set to a minimum of 15 cm (this is the
appropriate ratio if the floor and ceiling insulation are increased to a significant degree, in this case
calculation above shows that it has to be 25 cm of thickness for ceiling and 14 cm insulation thickness for
floor)), the space within the building is highly reduced, while the wall thickness is increased, with deep
openings appearing on the facade. In order to avoid this, thermal insulation which belongs to the group of
new and innovative insulating materials may be a good solution. There are many products that are used to a
varying degree, depending mainly on their price. Here, only some of them are specified.

THERMAL INSULATION

Thermal insulation products that are used in Serbia are mostly made of expanded polystyrene, styropor, and
extruded polystirol, styrodur. The energy consumed by conventional buildings in heating days amounts 200-
2
300 kWh/m . New buildings constructed based on standards of low energy consumption consume less that
100 кWh/m . This model of constructing enables high levels of residential comfort. Since conventional
2

styropor is not sufficient, there is a need for materials with better technical and thermal insulation properties.
(Polymers in structural engineering, 2010)

Use of transparent insulation materials

Aerogel

Concept and properties of the material

The nickname for aerogel is "frozen smoke" (Juran, 2012). Currently it is the lightest known solid material.
Aerogel was developed in 1931 at Stocktone University, California, in laboratory of professor Stiven Kistler.
"Actually, the structure of aerogel is relatively banal: it is simply an ultralight aerated foam that consists
almost 100% of nothing else than air (the exact figure varies between 95-99,9%)" (Laydecker, 2008, p. 128).
The remaining is a glass-like material – silicon dioxide, known as silica. The extremely high thermal
resistance is conceived by the low thermal conductivity 0,018 W/mK. By using this nanogel, the costs of
heating and cooling can be reduced to a significant degree. The insulation value of Aerogel is extremely
high. It can be used between panels of different materials, including plastics. "The insulation efficiency of a
1
single inch of Aerogel is equal to that of thirty-two layers of glass" (Mori, 2002, p.41). Main characteristics of
aerogel are listed below. (Table 9.)

Table 9. Main characteristics of the material

Thermal
Density
Material Group of Trade name conductivity Transparency Sustainability aspect
g/cm2
materials W/mK

1
2.5cm

455
 Pyrogel
Nanogel Outstanding heat
Aerogel Insulation material Okagel 0,02- 0,5 0,008-0,017 High transparency insulation properties,
Makrolon extremely low weight
Ambient

Use and processing

Aerogels can be used as filtering or insulation materials. For architectural purposes these gels are primary
used as transparent insulation panels (Fig. 2), additives in glass structures, as parts of roof construction or in
double webbed panels. This material is mostly available as granulate and it can be poured and used as high
performance insulation material for injection in narrow cavity layers from 1.5 cm thick. This material is
suitable for the core insulation of dual skin outside masonry and as an insulating material behind klinker
bricks facades. Aerogel has a high level of heat protection even in the case of thin layers. That is why it is
mentioned as one of the measures of improvement of energy efficiency in the context of heating to all kinds
of buildings.

Figure 2. Transparent heat insulation for industrial buildings, aerogel structure. Source: Peters, S. 2011. p.111.

In the mentioned residential building it can be applied as a dual skin masonry or behind gitter block.

Translucent aerogel insulating glass, is to be mentioned in this context, and is intended to be put to energy
efficient use mostly in museums, sports facilities and administration buildings. With nanogel of 60 mm as
2
intermediate layer G value of the material is even less than 0,3 W/m K. This kind of transparent insulation
panel is fulfilling the standard for passive houses.

Aerogel can also be a fill between two layers of polycarbonate panels (PC panels). This material has a high
heat-insulating qualities, high translucence and optimum light diffusion. It is worth to be mentioned in the
relation of heating although is used as roof glazing. These panels have a G value comparable to triple
glazing. (Peters, 2011)

Nansulate Shield

Nansulate Shield is a thin transparent insulation material with the G value many times higher than that of the
current best known insulation. This new product is the first of the Industrial Nanotech's company that is not a
coating, however has the same core nanotechnology as the Nansulate line of thermal insulation and
protective coatings.( Nansulate Shield, 2012). This nano-composed thermal insulation consists of 70%
Hydro-NM-Oxide, 30% of acril and other different additives (Juran, 2012).
This is a coating-like insulation of steady appearance and thickness, with the sole drawback being the
manufacturing costs, which makes its use highly limited. It can be easily applied at the outward surface of the
construction material.

Ceramic foams

Concept and properties of the material

These foams are made of ceramics as silicon carbide. Having in mind its high heat insulating properties
these foams are used widely in building industry especially in heating systems. Ceramic foams can be used
o
in temperatures up to 2000 C and in particular in light weight construction systems. For application in
building industry dual- component ceramic foams were developed. At room temperature these foams harden
within half an hour. Foams are water and steam resistant, fiber free and contain no toxic substances.

456
(Peters, 2011) We are mentioning these foams as one of the possibility to use as transparent insulation
systems.

Table 10. Main characteristics of the material

Sustainability
Material Group of materials Trade name Resistance Transparency
aspect
Duocel SiC Foam Water and steam 100% recyclable
Ceramic foam Insulation material Stelex resistant Optical Air purifying
Dual component Durable at 1000oC Heat insulating

RESULTS AND DISCUSSION

As indicated by the above, this housing unit cannot meet the required criteria in terms of the existing thermal
layer thickness. Achieving class C, i.e. the minimum criteria for the building to become energy-efficient,
requires the thickness of layers to be increased to a substantial degree. The building was initially insulated
using 5 cm for roof and 10 cm for wall insulation (although a 5 cm Styrofoam insulation is commonly used),
making it a class D building. If installing a thicker thermal insulation, consisting in some parts (e.g. in the
ceiling structure) of a 25 cm thick stone wool, we have a problem at hand that we cannot resolve using
current materials and construction techniques. Namely, due to the dimensions of roof wooden beams, this
thick insulation is unsuited for mounting the roof structure. Therefore, new approaches are needed using new
insulating materials. In this case, transparent insulating materials are suggested as the sole possibility. The
use of these materials is limited mainly as a result of their price; nowadays, they are used mostly for non-
residential buildings. They should be surely taken into account given the need for achieving the minimum
criteria when it comes to use of primary energy and buildings of higher energy-efficiency.
Further researches should be taken as calculations of heat transfers for the thermal envelope of the buildings
that have as a thermal layers new thermal insulation materials. These figures, as properties of these new
materials, are not indicated in the new Regulations on Energy Efficiency (Official Gazette of the Republic of
Serbia" no. 061/2011.). It shows that we are at the beginning process in researches of these kinds of
innovations.

CONCLUSION

Some of the measures for achieving the desired comfort and reducing the consumption of primary energy
are aimed at improving the performance of the building's thermal insulation. We have proposed new and
innovative insulating materials. If used properly, these materials can significantly contribute to improving the
energy efficiency of buildings. However, what is important for all materials, including insulation, is the
performance/price ratio. Namely, innovative products, including the new generation of thermal insulation
materials, are characterized by the fact that they can be obtained from a number of different materials, with
the choice ultimately depending on the desired quality and the price. Indeed, the price is the main limiting
factor to using these materials in the construction industry both in Serbia and abroad. Situation in Serbia is
so that these kinds of materials are not included in the Rules on energy efficiency of the structures. It shows
that we are about to work and learn about all of these materials as possibilities and proposals in the sense of
improvement of energy consumption.

NOMENCLATURE
2
A area, m
2
U Heat transfer coefficient W/m K
2
Umax maximum heat transfer coefficient W/m K
resistance of heat transfer coefficient (m ⋅K)/W
2
R
QH,nd annual energy for heating
θ temperature °C
λ thermal conductivity W/mK
g total sun energy trasfer
ρ
3
density, kg/m
T Light permeability

REFERENCES

457
Juran, D., (2012) Nanotechnology in Construction, Zagreb: Faculty of Electrical Engineering. Aerogel,
th
http://www.fer.unizg.hr/_download/repository/djuran_DS.pdf (accessed 20 July, 2012).
Laydecker, S. 2008. Nano materials in Architecture, Interior Architecture and Design, Berlin: Birkhauser
Verlag AG
Mori, T (ed) 2002. Immaterial-Ultramaterial, New York: Harvard Design School in association with George
Braziller, 1: 41.
Nansulate Shield,
th
http://www.industrialnanotech.com/INTK_press_release_11152006.htm, (accesed 25 August 2012.)
Peters, S. 2011. Material Revolution, Sustainable and multipurpose materials for design and architecture,
Munich: Birkhauser
Polymers in structural engineering, Used from http://static.oglasnik.hr/nekretnine/clanak/polimeri-u-
th
graditeljstvu,428 (accessed 09 August, 2010)
Radivojević, А., Rajičić, А. 2012. Architectural Structures, lecturing on Energy efficiency of the buildings,
Belgrade: Chamber of Engineers of Serbia
Regulation on Energy efficiency of the Buildings „Official Gazzete of RS“, no. 61/2011
Regulation on conditions, contents and certification of energy efficiency in buildings. „ Official Gazzete of
Republic of Serbia“ no. 61/2011, from 19.08.2011.
Laydecker, S. 2008. Nano materials in Architecture, Interior Architecture and Design, Berlin: Birkhauser
Verlag AG

458
 EVALUATION OF BIOHYDROGEN PRODUCTION POTENTIAL FROM SEWAGE
SLUDGE

ŞENTÜRK I.,BÜYÜKGÜNGÖR H.

Ondokuz Mayis University Engineering Faculty Environmental Engineering Department, TR 55139 Atakum /
Samsun, TURKEY

e-mails: ilknurg@omu.edu.tr; hbuyukg@omu.edu.tr

ABSTRACT

Owing to limited resources and emission of pollutants, hydrogen is a promising alternative to fossil fuels.
Hydrogen is a clean energy that produces water instead of greenhouse gases when combusted. Furthermore,
hydrogen has a high-energy yield (about 2.75 times that of hydrocarbon fuels). Biological hydrogen production
from biomass is considered one of the most promising alternatives for sustainable green energy production and
important solution to a sustainable power supply and is nowadays being seen as the versatile fuel of the future,
with potential to replace fossil fuels. This process is able to solve two problems: the reduction of pollution from the
uncontrolled degradation of waste and the generation of a clean alternative fuel.
The treatment and disposal of sewage sludge generated in urban wastewater treatment plants is an
important environmental problem. Unused, discarded biomass residues are a potential energy resource, which at
present are not well managed and thus pose significant environmental problems. Recently, some studies are
focusing on using the sludge to produce hydrogen by anaerobic fermentation.
Sewage sludge from a wastewater treatment plant is biomass that contains large quantities of
polysaccharides and proteins and thus is a potential substrate for producing hydrogen. Limited data have
addressed this topic show that the biohydrogen yield using waste sludge and anaerobic fermentation is rather low.
Due to low hydrogen yield from the raw sewage sludge, a number of pretreatment approaches have been
investigated. Pretreatment can increase the efficiency of anaerobic stabilization of sludge by hydrolyzing the
insoluble organic matter to water. The present study deals with the suitability of pretreated sewage sludge as the
primary substrate for microbial H2 production and also may overcome certain important limitations of biological H2
production.

Keywords: Biohydrogen production, sewage sludge, pretreatment

1. INTRODUCTION

Over the past 50 years, the world’s population has more than doubled, coupled with an expectation
of a higher standard of living and an ever increasing economic output this has resulted in a large increase in
primary energy consumption, particularly the use of fossil fuel derived energy. In 2010, world primary energy
consumption grew by 5.6%, the largest percentage growth in almost 40 years. This growth included an
increase in the consumption of all major fossil fuels including oil, natural gas and coal. This trend in
increasing energy consumption is expected to continue as the world’s population is projected to increase by
an additional 1.4 billion people by 2030, and have an increase of 100% of the world’s real income. These
increases will put enormous pressure on the finite supply of fossil fuel based energy, exacerbating global
concerns over energy security, fossil fuel based environmental impacts such as climate change, and the
rising cost of energy and food. The utilization of current energy sources has been generating environmental
pollution of air, water and soil through the years. These negative effects have increased interest in the
development of new technologies to obtain clean energy, mainly through the utilization of renewable energy
sources. Currently, the world consumes about 15 terawatts of energy per year and only 7.8% of this is
derived from renewable energy sources (Senturk and Buyukgungor, 2012 and 2013).
Problems of deficiencies in fossil fuel sources have become apparent in many countries. Moreover,
problems of the greenhouse effect, caused by the enormous combustion of fossil fuels, have become
gradually severe. Therefore, it is necessary to explore clean and renewable energy sources for use in the
near future (Chairattanamanokorn et al., 2009). Hydrogen is a promising alternative energy to fossil fuels. It
is environmental friendly renewable energy in which the by product from its combustion with oxygen is water.
(Lin et al., 2012; Sittijunda et al., 2010; Guo et al., 2010). Hydrogen has high energy yield of 142.35 kJ/g
which is about 2.75 times greater than the hydrocarbon fuel (Guo et al., 2010; Cai et al., 2004).

459
 Hydrogen can be generated by thermochemical, electrochemical or microbial fermentation
(biological) processes. However, the thermochemical process requires hydrocarbon to be used as
feedstocks, which mostly comes from fossil fuels while electrochemical process needs supply of electricity.
Biohydrogen production from organic waste or wastewater through fermentation by anaerobic acidogenic
bacteria with highly diverse fermentation characteristics and hydrogen production capabilities does not
require input of external energy (Lin et al., 2012). Recently, biohydrogen production has received an
increasing attention due to a high demand of sustainable energy. Biologically, hydrogen can be produced by
photo and dark fermentations (Sittijunda and Reungsang, 2012). Hydrogen production through photo
fermentation is not favorable due to the limitations of light conversion efficiencies, photo inhibition at high
solar light intensities and the design of efficient photoreactors. Dark fermentative hydrogen production is a
ubiquitous phenomenon under anoxic or anaerobic condition (Lin et al., 2012). Dark fermentation has
advantages over photo fermentation process in terms of better cost-effective, higher rate of hydrogen
production, and the ability to utilize various kinds of substrates (Sittijunda and Reungsang, 2012).

Dark fermentation under anaerobic conditions can be operated without light and oxygen limits.
Hydrogen production through dark fermentation has advantages over other processes due to its ability of
continuous hydrogen production without an input of external energy and the stabilization of the human-
derived organic wastes and biomass. Among the hydrogen production methods, the most promising,
relatively economical and also environmentally friendly one seems to be the dark fermentation of organic
wastes as it solves the problems of energy production and waste disposal simultaneously
(Chairattanamanokorn et al., 2009; Lin et al., 2012). These human-derived organic wastes for hydrogen
production mainly include milk industry wastewater, lactic acid fermentation plant wastewater, distillery
wastewater and sewage sludge as well as municipal solid waste (Lin et al., 2012). Biomass (waste activated
sludge, plants, animals, and their organic waste products) provides approximately 14% of the world’s energy
needs. In developing countries, which often have limited supplies of fossil fuels but abundant biomass
resources, as much as 35% of the energy demand is supplied by biomass. However, in developed countries
with still abundant fossil fuel supplies, reliance on biomass as an energy resource is very limited, despite its
huge energy potential (Guo et al., 2008).
3
In 2008, about 2.38 billion m /year of municipal and industrial wastewater was treated in wastewater
treatment plants in Turkey, producing about 1.075.000 t/year of sewage sludge (TÜİK, 2008). Landfill is one
of the most common methods that are used in the final disposal of sludge worldwide (Ozsoy et al., 2006).
The disposal of sludge generated in urban wastewater treatment plants is an important environmental
problem. Unused, discarded biomass residues are a potential energy resource, which at present are not well
managed and thus pose significant environmental problems (Guo et al., 2008). Using sewage sludge as the
substrate for fermentative hydrogen production offers several advantages over the use of other biomass
sources. It is available at little or no cost. The supply is plentiful and can be found wherever there are human
settlements (M. Nicolau et al., 2008).
The strict requirement of wastewater effluent standard and a good design for wastewater treatment
process increased the quantity of waste activated sludge (WAS), commonly known as sewage sludge. WAS
is a byproduct of the wastewater treatment process. Sewage sludge is mainly composed of microorganisms,
and its organic substance composition is therefore different from carbohydrate rich substrates such as
glucose or starch. The sewage sludge consists of 41% protein, 25% lipid, 14% carbohydrate, and 20%
unknown components on the basis of chemical oxygen demand (COD) or 32-41% protein and 5-12% fats on
the basis of total dry solids (TDS). Thus, represents a potential substrate for producing hydrogen (Cai et al.,
2004; Guo et al., 2010; Wang et al., 2003). Recently there are interests toward a utilization of activated
sludge as the biological material for a production of renewable energy i.e. hydrogen and methane using
anaerobic digestion process (Sittijunda et al., 2010).
But organics components in sewage sludge are mostly nonutilizable; hence pretreatment becomes
essential to render them suitable for H2 fermentation. Another credible option is co-digestion with other
carbohydrate rich biomass like food waste or molasses (Kotay and Das, 2010). Kim et al. (2004) has
reported higher H2 yield (122.9 mL H2/g carbohydrate-COD) from co-digestion of 87:13 (food waste:sewage
sludge). However, food waste rich in carbohydrates was the major component in that study not sewage
sludge. In this study, it is expected to address the problems associated with waste activated sludge disposal
through simultaneous generation of clean gaseous energy in the form of hydrogen. A synchronous objective
was to investigate the influence of different pretreatment on the hydrogen production using waste sludge by
anaerobic fermentation.
2. APPLICABILITY OF BIOHYDROGEN PRODUCTION FROM SEWAGE SLUDGE

460
 The sludge is composed of largely organic matters (59–88%) that can decompose and produce
offensive odors. These organics are mainly the microbial matters and the microorganisms include hydrogen
producing ones and hydrogen consuming ones. The treatment and disposal of the excess sludge has
become an important problem and a great challenge for many plants. Anaerobic digestion is an appropriate
technique for reduction in the volume and weight of excess sludge before final disposal, and it is employed
worldwide as the oldest and most important process for sludge stabilization. Additionally, anaerobic digestion
can recover partly the bioenergy of sludge through producing methane. Anaerobic digestion process
generally consists of 4 stages i.e. hydrolysis, acidogenesis, acetogenesis and methanogenesis (Figure 1).
Hydrogen is an intermediary metabolite of anaerobic digestion, which is rapidly taken up and converted to
other products by the hydrogen consuming microorganisms in the third stage of anaerobic digestion. On the
other hand, the use of hydrogen is more extensive than that of methane. So, it is beneficial to get hydrogen
in the anaerobic digestion of sludge. In order to harvest hydrogen, the anaerobic digestion of sludge must be
blocked at the hydrogen and acetic acid formation stage, namely, the second stage of anaerobic digestion.

Fig. 1. Anaerobic Metabolism (Diagman and Kim, 2008)

The major reason why sewage sludge has been focused as source of inoculum is that it is bountiful
with enteric bacteria, which are potential H2 producers. Several pure cultures and mixed microbial cultures
have been successfully enriched and tested for H2 producing potential. Nevertheless, the pursuit for ideal
microbe(s) for H2 production has thrust the researchers to screen various sources (Kotay and Das, 2009).
Showed that anaerobic digestion processes using mixed cultures are more practical, because they are
simpler to operate and easier to control, and may have a broader choice of feedstock (Valdez-Vazquez et
al., 2005). On these grounds, nowadays researchers are more interested in using mixed cultures, instead of
pure cultures, for wastewater and waste treatment. In mixed cultures, hydrogen produced by hydrogen
producing bacteria (such as Clostridium sp. and Enterobacter) are consumed immediately by
hydrogenotrophic methanogens (Ray et al., 2008). Thus, in order to produce hydrogen from mixed anaerobic
cultures, hydrogen consuming bacterial activity (methanogens) should be inhibited or eliminated while
preserving the activity of the hydrogen producing bacteria. Many hydrogen producing microorganisms can
form endospores, which can be considered “survival structures” developed by these organisms when
unfavorable environmental conditions are encountered, e.g., high temperature, and harmful chemicals
including acid and alkaline. When favorable conditions return, the spores germinate and become vegetative
cells. However, most of hydrogen consuming microorganisms, e.g., methanogens, do not have this
characteristic (Xiao and Liu, 2009). Several methods have been reported to inhibit methanogenic activity and
selectively enrich hydrogen producing acidogenic bacteria from mixed anaerobic cultures are heat treatment,

461
aeration, acid and base treatment, inhibiting chemical addition such as 2-bromoethanesulfonate (BES),
iodopropane, chloroform, acidogenic culture preparation. The physiological differences between hydrogen
producing bacteria (acidogenic bacteria) and hydrogen consuming bacteria (methanogenic bacteria) form the
fundamental basis behind the development of the various pretreatment methods proposed for preparation of
hydrogen producing seeds. Additionally, the growth rate of most hydrogen producing microorganisms is
faster than that of hydrogen consuming ones.
In the case of utilization of sludge as substrate for H2 production, the pretreatment methods not only
reduced the particle size but also increased the level of soluble protein. In particular, thermal and
thermochemical pretreatments resulted in significant particle size reduction. The level of soluble protein was
increased by chemical and thermochemical pretreatment methods in particular. Proteins are principle
constituents of organisms and they contain carbon, which is a common organic substance, as well as
hydrogen, oxygen and nitrogen. For this reason, it was considered that as the level of soluble protein
increased, the efficiency of anaerobic digestion would be improved. The differences in the amount of biogas
produced showed that the impact of the rate limiting step could be reduced by pretreatment and digestion
efficiencies of the WAS were consequently improved (Kim et al., 2003).
So far, very little work on fermentative hydrogen production from sewage sludge has been
conducted. Limited data show that the biohydrogen yield using waste activated sludge and anaerobic
fermentation is rather low at:/0.08 mmol H2/g dry solid (DS) (Huang et al., 2000). Pretreatment can increase
the efficiency of anaerobic stabilization of sludge by hydrolyzing the insoluble organic matter to water (e.g.
see Lee and Mueller (2001) and the references cited therein). The use of thermal boiling to hydrolyze sludge
has been demonstrated to promote the anaerobic production of methane (Chu et al., 2002). Cheng et al.
(2000) reported a yield of 0.7 mmol H2/g COD by anaerobically fermenting the thermally boiled waste
activated sludge. This value is close to that obtained for fermenting protein containing wastewater, but is still
much lower than that for polysaccharides. The production of hydrogen from waste sludge does not seem
very promising, given the low yields reported in these preliminary studies (Wang et al., 2003).
_1
Huang et al. (2000) have reported yields of 0.16 g H2 kg from waste activated sludge, whereas
_1
Wang et al. (2003a,b) reported yields equivalent to 0.9–6.3 g H2 kg DS depending on the pretreatment
applied to the sewage sludge. However, it should be noted that so far all these studies used pure cultures of
bacteria as inoculum. Although this may result in higher yields it would be expensive and impractical to
attempt at full scale due to the time, cost and energy requirements of preparing and storing enough bacterial
biomass to seed a digester containing thousands of liters of sewage sludge. Despite this limitation however,
work published so far has demonstrated the feasibility of producing hydrogen fermentatively from sewage
sludge. Wang et al. (2003a) also demonstrated that applying different pretreatment to sewage sludge prior to
fermentation can have a significant effect on hydrogen yield. In particular, sterilization of the sewage sludge
at 121°C consistently improved hydrogen production and reduced the production of methane which can have
a negative impact on hydrogen yield (M. Nicolau et al., 2008). Hydrogen yields of 1.2 mg H2/g COD (Wang et
al. 2003b) and 0.6 mol/kg COD (Wang et al. 2004) were reported when sludge was used as the raw material.
However, higher hydrogen yields (15 mg H2/g COD) were obtained from the filtrate (Wang et al. (2004).
Pretreatment of the sludge increased the soluble COD enhancing the hydrogen yield (0.9 mmol/g dried
sludge) Wang et al. (2003a) (Kapdan and Kargi, 2006; Senturk and Buyukgungor, 2010). Consequently, in
order to further increase the yield of hydrogen production, the impacts of pretreatments on hydrogen
production using sewage sludge should be examined. The reason that the same sludge, using different
pretreatment had different hydrogen yield and lag time was demonstrated (Guo et al., 2008). Table 1 lists
some results from the literature. Various means of producing hydrogen using numerous substrates and seed
bacteria have been reported, such as g H2/g VSS (Volatile suspended solids), mL H2/g COD and others.
Experimental results can almost never be compared because details for unit conversion are nonavailable.

462
Table 1. Comparison of the literature data on biohydrogen production using waste sludge
Operating Conditions
Substrate pH T (°C) rpm Reactor Seed H2 yield References
type
Waste activated sludge 7.0 55 250 Batch Anaerobic sludge (Heat 7.81 mmol Lin et al., 2012
treated at 100°C for 2h)
Municipal sludge 6.7~7.9 36 1,050 Batch - 34.2~34.9 mL/g Volatile Wu et al., 2009
pretreatment with solid (VS)
ultrasonic treatment
Aerobic thermophilic 5.5 55 50 Batch - 136.9 mL H2/g TS Sittijunda et al.,
digestion sludge 2010
Waste sludge sterilization 35 125 Batch Pseudomonas sp. GZ1 1.34 mg H2/g TCOD Guo et al., 2010
for 20 min at 121°C
Wastewater sludge ~ 6.4 35 - Batch Clostridium bifermentans 1.5-2.1 mmol H2/g COD Wang et al., 2003
Sterilized sludge 7.38 35 125 Batch Pseudomonas sp. GZ1 15.02 mL H2/g TCOD Guo et al., 2008
Sterilized sludge - 37±1 140-150 Batch Consortiums in sterilized 11.77 mL H2/g DS Xiao and Liu, 2009
sludge
Alkaline treated sludge 11.0 36±1 150 Batch Alkaline treated sludge 16.59 mL H2/g substrate Cai et al., 2004
Sterilized sludge Defined consortium (1:1:1) 41.23 mL H2/gCODreduced Kotay and Das,
2010
Raw sewage sludge 11.5 7.57 mL H2/gVS

Acid pretreatment 7±0.1 3.25 mL H2/gVS

BenYi and JunXin,
Alkaline pretreatment 11.5 37±1 140-150 Batch - 11.68 mL H2/gVS 2009

Thermal pretreatment 6.8 8.62 mL H2/gVS

Ultrasonic pretreatment 6.9 3.83 mL H2/gVS

463
2.1. Impact of ultrasonication pretreatment on hydrogen production
Ultrasonic could produce cavitation in waste sludge, and cell lysis and particle size reduction would
be caused which the digestion process was enhanced (Clark and Nujjoo, 2000). It appeared from the study
that the mechanical shear forces caused by ultrasonic capitations could be a key factor for sludge
disintegration and collapse of cavitations bubbles could significantly alter the sludge characteristics (Mao et
al., 2004). Tiehm et al. (2001) showed that applying ultrasound (3.6 kW, 31 kHz, 64 s) to sludge
disintegration can release the organic substances into the sludge, so that the soluble chemical oxygen
demand (SCOD) in the supernatant increases from 630 to 2270 mg/L. Moreover, the digestion time could be
reduced to 8 days. Chu et al. (2002a) demonstrated that “weak” ultrasound pretreatment greatly increased
both the production rate and ultimate yield of methane. Chiu et al. (1997) observed simultaneous alkaline
and ultrasound pretreatment was more effective in releasing SCOD.
As for the mechanisms of ultrasonic disintegration, Tiehm et al. (2001) noted that hydromechanical
shear forces produced by ultrasonic cavitation were predominantly responsible for sludge disintegration.
However, it is still unclear about the other contributions to the ultrasonic waste activated sludge
disintegration. There are four paths, which are shown as following, responsible for the ultrasonic activated
sludge disintegration:
 Hydro-mechanical shear forces;
 Oxidizing effect of •OH, •H, •N and •O produced under the ultrasonic radiation;
 Thermal decomposition of volatile hydrophobic substances in the sludge;
 Increase of temperature during ultrasonic activated sludge disintegration.

If the sludge temperature improved, the lipid on the cytoplasmic membrane can be decomposed,
which results in the generation of little holes on the membrane. The intracellular substances can release
through the holes, which causes the increase of SCOD in the supernatant (Wang et al., 2005).

Because the quantity of volatile hydrophobic substances in the sludge is very low, so the third path
can be ignored. The increment speed in sludge temperature is slow. Wang et al. (2005) was reported that
the longest disintegration time is 60 min and the disintegrated sludge temperature is increased to 82◦C.
According to this study on thermal sludge disintegration, improving the sludge temperature to 80◦C in 1 h
produces low SCOD+ and SCOD%. Therefore, the forth can also be neglected. As for the second path, Hart
(1986) and Hart and Henglein (1986) has demonstrated that the amount of •OH is much more than the one
of •H, •N and •O. Accordingly, the oxidizing effect of •H, •N and •O can be ignored. From the above analysis,
we know there are two main paths attributed to the ultrasonic sludge disintegration, namely hydro-
mechanical shear forces and the oxidizing effect of •OH (Wang et al., 2005).

2.2. Impact of thermal pretreatment on hydrogen production

Thermal treatment results in the breakdown of the gel structure of the sludge and the release of
intracellular bound water. Therefore, this treatment allows a high level of solubilisation, an improvement in
biogas production, modification in sludge characteristics (increase in filtreability and viscosity reduction) and
reduction of pathogen microorganisms. The main parameter for thermal treatment is temperature; time of
treatment has less influence. According to several authors, optimal temperature is around 170–200◦C.
Indeed, for higher temperature, biodegradability of sludge is no more improved and can decrease. This can
be due to the formation of refractory compounds linked to Maillard reactions. But, a thermal treatment around
175◦C, combined with anaerobic digestion, can highly reduce sludge production; this reduction can reach
50–70% according to the process (Bougrier et al., 2007; Genc, 2010).

Kim et al. (2003) reported that the organic particulates in WAS were liquidized to soluble
carbohydrates, lipids and proteins or converted into lower molecular weight compounds by thermal
pretreatment. Li and Noike (1992) reported that the optimum pretreatment temperature and contact time for
improving the anaerobic digestion of WAS were 170◦C and 60 min, respectively. Through thermal
pretreatment, the gas production from WAS was greatly increased (Kim et al., 2003).

Sterilized sludge was used to produce hydrogen by anaerobic self-fermentation and raw sludge was
used as control by Xiao and Liu (2009). Owing to the co-existence of hydrogen producing microorganisms
and hydrogen consuming ones and their different growth rates, the hydrogen production and hydrogen
consumption (methane production) could be detected in the anaerobic self-fermentation of the raw sludge.
However, most of the organics presented in the raw sludge were of microbial cells, which were enwrapped
by cell walls, and such few available organics in raw sludge for hydrogen producing microorganisms and
hydrogen consuming ones resulted in their slow growth and metabolism. As a result, there was little

464
hydrogen detected in the anaerobic self-fermentation of raw sludge. Apart from raw sludge, the sterilization
had disrupted some microbial cells and released some intracellular organics from cell, and thus the thermally
screened hydrogen producing microorganisms in the sterilized sludge could quickly use those organics and
produce hydrogen. Xiao and Liu (2009) reported the maximal hydrogen yield of the sterilized sludge (16.26
mL H2/g VS) was much higher than that of the raw sludge (0.35 mL H2/g VS). Guo et al. (2008) was also
produced 1.34 mg H2/g TCOD with sterilization pretreated sludge. This value is comparable to that of protein
fermentation, but is still far lower than that of polysaccharides.
Although most hydrogen consuming microorganisms (mainly methanogens) died during sterilization,
hydrogen consumption still occurred in the later stage (after 36 h) of anaerobic self-fermentation of the
sterilized sludge. Interestingly, no methane was detected in the anaerobic self-fermentation of sterilized
sludge. There were three reasons for the hydrogen consumption: (1) there are a few hydrogen consuming
microorganisms (for example, Homoacetogenic bacteria) in the sludge that can also form spores and resist
autoclaving. In the initial stage of anaerobic fermentation, the hydrogen consuming action was weak and
inconspicuous because of fewer hydrogen consuming microorganisms and their lower growth rate compared
with hydrogen producing ones. In the later stage of anaerobic fermentation, the hydrogen consuming action
became obvious along with the increase in hydrogen consuming microorganisms. (2) The increase of volatile
fatty acids (VFA), particularly, acetic acid could inhibit further growth of hydrogen producing microorganisms
and then result in the decrease of the hydrogen production rate. (3) The anaerobic fermentation of proteins
could not only produce hydrogen but also consume hydrogen, and high content of proteins in the sterilized
sludge may also result in the consumption of hydrogen (Xiao and Liu, 2009).
In spite of these three matters, there were some advantages of using sterilized sludge to produce
hydrogen: (1) without any extra-seeds and any extra feeds, it is simple and convenient to produce hydrogen
by anaerobic self-fermentation; (2) without any extra-added materials (such as alkaline and acid used in
alkaline or acid pretreatment), the sterilized sludge was easy to be treated after hydrogen production; (3) no
methane was produced in the process of anaerobic self fermentation because most of the methanogens
were killed by autoclave (Xiao and Liu, 2009).

Heat treatment temperatures ranging from 75°C to 121°C and exposure times ranging between 15
min to 2 h (Li and Fang, 2007). The most common condition in heat treatment is 100°C for 15 min. Mohan et
al. (2008) evaluated the influence of the heat treatment of anaerobic mixed inoculum at 100°C for 1 h on
hydrogen production. Hydrogen production yield increased from 0.002 mmol H2/g COD to 0.122 mmol H2/g
COD with heat treatment procedure. Mu et al. (2007) used heat, acid and base treatment methods to
suppress methanogenesis in mixed cultures and to enrich H2 producing inoculum. Thus, highest H2 yield of
(2.0 mol-H2/mol-glucose) was achieved with the heat treated sludge. Ren et al. (2008) studied four treatment
methods including heat shock treatment, acid treatment, alkaline treatment and repeated aeration treatment
to enrich hydrogen producing bacteria. Thus, hydrogen production increased from 180.4 mL (control) to
189.5 mL by heat treatment of sludge (Ozkan, 2009).

2.3. Impact of acidic or basic pretreatment on hydrogen production

Hydrogen production occurs in the acidification stage, and the pH value is one of important factors
affecting the biohydrogen fermentation (Cai et al., 2004). In the conventional methanogenic process, the pH
is controlled at near pH 7. At pH below 6.3 or above 7.8, methane production rate would drop sharply (Van
Haandel and Lettinga, 1994; Chen et al., 2002). Thus, the bioactivity of methanogens can be inhibited by
adjusting the pH of anaerobic sludge away from pH 7. On the other hand, the hydrogen producing
Clostridium can resist to extreme acidity and alkalinity by forming protective spores. Wang and Wan (2008)
indicated that the hydrogen production potentials of the sludge treated by acid (96.8 mL), base (125.9 mL)
were higher than hydrogen production potential of the control test (65.7 mL) during the fermentative
hydrogen production using glucose as the substrate (Ozkan, 2009). Alkaline pretreated sludge improved the
hydrogen yield to 1.48 mg H2/g DS in comparison to 0.81 mg H2/g DS when the raw sludge was used as
substrate (Cai et al., 2004).
Cai et al. (2004) were analyzed soluble organic matter concentrations of the raw sludge an alkaline
pretreated sludge at different initial pH values. Obviously, all concentrations of three soluble organic matters
in the alkaline pretreated sludge were much higher than those in the raw sludge. The soluble protein was the
major part of all three (protein, carbohydrates, and lipids) soluble organic matter. For the raw sludge, alkaline
pretreatment was more effective for solubilizing organic matter than acidic pretreatment. For the alkaline
pretreated sludge, the adjustment of pH had impact on soluble organic matter, and all three soluble organic
matters at high initial pH of 11.0 were more than those at low initial pH of 5.0. So, alkaline pretreatment of
sewage sludge could provide more soluble organic matter for biohydrogen production from sewage sludge.

465
2.4. Impact of microwave pretreatment on hydrogen production
The main advantages of microwave heating are the uniformity of heating and the precise control of
the process temperature. Guo et al. (2008) was reported the maximal hydrogen production yield obtained
using microwave pretreated sludge was 14.65 mL (11.44 mL/gTCOD), and lag time was 10 h. Nakazato et
al. (2006) was found out that microwave pretreatment improve sludge dewaterability. This effect was efficient
for hydrolyzing sludge and producing hydrogen. The releasing of nutrient was also found in this study.
Otherwise, the anaerobic digester fed with microwave irradiated sludge was more efficient in inactivation of
fecal coliforms than the other two digesters fed with raw sludge and externally heated sludge, respectively
(Hong et al., 2006). So microwave pretreated sludge could reduce the pathogen (Guo et al., 2008).

3. CONCLUSIONS
The major problems in biohydrogen production from wastes are the low rates and yields of hydrogen
formation. Large reactor volumes are required for biohydrogen production due to low hydrogen production
rates. Low yields and the rates of hydrogen formation may be overcome by selecting and using more
effective organisms or mixed cultures, developing more efficient processing schemes, optimizing the
environmental conditions, and developing more efficient bioreactors. Considerable research and
development studies are needed to improve the ‘state of the art’ in biohydrogen production.
Utilization of sewage sludge as substrate for H2 production hasn’t received much attention than what
it actually deserves. The feasibility studies on fermentative H2 production from sludge as demonstrated by
few researchers suggest that lower yield is the major limitation which precludes the technology from being
commercialized. However, these studies were successful in concluding certain critical findings viz.;
i) Various pretreatments can have significant effect of on H2 production. The main groups of the
organic solids discovered in sewage sludge are proteins, carbohydrates, fats and oils. But these organics are
mostly complex and under-utilized; hence pretreatment becomes essential to render them suitable for H2
fermentation. These pretreatments include alkaline pretreatment, acidification, sterilization, ultrasonication,
microwave, enzyme and heating. Pretreatment not only released insoluble organic matter into water to
increase the efficiency of fermentation, but also inactivated methanogenic bacteria in the substrate to reduce
their consumption of hydrogen.
ii) Nutrient formulation is necessary for amelioration of H2 yield
iii) Mixed cultures are more efficient than pure cultures with respect to H2 production from sewage
sludge and
iv) Sequential H2 and CH4 production system using mixed consortia can augment the total gaseous
energy recovery.
From the results obtained we conclude that biohydrogen yield of sewage sludge using pretreatment
methods with mixed cultures at proper bioreactor conditions can improve.

4. REFERENCES
BenYi, X. and L. JunXin. 2009. Effects of various pretreatments on biohydrogen production from
sewage sludge. Chinese Science Bulletin 54:2038-2044.
Bougrier, C., J. P. Delgenes, H. Carrere. 2007. Impacts of thermal pre-treatments on the
semicontinuous anaerobic digestion of waste activated sludge. Biochemical Engineering Journal 34:20-27.

Cai, M., J. Liu, Y. Wei. 2004. Enhanced biohydrogen production from sewage sludge with alkaline
pretreatment. Environmental Science & Technology 38:3195-3202.

Chairattanamanokorn, P., P. Penthamkeerati, A. Reungsang, Y. C. Lo, W. B. Lu, J. S. Chang. 2009.

Production of biohydrogen from hydrolyzed bagasse with thermally preheated sludge. International Journal
of Hydrogen Energy 34:7612-7617.
Chen, C. C., C. Y. Lin, M. C. Lin. 2002. Acid-base enrichment enhances anaerobic hydrogen
production process. Applied Microbiology and Biotechnology 58(2):224–228.
Cheng, S. S., M. D. Bai, S. M. Chang, K. L. Wu, W. C. Chen. 2000. Studies on the feasibility of
hydrogen production hydrolyzed sludge by anaerobic microorganisms. The Twenty-fifth Wastewater
Technology Conference, Yunlin, Taiwan (in Chinese).
Chiu, Y. C., C. N. Chang, J. G. Lin, S. J. Huang. 1997. Alkaline and ultrasonic pretreatment of sludge
before anaerobic digestion. Water Science and Technology 36 (11):155–162.

Chu, C. P., D. J. Lee, C. S. You, J. H. Tay. 2002a. Weak ultrasonic pretreatment on anaerobic
digestion of flocculated activated biosolids. Water Resources 36:2681–2688.

466
 Chu, C. P., W. W. Lin, D. J. Lee, B. V. Chang, X. F. Peng. 2002. Thermal treatment of waste activated
sludge using liquid boiling. Journal of Environmental Engineering (ASCE) 128:1100-1103.
Clark, P. B. and I. Nujjoo. 2000. Ultrasonic sludge pretreatment for enhanced sludge digestion.
Journal of the Institution of Water & Environmental Management 14:66–71.
Digman, B. and D. S. Kim. 2008. Review: Alternative energy from food processing wastes.
Environmental Progress 27:524–537.
Genc, N. 2010. Biohydrogen production from waste sludge. Sigma Journal of Engineering and Natural
Sciences 28:235-248.
Guo, L., X. M. Li, G. M. Zeng, Y. Zhou. 2010. Effective hydrogen production using waste sludge and
its filtrate. Energy 35:3557-3562.
Guo, L., X. M. Li, X. Bo, Q. Yang, G. M. Zeng, D. Liao, J. J. Liu. 2008. Impacts of sterilization,
microwave and ultrasonication pretreatment on hydrogen producing using waste sludge. Bioresource
Technology 99: 3651–3658.
14,14
Hart, E. J. 1986. Isotropic exchange in the sonolysis of aqueous solutions containing N2 and
15,15
N2. The Journal of Physical Chemistry 90:5989–5991.
Hart, E. J. and A. Henglein. 1986. Sonolytic decomposition of nitrous oxide in aqueous solutions. The
Journal of Physical Chemistry 90:5992-5995.
Hong, S. M., J. K. Park, N. Teeradej, Y. O. Lee, Y. K. Cho, C. H. Park. 2006. Pretreatment of sludge
with microwave for pathogen destruction and improved anaerobic digestion performance. Water
Environment Research 78:76–83.
http://www.tuik.gov.tr (Turkish Statistical Institute, 2008).

Huang, C. H., H. Y. Lin, Y. Y. Tsai, Y. K. Hsie. 2000. The preliminary studies of hydrogen production
from anaerobic digestion with different substrates and cultivations. The Twenty-fifth Wastewater Technology
Conference, Yunlin, Taiwan (in Chinese).
Kapdan, I. K. and Kargi, F. 2006. Biohydrogen production from waste materials. Enzyme and Microbial
Technology 38:569-582.

Kim, J., C. Park, T. H. Kim, M. Lee, S. Kim, S. W. Kim, J. Lee. 2003. Effects of various pretreatments
for enhanced anaerobic digestion with waste activated sludge. Journal of Bioscience and Bioengineering
95:271-275.

Kim, S. H., S. K. Han, H. S. Shin. 2004. Feasibility of biohydrogen production by anaerobic co-
digestion of food waste and sewage sludge. International Journal of Hydrogen Energy 29:1607-1616.
Kotay, S. H. and D. Das. 2010. Microbial hydrogen production from sewage sludge bioaugmented with
a constructed microbial consortium. International Journal of Hydrogen Energy 35:10653-10659.
Kotay, S. M. and D. Das. 2009. Novel dark fermentation involving bioaugmentation with constructed
bacterial consortium for enhanced biohydrogen production from pretreated sewage sludge. International
Journal of Hydrogen Energy 34:7489-7496.
Lee, D. J. and J. A. Mueller. 2001. Preliminary treatments. In: Spinosa, L., Vesilind, A. (Eds.), Sludge
into Biosolids—Processing, Disposal, Utilization, IWA Publishing, London, UK.
Li, C. and H. P. Fang. 2007. Fermentative hydrogen production from wastewater and solid wastes by
mixed cultures. Environmental Science and Technology 37:1–39.
Li, Y. Y. and T. Noike. 1992. Upgrading of anaerobic digestion of waste activated sludge by thermal
pretreatment. Water Science and Technology 26:857-866.

Lin, Y. H., H. X. Zheng, M. L. Juan. 2012. Biohydrogen production using waste activated sludge as a
substrate from fructose-processing wastewater treatment. Process Safety and Environmental Protection
90:221–230.
M. Nicolau, J., R. Dinsdale, A. Guwy. 2008. Hydrogen production from sewage sludge using mixed
microflora inoculum: Effect of pH and enzymatic pretreatment. Bioresource Technology 99:6325–6331.
Mao, T., S. Y. Hong, K. Y. Show, J. H. Tay, D. J. Lee. 2004. A comparison of ultrasound treatment on
primary and secondary sludges. Water Science and Technology 50:91–97.
Mohan, V. S., V. L. Babu, P. N. Sarma. 2008. Effect of various pretreatment methods on anaerobic
mixed microflora to enhance biohydrogen production utilizing dairy wastewater as substrate. Bioresource
Technology 99:59–67.
Mu, Y., H. Yu, G. Wang. 2007. Evaluation of three methods for enriching H2-producing cultures from
anaerobic sludge. Enzyme and Microbial Technology 40:947–953.

467
 Nakazato, T., M. Akasoka, H. Tao. 2006. A rapid fractionation method for heavy metals in soil by
continuous flow sequential extraction assisted by focused microwaves. Analytical and Bioanalytical
Chemistry 386:1515–1523.
Ozkan, L. 2009. Dark fermentative biohydrogen production from sugar-beet processing wastes. M.Sc.
Thesis. The Graduate School of Natural and Applied Sciences of Middle East Technical University. 135 p.

Ozsoy, G., F. B. Dilek, F. D. Sanin. 2006. An investigation of agricultural use potential of wastewater
sludges in Turkey case of heavy metals. Water Science and Technology 54:155–161.
Ray, S., N. Chowdhury, J. A. Lalman, R. Seth, N. Biswas. 2008. Impact of initial pH and linoleic acid
(C18:2) on hydrogen production by a mesophilic anaerobic mixed culture. Journal of Environmental
Engineering (ASCE) 134(2):110–117.
Ren, N., W. Guo, X. Wang, W. Xiang, B. Liu, X. Wang, J. Ding, Z. Chen. 2008. Effects of different
pretreatment methods on fermentation types and dominant bacteria for hydrogen production. International
Journal of Hydrogen Energy 33:4318–4324.
Senturk, I. and H. Buyukgungor. 2010. An examination of used different waste materials and
biohydrogen production methods. Sigma Journal of Engineering and Natural Sciences 28:369-395.
Senturk, I. and H. Buyukgungor. 2012. Evaluation of biohydrogen production potential from marine
macro algae. NATO Advanced Research Workshop 984302 “Black Sea: Strategy for Addressing its Energy
Resources Development & Hydrogen Energy Problems, 7-10 October 2012, Batumi, Georgia.
Senturk, I. and H. Buyukgungor. 2013. Evaluation of biohydrogen production potential from marine
macro algae. Black Sea Energy Resource Development and Hydrogen Energy Problems (Nato Science for
Peace and Security Series-C: Environmental Security), edited by A. Veziroglu, and M. Tsitskishvili. Springer,
450 p.
Sittijunda, S. and A. Reungsang. 2012. Biohydrogen production from waste glycerol and sludge by
anaerobic mixed cultures. International Journal of Hydrogen Energy 37:13789–13796.
Sittijunda, S., A. Reungsang, S. O-thong. 2010. Biohydrogen production from dual digestion
pretreatment of poultry slaughterhouse sludge by anaerobic self-fermentation. International Journal of
Hydrogen Energy 35:13427-13434.
Tiehm, A., K. Nickel, M. Zellhorn, U. Neis. 2001. Ultrasonic waste activated sludge disintegration for
improving anaerobic stabilization. Journal of Water Resources 35 (8):2003–2009.
Valdez-Vazquez, I., R. Sparling, D. Risbey, N. Rinderknecht-Seijas, H. M. Poggi-Varaldo. 2005.
Hydrogen generation via anaerobic fermentation of paper mill wastes. Bioresource Technology 96:1907–
1913.
Van Haandel, A. C. and G. Lettinga. 1994. Anaerobic sewage treatment—A practical guide for regions
with a hot climate. Wiley, New York.
Wang, C. C., C. W. Chang, C. P. Chu, D. J. Lee, B. V. Chang, C. S. Liao. 2003. Producing hydrogen
from wastewater sludge by Clostridium bifermentans. Journal of Biotechnology 102:83-92.
Wang, C. C., C. W. Chang, C. P. Chu, D. J. Lee, B. V. Chang, C. S. Liao. 2004. Efficient production of
hydrogen from wastewater sludge. Journal of Chemical Technology and Biotechnology 79:426–427.
Wang, C. C., C. W. Chang, C. P. Chu, D. J. Lee, B. V. Chang, C. S. Liao. 2003a. Producing hydrogen
from wastewater sludge by Clostridium bifermentans. Journal of Biotechnology 102 (1):83–92.
Wang, C. C., C. W. Chang, C. P. Chu, D. J. Lee, B. V. Chang, C. S. Liao, J. H. Tay. 2003b. Using
filtrate of waste biosolids to effectively produce biohydrogen by anaerobic fermentation. Water Research 37
(11):2789–2793.
Wang, F., Y. Wang, M. Ji. 2005. Mechanisms and kinetics models for ultrasonic waste activated
sludge disintegration. Journal of Hazardous Materials B123:145-150.

Wang, J. and W. Wan. 2008. Comparison of different pretreatment methods for enriching hydrogen
producing bacteria from digested sludge. International Journal of Hydrogen Energy 33:2934-2941.

Wu, F., S. Q. Zhou, Y. L. Lai, W. J. Zhong. 2009. Studies on the effects of pretreatment on production
hydrogen from municipal sludge anaerobic fermentation. Natural Science 1:10-16.
Xiao, B. and J. Liu. 2009. Biological hydrogen production from sterilized sewage sludge by anaerobic
self-fermentation. Journal of Hazardous Materials 168:163-167.

468
 DIRECT PRODUCTION OF HYDROGEN AND SYNTHESIS GAS FROM SOUR NATURAL GAS:
NOVEL TRENDS

H. K. Abdel-Aal
Professor Emeritus, National Research Center, Cairo, Egypt

ABSTRACT
Current technology for hydrogen production from natural gas is mature and practiced on a large scale for the
production of hydrogen. The major part of the world’s hydrogen production is accomplished by the steam
methane reforming (SMR) of natural gas. This is a catalytic process that uses sulfur-free feed stocks of natural
gas.

In this paper, novel trends in using non-catalytic reforming of sour natural gas are described. This will involve
explaining the theory of non-catalytic reforming of methane. The methods include:
Category 1: The application of High Temperature Heat (H.T.H.) for the direct production of hydrogen or synthesis
gas from sour natural gas using two distinctive approaches:
st
1 Chamber Combustion or Non-Catalytic Partial Oxidation (NCPO): It involves the combustion of sour natural
gas under partial oxidation conditions in the absence of catalysts.
nd
2 Rich Combustion in Porous Media under super adiabatic combustion (SAC): This leads to the partial
oxidation of methane.
Category 2: Pyrolysis of natural gas, where natural gas is subjected to a plasma source, producing hydrogen,
carbon black and sulfur, when using sour feed stocks.

1. INTRODUCTION
Natural gas unanimously occupies a place ahead of other fossil fuels for its superior hydrogen content and lower
emissions. Globally, there will be no shortage of gas anytime soon, with reserves estimated at around 165 (t. c.
m.) Unfortunately, one third of the natural gas reserves is classified as “Sour” It contains high carbon dioxide
and hydrogen sulfide concentrations, which considerably hinder the development of such deposits for industrial
utilization
Hydrogen Sulfide is a naturally occurring gas mixed with natural gas or dissolved in the oil or brine and released
upon exposure to atmospheric conditions. The principal source of anthropogenic hydrogen sulfide is as a by-
product in the purification of natural gas. The cost of sulfur contents removal from natural gas and petroleum has
been estimated to be $10 billion/year in USA. Worldwide, increasing energy costs and growing demand for
natural gas along with deep gas-drilling have driven the development of sour gas fields around the world.

Current technology employed to manufacture hydrogen or synthesis gas (hydrogen/ carbon monoxide) from
natural gas is based on one of three major thermo-chemical reforming techniques:

1-Steam reforming of methane (SRM)

2- Partial oxidation (PO)
3- Auto thermal reforming (ATR)

These are catalytic processes, which necessitates sulfur-free feedstock of natural gas to avoid catalyst
poisoning.

Acidic gas removal is a pre-request step for these processes. This is a costly operation involving the use of
amine solvents with subsequent regeneration

In the production of hydrogen and/or synthesis gas from natural gas, three challenging problems are normally
encountered, particularly when using sour natural gas:-

469
 Problem 1- Sour natural gas has to be free from the acidic sulfur compounds. In other words, the natural gas has
to be sulfur-free.
Problem 2- The selection of one of the current technology employed to manufacture hydrogen or synthesis gas
(hydrogen/ carbon monoxide) from natural gas (as shown above).
Problem 3- The hydrogen content in sour natural gas is the sum of hydrogen combined with both the
hydrocarbons and the hydrogen sulfide. In order to optimize the recovery of hydrogen, both sources should be
targeted. When dealing with problem 1 above, we should concern ourselves with the recovery of the hydrogen
combined with the sulfur in forming the sulfur compounds, most important H2S. This is illustrated next in Figure
1.

Figure1- Current Technology of Sweeting Sour Natural Gas

Loss of H2 as H2O

In the combustion of:

H2S +3/2 O2  H2O + SO2

Claus Process Sulfur

Amine Solvents
H2S

Gas Catalytic Hydrogen

Sour N. Gas Sweet N. Gas Reforming
Treatment

Figure 1- Current Technology: Catalytic Reforming of Sour Natural gas to Produce Hydrogen Requires Amine Treatment

2. HOW DO WE GET THE SWEET FROM THE SOUR?

The strength of the carbon-hydrogen covalent bond in methane is among the strongest in all hydrocarbons.
Despite the high activation barrier for breaking the tight C–H bond, CH4 is still the principal starting material for
the manufacture of hydrogen in the reforming of natural gas.

To overcome this problem, scientists are indulged into the development of catalytic reforming processes. The
other promising option is the non-catalytic approach recommended to handle sour feed stocks as illustrated in
Figure 2.

OPTIONS IN REFORMING

SOUR NATURAL GAS

470
 [Type
Gas a quote from
Sweetening

NON- CATALYTIC

CATALYTIC REFORMING

REFORMING

HIGH TEMP. HEAT PYROLYSIS

(HTH) (BY PLASMA)

CHAMBER COMBUSTION RICH COMBUTION IN

NON CATALYTIC PARTIAL POROUS MEDIA: SUPER
OXIDATION (NCPO) ADIABATIC COMBUS. (SAC)

Figure 2- Options in Reforming Sour Natural Gas

The non -catalytic option, as proposed to handle the reforming of sour natural gas, include the following
methods. The reforming processes would eventually lead to recover hydrogen gas (the sweet) from the sour
natural gas.
• The application of High Temperature Heat (H.T.H.) for the direct production of hydrogen and synthesis
gas from sour natural gas using two distinctive approaches.
a) Chamber Combustion or Non-Catalytic Partial Oxidation (NCPO): It involves the combustion of sour
natural gas under partial oxidation conditions in the absence of catalysts.
b) Rich Combustion in Porous Media under super adiabatic combustion (SAC): This leads to the partial
oxidation of methane.
• Pyrolysis of natural gas, where natural gas is subjected to a plasma source, producing hydrogen,
carbon black and sulfur, when using sour feed stocks.

3. CHAMBER COMBUSTION: Non-Catalytic Partial Oxidation (NCPO) OF SOUR GAS

Sour natural gas is simply exemplified by methane containing hydrogen sulfide. The NCPO for synthesis gas
production was introduced by Abdel-Aal and coworkers[1,2] is illustrated as shown in Figure 3. The basic step in
NCPO process is the simultaneous partial oxidation of methane along with complete combustion of H2S
according to equations 1 and 2. This occurs by the complete reaction of the oxygen with some of the merthane
to give water and carbon dioxide, followed by the slower reactions of these products with the excess methane
forming a mixture of hydrogen and carbon dioxide. The reaction of H2S with oxygen takes place simultaneously.

471
 CH4 + 1/2 O2 CO + 2H2 ......(1)
H2S + 3/2 O2 SO2 + H2O ......(2)

Figure3- NCPO of sour natural gas producing hydrogen and sulfuric acid

It is of interest to compare between the steam reforming method (SRM) and the proposed NCPO as far as the
production of sulfuric acid is concerned. In the first method, elementary sulfur, once produced by the Claus
process, is converted to sulfuric acid via the steps: SSO2SO3  H2SO4. In the latter method, steps are
2
represented by: the absorption of SO by water to give  H2SO3 (then apply electric current) H2SO4 + H2

4. RICH COMBUSTION IN POROUS MEDIA

Super-adiabatic combustion (SAC), also known as filtration combustion (FC) of a fuel gas/oxidant mixture, is
one of the well-known techniques that have been reported in the literature by Marras et al [3] to produce
hydrogen. A porous ceramic medium with a high thermal capacity is used in the combustion reactor. Two stages
are described in this process:
• First: The super adiabatic temperature drives the reactions of the mixture beyond the conventional
flammability limit, making ultra-rich combustion possible. Methane is thermally decomposed when the
porous media is at high temperature. Reaction is represented by: CH4  C + 2H2.
O2 will then be rapidly consumed, followed by fast production of H2 and H2O. This stage accounts for the
majority of the fuel (methane) conversion.
• Second: This stage is recognized as the reforming stage, where the following reaction takes place:
CH4 + H2O  3H2 + CO
It should be noted that in the literature surveyed on rich combustion in porous media of natural gas, sweet
methane was the only fuel examined so far.

5. HYDROGEN PRODUCTION BY PYROLYSIS (PYR)

The pyrolysis of natural gas produces two main products: Hydrogen and Carbon Black. Since 1992, a full-scale
pilot plant has been operating and the development has resulted in a technology which would lead to
commercialization as reported by Lynum [4].

In case of sour natural gas, elementary sulfur will be produced as an additional product.
The process is simplified schematically as shown in Figure 4.

472
 Plasma torch
Hydrogen

Feed (Sour natural Gas) Carbon black

REACTOR
Sulfur

Figure4-Schematic Presentation of a Pyrolysis Unit

CONCLUSIONS:

When it comes to utilize the large reserves of natural gas found in many oil-producing countries- much of it is
sour, it is found that separation of the acidic gases is a current practice to produce hydrogen or synthesis gas by
SMR.
The non-catalytic reforming of methane offers promising means to overcome the SMR costly operation. The
potential of exporting natural gas from its resources in the form of hydrogen to the international market is very
encouraging. Keeping in mind that one trillion cubic feet of Natural gas, produces 8700 Tons of Liquid Hydrogen.
Imagine how much hydrogen could be produced from the proven reserves of natural gas available in the Middle
East, It is estimated to be about 40% of the total world reserves.
From the economic point of view, the cost of removal of sulfur contents (12-15%) from natural gas and petroleum
has been estimated to be $10 billion/year in USA. Worldwide, increasing energy costs and growing demand for
natural gas along with deep gas-drilling have driven the development of sour gas fields around the world. In
addition, about forty percent of the world’s natural gas reserves are in the form of sour gas where H2S and CO2
compositions exceed 10% by volume.

REFERENCES
1. Abdel-Aal, H.K.; Shalabi M.A., Noncatalytic Partial Oxidation of Sour Natural gas…Ind. Eng. Chem. Res.
1966, 35, 1785-1787.
2. Abdel-Aal et al, Simulation of the Direct Production of Synthesis gas from Sour Natural Gas by NCPO, Ind.
Eng. Chem. Res.,Vol 38,No 3,pages 1069-1074, 1999.
3. Simon Marras et al., Non-catalytic fuel reforming of a rich hydrocarbon fuel/air mixture in a porous medium
burner, paper presented at the Westerns States Section of the Combustion Institute, Boise State University,
Idaho, USA, March 2006
4. Lynum, S; CO2- Free Hydrogen from Hydrocarbons: The Kvaerner CB & H Process; 5 th Annual US
Hydrogen Meeting, National Hydrogen association, 1994

473
A DYNAMIC SIMULATION STUDY OF A SMALL SCALE HYDROGEN PRODUCTION SYSTEM
FOR A HIGH TEMPERATURE PROTON EXCHANGE FUEL CELL

1
A. Ersoz
1
TUBITAK Marmara Research Center Energy Institute, 41470, Gebze KOCAELI, TURKIYE
email: atilla.ersoz@tubitak.gov.tr, Tel: 0090 262 6772707, Fax: 0090 6412309.

SUMMARY

Fuel cell-based micro-combined heat and power (micro-CHP) systems have received increasing attention recently.
In this study, a small scale hydrogen production system for 1 kW e High Temperature Fuel Cell (HTPEM) is
presented and analyzed to investigate its dynamic behavior. The simulated hydrogen production system (HPS)
can be integrated with a fuel cell to produce 1 kW e power for residential uninterruptable power supply (UPS)
application. Within this frame, HPS is modeled as a steady state process and then it is operated at dynamic
conditions. The system is simulated with ASPEN HYSYS process simulation software. A dynamic simulation study
has been conducted to investigate the changes on temperatures, compositions, feed conditions and electrical
loads during start-up period of HPS. Hydrogen and CO have been observed in HTPEM feed stream as 34% and
0.3% respectively within these optimum ratio conditions. The results show that electrical efficiency is in the range
of 26-30% and fuel processing efficiency is about 81-83% at this load level.

Keywords: Fuel Processing, Hydrogen, Reforming, Microcogeneration, High Temperature Proton Exhange
Membrane Fuel Cell, Uninterruptible Power Supply

INTRODUCTION

Fuel cell-based micro-combined heat and power (micro-CHP) systems have received increasing attention recently.
This is due to a number of reasons. First of all distributed generation minimizes transmission losses. Secondly, a
well developed infrastructure for distributed natural gas is already presented in several countries. Therefore, the
step to produce electricity in domestic households is realistic provided that the technical and financial issues of the
CHP units can be resolved. Along with the technological progress in small-scale energy conversion devices such
as Stirling engines, micro-turbines and fuel cells, small-scale applications from just a few kWs up to some tens of
kWs have become available (Lund, 2011). Among fuel cells, a range of different types exists including proton
exchange membrane (PEM) and solid oxide fuel cells (SOFC) (Korsgaard et al., 2008).

Stationary fuel cell-based power generation systems also offers a great market opportunity, because the
technology is capable of achieving higher efficiencies, with lower emissions as compared to conventional power
systems (Arsalis et al., 2011). A micro-CHP system provides electricity and heat (space heating and hot water) for
a residential building demand. These systems are designed to convert the chemical energy in a hydrocarbon
feedstock into both electrical power and useful heat (Nielsen et al., 2011).

Fuel cells can be one of the ideal technologies for micro-CHP generation in terms of some characteristics such as
high reliability, low environmental impact and size flexibility. In the last decade, micro-CHP based on PEM fuel
cells has been intensively developed. Most of these systems are based on low temperature PEM fuel cells which
operate at less than 100 °C. Several studies about PEMFC based CHP processes can be found in open literature.
The development of PBI based MEA seems to have introduced with some improvements as operation with natural
gas in recent years. These MEAs withstand higher temperatures (up to 180°C) and therefore tolerance to carbon
monoxide (CO) is improved. As a result of these improvements, simplified purification processes in reformers are
required. The fuel cells based on these MEAs are normally referred as High Temperature Proton Exchange
Membrane Fuel Cells (HTPEM) in contrast with Nafion based Low Temperature Proton Exchange Membrane
(LTPEM) fuel cells which require high purity hydrogen. Although in the last year’s attention to HTPEM fuel cells
has been growing, most research activities concern with membrane modeling or experimental characterization of
stacks and only few with CHP systems (Zuliani et al., 2012).

Currently the fuel cell technology based on hydrogen rich reformate gas has not made the market penetration.
This depends on several parameters. The current systems are too complex and have a too long start up time for
the application in the uninterruptible power supply market. Hence only prototypes exist based on reforming of
natural gas and high temperature proton exchange membrane fuel cells (HTPEMFC).

474
Recent progress in H2/air PEM fuel cells has focused to develop PEM fuel cells that operate above 100 °C. There
are several reasons for operating at a higher temperature: (i) Electrochemical kinetics for both electrode reactions
are enhanced; (ii) Water management can be simplified because only a single phase of water need to be
considered; (iii) The cooling system is simplified due to the increased temperature gradient between the fuel cell
stack and the coolant; (iv) Waste heat can be recovered as a practical energy source; (v) CO tolerance is
dramatically increased thereby allowing fuel cells to use lower quality reformed hydrogen (Zhang et al., 2006).

Propane, natural gas as well as liquid hydrocarbons are currently the preferred fuels because of their high storage
density for hydrogen generation in fuel cell applications. Using these fuels the existing infrastructure can be used
until hydrogen widely available.

Currently, the most promising fuel processing system design is to use the auto thermal reforming of natural gas in
combination with a high temperature fuel cell. The high temperature fuel cell can tolerate CO contents of about 3%
vol. at operation temperatures around 180°C compared to a common PEM fuel cell (Heinzel et al., 2006).
Unfortunately systems based on this technology are currently too complex in the system design. Fuel cell systems
using hydrocarbons as hydrogen source are mainly combined with fuel cells working either on a very low or a very
high temperature level. Fuel cells working at temperatures below 100°C (common PEM technology) can only be
operated with CO contents that are lower than 100 ppm (Smolinka,2005) Therefore extensive hydrogen
purification steps and a very high control effort are necessary. Fuel cells which are working at very high
temperatures larger than 750°C (e.g. SOFC) don’t need any hydrogen rich gas clean-up steps after the reforming
process but they have problems because of thermal stress of the materials and sealing. Those fuel cells cannot be
used in applications where quick start ups or dynamic operation modes are required.

The CHP plants based on fuel cells represent an interesting alternative to traditional technologies for the combined
production of electrical energy and heat; this is due to the high efficiency of the FCs in partial load, the modularity,
the ability to ensure substantial autonomy of the user from the electricity transmission grid and, last but not least,
the possible reduction of environmental impact (Barelli et al., 2012). A dynamic model is essential for the fuel
processor operation for the following reasons: (1) discriminating control system design for improved load rejection,
and (2) evaluating start-up strategies for fast start-up (Chena et al., 2006).

DYNAMIC SIMULATION

Dynamic simulation can help to create a better design, optimization and operation of a chemical process.
Chemical processes are never truly at steady state. Feed and environmental disturbances, heat exchanger fouling,
and catalytic degradation continuously upset the conditions of a smooth running process. The transient behavior of
the process system is the best studied using a dynamic simulation package. The design and optimization of a
chemical process involves the study of both steady state and dynamic behavior. Steady state models can perform
steady state energy and material balances and evaluates different plant scenarios. With dynamic simulation, it
can be confirmed that the plant can produce the desired product in a manner that is safe and easy to operate
(Aspen HYSYS, 2009).

In Aspen HYSYS software, the dynamic analysis of a process system can provide insight into the process system
when it is not possible with steady state modeling. The following conditions can be investigated with dynamic
simulation:
• Process optimization
• Controller optimization
• Safety evaluation
• Transitions between operating conditions
• Startup/Shutdown conditions

The Aspen HYSYS dynamic model shares the same physical property packages as the steady state model. The
dynamic model simulates the thermal, equilibrium, and reactive behavior of the chemical system in a similar
manner as the steady state model (Aspen HYSYS, 2009).

On the other hand, the dynamic model uses a different set of conservation equations which account for changes
occurring over time. The equations for material, energy, and composition balances include an additional
“accumulation” term which is differentiated with respect to time. Non-linear differential equations can be formulated
to approximate the conservation principles; however, an analytical solution method does not exist (Aspen HYSYS,
2009).

Therefore, numerical integration is used to determine the process behavior at distinct time steps. The smaller the
time step, the more closely the calculated solution matches the analytic solution. However, this gain in rigour is

475
offset by the additional calculation time required to simulate the same amount of elapsed real time. A reasonable
compromise is achieved by using the largest possible step size, while maintaining an acceptable degree of
accuracy without becoming unstable (Aspen HYSYS, 2009).
The aim of this work is to investigate the dynamic characteristics and to realize the dynamic process simulation
studies for hydrogen production system (HPS) at different loads. Dynamic process simulation studies were
performed for the integration of HPS including ATR, HTS and LTS reactors and HTPEMFC with a significant
reduced start up time. System startup period phases envisaged for the dynamic simulation parameters are within
the scope. Different operating conditions were accepted. ASPEN HYSYS process simulation software was used
for the dynamic simulation studies.

PROCESS LAYOUT

The system is based on an autothermal reforming (ATR) process of natural gas and a high temperature fuel cell
stack (HTPEMFC). A combination of both the processes steam reforming (SR) and partial oxidation (POX), called
autothermal reforming (ATR) has been used for the main HPS section as shown in Fig.1.

Fig. 1. Schematic view of the 1 kWe HTPEM FC process

The main characteristics of ATR are;

- Low energy requirement due to the complementary steam reforming (SR) and partial oxidation (POX)
reactions,
- Low energy consumption,
- High Gas Space Velocity (GSV), at least one order of magnitude larger than traditional SR,
- Preset H2/CO ratio easily regulated by inlet reactant ratios
- CO2 recycling.

Natural gas, steam and air streams are introduced to ATR inlet at the beginning of the envisaged process. The
autothermal reforming reactor is controlled at an outlet temperature around 700-800°C which is optimized for
desired hydrogen yield. Then, the outlet stream is cooled down prior to HTS inlet temperature around 350°C.
Similarly, HTS outlet is cooled down to about 250°C for LTS inlet stream. The exothermic water gas shift reactions
(HTS and LTS) convert the entire CO generated by SR and POX into CO2, while producing the maximum yield of
H2 by utilizing the steam as a co-reactant. Finally, the syngas coming from LTS reactor cooled down again for the
HTPEM inlet operating conditions around 180°C. The following main reactions (Equation 1 to 5) can be described
in the overall reforming process:

CH4 + H2O → CO + 3H2 ∆H 298°C = 206 kJmol -1 (1)

CO + H2O → CO2 + H2 ∆H 298°C = - 41 kJmol -1
(2)
CH4 + 2H2O → CO2 + 4H2 ∆H 298°C = 165 kJmol -1
(3)

476
 CH4 + 1/2 O2 → CO + 2H2 ∆H 298°C = -35.7 kJmol -1 (4)
CH4 + CO2 ⇔ 2CO + 2H2 ∆H 298°C = 247.4 kJmol -1
(5)
This process includes the main units implemented in the dynamic simulation study. The main units constituting the
process model are as follows:

- HTPEM fuel cell for the production of electricity

- Auto thermal reformer (ATR) for the hydrogen production starting from natural gas
- High and low temperature water gas shift reactors (HTS & LTS) for hydrogen rich gas purification
- Heat exchangers to maintain desired temperature conditions
- Control valves
- Control spreadsheet
- Flow controllers

The main issue is a fast start up of this HTPEMFC system. This requirement is essential for the applications of the
uninterruptible power supply (UPS) unit as well as for micro cogeneration (CHP) systems. One of the main issues
for the UPS systems is to serve power during the electrical power outage. This challenge also has to be solved for
the CHP market especially for dynamic loads and for standby modes.

RESULTS AND DISCUSSION

Dynamic simulation results of the single ATR reactor for start-up period

Simulations are based on to carry out primarily POX mode initialization of the ATR reactor. Within this assumption,
only air and natural gas should be fed to first ATR reactor at the start-up phase and then the water vapor supply is
fed after a period of time (Fig.2).

The reactor feed stream valves are adjusted under dynamic simulation conditions. Using the simulation integrator
module, the ATR exit temperature, outlet compositions, steam/carbon and oxygen/carbon ratios, natural gas,
steam and air flows is recorded versus time (Fig.3). Electrical load changes (kWe) have also been observed in the
light of these data.

Fig. 2. Dynamic simulation of ATR

477
Fig. 3. Integrator module and flow control modules of dynamic simulation study

The temperature range is determined as 700-800° C to be appropriate at the optimum operating conditions of the
ATR as can be seen from Fig.4. In particular, suitable choice of operating temperature range of ATR affects the
dynamic behavior, strength and life time of the commercial catalyst and the reactor outlet compositions. Thus, it
can be concluded that the targeted electrical load of the whole integrated process is affected by operation
temperature and the reactor outlet compositions.

Fig. 4. ATR temperature profile during start-up period Fig. 5. Composition profile of ATR exit stream at start- up period

ATR reactor is being simulated within a dynamic equilibrium reaction model approach. Thermodynamic equilibrium
takes place in the reactor considered during dynamic simulation work. ATR outlet compositions are calculated
separately for each component. For the star-up period, the composition profile at ATR product stream is shown in
Fig.5. CO concentration is about the order of 20% at the partial oxidation (POX) mode during initiation period. It is
reduced to approximately 0.4% after reaching steady state (approximately 10 min. later) due to the steam supply
to ATR reactor. Similarly, while CO2 is produced approximately in the range of 0.1% – 0.2% during initial stage, it
is increased to 0.7% – 0.75% at steady state conditions.

The reactor exit stream includes about 35% H2O at the end of 10 minutes after starting feeding of water vapor
approximately 100 seconds following the initial phase. A portion of the steam feed is consumed in the endothermic
steam reforming reactions.

Hydrogen production is an extremely important step mainly through this process with the targeted fuel cell power
generation. It seems that the order of 26 - 28% wet basis of hydrogen is produced prior to the water gas shift
reactor during start-up period.

478
The oxygen feed is given by the airflow which is especially consumed in POX reactions during the initiation and
progression of the ATR step. The inert nitrogen which is not involved in the reaction is located in the order of
approximately 26% at reaching steady state point while the oxygen in the air is worked in the formation of
exothermic reactions of partial oxidation.

Steam / carbon ratio is especially critical in terms of avoiding carbon deposition on the catalyst and the realization
of endothermic steam reformer reactions. It can be said that the optimum operating ratio is about 3 during the
initial startup period, except for POX mode (Fig.6). The reactor temperature increases but also the efficiency of
hydrogen production reduces for the lower values of S/C ratios.

Fig. 6. Steam/Carbon (S/C) ratio at start-up of ATR Fig. 7. Oxygen/Carbon (O2/C) ratio at start-up of ATR

Oxygen / carbon ratio has a critical importance in the progression of the exothermic partial oxidation reactions. The
efficiency of hydrogen production reduces if this rate is high. At the same time higher O2/C ratios also causes an
increase of the reactor temperature. The best data is observed between 0.6 and 0.7 at steady state condition
(Fig.7). In particular, it is very important to keep this ratio at the required the optimum conditions to prevent the
undesirable adiabatic reaction temperature peaks or hot spots (runaway criteria). The life of the commercial
catalyst used in the ATR reactor is adversely affected by unexpected temperature rises.

Natural gas feed increases up to 0.35 kgh-1 approximately at steady state conditions while it is about 0.05 kgh-1 at
the beginning the starting period. Similarly, it can be said that the value of the water vapor is approximately equal
to 1.2 kgh-1 with a suitable S/C ratio adjustments. Air flow value is observed approximately equal to 2 kgh-1. at
steady state conditions with a suitable O2/C adjustment while it is about 0.5% initially used in the reaction. The
feed flow rates profile has been shown in Fig.8 for start-up period of ATR reactor.

Fig. 8. Feed flow rates profile during start-up period Fig. 9. Load change profile at start-up period of ATR
of ATR reactor

So far, all of the above-mentioned process variables and operating parameters are critical to achieving the target
value of 1 kW e fuel cell power generation system with the application of uninterrupted power supply. The load
increases from the initial value of 0.2 kW e to 1.2 kW e over the system start-up period in10 min as can be seen from
dynamic simulation study (Fig.9). System startup time is directly related to fuel cell behavior at high temperatures,
heat integration, technical limitations etc. So, it is especially targeted to reach steady state for ATR reactor
conditions under a period of 10 minutes.

479
Dynamic simulation results for the ATR + HTS + LTS integration at start-up period

The simulation of a combination of ATR-HTS-LTS reactors was carried out following the study of the singular
dynamic simulation of the ATR reactor (Fig.10). The above results are only given for ATR as well as to work within
this integration which is carried out under the same conditions. ATR-HTS-LTS integration procedure is similar to
the work of individual simulated ATR reactor during the dynamic simulations of startup period. The ATR graphs
have very similar data which is provided for the previous work for ATR reactor. Because of this, these graphs are
not mentioned again in this section.

HTS and LTS reactor interactions based on the temperature and composition were investigated and are reported
below.

Fig. 10. Dynamic simulation work for ATR-HTS-LTS integration

Table 1 summarizes the operating conditions and the result of the efficiency calculations. The total thermal
efficiency of hydrogen production system (HPS) is calculated as the ratio of the heating value of the HTPEM
inlet fuel (hydrogen energy) and natural gas inlet stream. The heating value of a stream is calculated by multiplying
of its lower heating value (LHV) with its mass flow rate. HPS represents the ratio of heating value of the
HTPEM inlet stream to the heating value of the total fuel feed stream. The efficiency definitions are given in
equation 6 and 7. The net electrical efficiency of the fuel cell system, net elec , is the ratio of current power load
and the energy content of NG feed.

Table 1 Process data for 1,097 kW e power demand

Steam feed (kmolh-1) 0,050 LHV of NG feed (kJkg-1) 50035

-1
NG feed (kmolh ) 0,016 Energy content of NG Feed (kW) 3,609
-1
S/C 3,078 NG feed (kgh ) 0,26
-1 -1
Air feed (kmolh ) 0,048 NG feed (kgs ) 0,0001

O2/C 0,614 Energy Content of H2 produced (kW) 2,959

O2 in air 0,208 HTPEM in - H2 produced (kgh-1) 0,089

LHV of H2 (kJkg-1) 120.000 Current power load (kWe) 1,097

PEM feed (kmolh-1) 0,124 HTPEM-Feed Molar flow (kmolh-1) 0,136
η HPS 81,991 % η net.elec 30,392

The fuel processing efficiency and net electrical efficiency are calculated as:

η HPS = (m HTPEM inlet fuel x LHV HTPEM Feed ) / (m Fuel feed (NG) x LHV Fuel feed (NG) ) (6)
η net electric = P current / LHV NG feed (7)
-1 -1
in which m is the mass flow rate of the feeds (kgh ), LHV is the lower heating value (kJkg ), P is the power load
(kW e). The number of cells in the fuel cell stack is one of the parameters, which affects the fuel cell electrical
efficiency, hence, also the net electrical efficiency. The number of HTPEM cells has been taken as 50 in this study.
The details of the stream data of this integrated simulation study is given in Table 2.

480
Table 2 Stream data of the integrated simulation study for ATR+HTS+LTS
Name Steam NG Air Steam NG Feed Air Feed 9 LTS Out
Feed
Vapour Fraction 1,000 1,000 1,000 1,000 1,000 1,000 0,318 1,000
Temperature (C) 350 150 250 350 150 250 150,000 258
Pressure (kPa) 120,000 120,000 120,000 110,306 110,306 110,306 50,000 105,393
-1
Molar Flow (kmolh ) 0,050 0,016 0,048 0,050 0,016 0,048 0,000 0,136
-1
Mass Flow (kgh ) 0,898 0,260 1,381 0,898 0,260 1,381 0,000 2,538

Name 5 HTPEM in 5_1 4 6 7 8 Gas out

Vapour Fraction 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
Temperature (C) 258 180 267 180 180 180 18 672
Pressure (kPa) 105,393 102,995 68,942 101,325 102,995 102,994 102,995 108,904
-1
Molar Flow (kmolh ) 0,000 0,136 0,000 -0,136 -0,136 0,000 0,000 0,136
-1
Mass Flow (kgh ) 0,000 2,538 0,000 -2,538 -2,538 0,000 0,000 2,538

Name 8_1 HTS Out Product 1 3_1 HTS Out- 3 2

1
Vapour Fraction 0,885 1,000 1,000 1,000 1,000 1,000 1,000 1,000
Temperature (C) 25 384 673 350 407 384 384 250
Pressure (kPa) 60,000 106,820 110,306 106,821 66,859 106,459 106,821 105,393
-1
Molar Flow (kmolh ) 0,000 0,136 0,136 0,136 0,000 0,136 0,000 0,136
-1
Mass Flow (kgh ) 0,000 2,538 2,538 2,538 0,000 2,538 0,000 2,538

HTS is the first reactor which is integrated at the ATR optimum operating conditions. HTS input stream reduced to
a value of approximately 350°C via cooling of the ATR effluent. According to the system initialization period, the
temperature of outlet stream of HTS is expected approximately to be the order of 390°C at steady state (Fig.11).
An appropriate HTS operating temperature range affects the dynamic behavior and the strength of the commercial
catalyst and the reactor outlet compositions. The electrical load of integrated process is influenced by the reactor
outlet compositions and operating temperatures.

Fig. 11. HTS temperature profile at start-up period

Water-gas shift equilibrium conversion reactions have been assumed in the reactor module during the dynamics
simulations of HTS reactor. Outlet compositions have been calculated separately for each component. The
composition profile at HTS product stream is shown in Fig.12 for the star-up period. CO concentrations at the ATR
exit can be reduced to approximately 0.1% through steady state due to supply steam while it is about the order of
14% during the startup period .

This composition is the input value for the LTS reactor which will take place in the next stage. Similarly, the initially
produced CO2 in the order of 4%, reaches up to about 10% at steady state conditions. The water vapor coming
from ATR exit is consumed in HTS reactions then it reaches in the order of 30% between 100 and 200 sec.. A

481
portion of the water vapor is consumed in exothermic water gas shift reaction for the additional CO and H2
production.

The amount of hydrogen produced is reached to the range of 29% while it was in the order of 26 to 27% wet based
prior to the water gas shift reactor during the startup period.

Inert nitrogen which is not taking place in the reactions has been abandoned the HTS reactor after the ATR
reactions. The concentration of N2 is in the order of about 30% after reaching steady state point.

Fig. 12. Composition profile of HTS exit stream at Fig. 13. Temperature profile of LTS at start-up period
start-up period

LTS inlet temperature is determined as 250°C which is required for the normal operating range of the catalyst. The
temperature of LTS input stream is reduced to 250°C by cooling the HTS exit stream. The steady state
temperatures of LTS output stream is reached to the value of 257°C due to exothermic reaction according to the
system startup period (Fig.13). LTS operating temperature also affects the dynamic behavior and the strength of
the catalyst and also the outlet compositions.

Similarly to HTS reactor, water-gas shift equilibrium conversion reactions have also been assumed during the
dynamics simulations of LTS reactor. Outlet compositions have been calculated separately for each component.
The HTS composition profile at the product stream is shown in Fig.14 for the star-up period.

CO concentrations at the HTS exit can be reduced to approximately 0.01% after reaching steady state due to
supply steam while it was about the order of 0.1% at the exit of HTS during the startup period. This composition
value is desired as input composition of the anode inlet fuel for the HTPEM fuel cell. The effect of CO is extremely
important for PEM fuel cells as it is known. The high amount of CO which can be higher than desired value
reduces the effectiveness of the fuel cell catalyst and provides the loss of the activation in a short time. Similarly,
the initially produced CO2 of the order of 4% reaches up to about 11% at steady state.

The water vapor coming from HTS exit is consumed in LTS reactions then it reaches in the order of 30% between
100 and 200 sec.. A portion of the water vapor is consumed in exothermic water gas shift reaction for the
additional CO and H2 production. The amount of wet based hydrogen is reached to the range of 30% while it was
in the order of 26 to 27% wet based prior to the water gas shift reactors during the startup period. Inert nitrogen
which is not taking place in the reactions has been abandoned the LTS reactor after the HTS reactions. The
concentration of N2 is in the order of about 30% after reaching steady state point.

482
Fig. 14. Composition profile of LTS exit stream at Fig. 15. Load change profile at start-up period (ATR-HTS-LTS
integration) start-up period

The electrical load increases from the initial value of 0.17 kW e up to the value of 1.02 kW e as can be seen from a
dynamic simulation study over the period the system startup. The power variation is then fixed after 200 sec.
(Fig.15).

CONCLUDING REMARKS

In this study a process simulation model has been used to realize an analysis of a system consist of a fuel
processing system for hydrogen rich gas production using natural gas. An HTPEM fuel cell stack of 1 kW e has
been integrated to HPS. The dynamic simulation model includes units such as HTPEM fuel cell, auto thermal
reformer (ATR), high and low temperature water gas shift reactors (HTS & LTS), heat exchangers, control valves,
control spreadsheet and flow controllers.

Several observations have been noted which are based on the results as presented in the study. The dynamic
simulation model of hydrogen production process using autothermal reforming of natural gas has been
successfully developed using Aspen HYSYS software. The results show that electrical efficiency is in the range of
26-30% and fuel processing efficiency is about 81-83% at this load level.

Steam/carbon (S/C) ratio is a very critical parameter in terms of preventing carbon formation on the real catalytic
conditions in the reactors during the endothermic reactions. It can be said that the optimum operating conditions
for S/C ratio is about 3, during the initial startup period, except for POX mode. The reactor temperature increases
but also the efficiency of hydrogen production reduces at below this S/C value.

Oxygen/carbon (O2/C) ratio has also a critical importance in the progression of the exothermic partial oxidation
reactions. The efficiency of hydrogen production reduces if this rate is high. At the same time this also causes an
increase of the reactor temperature. The best condition for O2/C ratio is observed between 0.6 and 0.7 at steady
state conditions. Hydrogen and CO have been observed in HTPEM feed stream as 34% and 0.3% respectively
within these optimum ratio conditions.

The system efficiencies can be improved choosing fuel cell stack operation modes at higher voltages. The balance
of plant can components can be used with a less complicated case and more reliably in HTPEM fuel cells in
comparison with LTPEM systems.

ACKNOWLEDGEMENT

This work is a part of a project under award number “109R036” which is supported by TUBITAK (The
Scientific And Technological Research Council of Turkey). We would like to thank TUBITAK for funding
this project work.

483
REFERENCES

1. Lund PD. Energy relevance of microgeneration - Case advanced fuel cells. International Journal of Energy
Research 2011;35: 1100–1106.
2. Korsgaard AR, Nielsen MP, Kær SK. Part two: Control of a novel HTPEM-based micro combined heat and
power fuel cell system. International Journal of Hydrogen Energy 2008; 33 : 1921-1931.
3. Arsalis A, Nielsen MP, Kær SK. Modeling and off-design performance of a 1 kW e HT-PEMFC (high temperature-
proton exchange membrane fuel cell)-based residential micro-CHP (combined-heat-and-power) system for Danish
single-family households. Energy 2011; 36 : 993-1002.
4. Arsalis A, Nielsen MP, Kær SK. Modeling and parametric study of a 1 kW e HT-PEMFC-based residential micro-
CHP system. International Journal of Hydrogen Energy 2011; 36: 5010-5020.
5. Zuliani N, Taccani R. Microcogeneration system based on HTPEM fuel cell fueled with natural gas:
Performance analysis. Applied Energy 2012; 97 : 802–808.
6. Zhang J, Xie Z, et.al. High temperature PEM fuel cells. Journal of Power Sources 2006; 160 : 872–891.
7. Heinzel A, Burfeind J, Bandlamudi G, Kundler I. Betriebserfahrungen mit Hochtemperatur-PEM von PEMEAS,
Riesaer Brennstoffzellen-Workshop 2006; http://brennstoffzellen.pitcom.net/upload/dokument105.pdf (Date of
access: 02.01.2013).
8. Smolinka T. Untersuchungen an einer mit Reformat betriebenen PEM-Brennstoffzelle, Universität Ulm,
Dissertation 2005; http://publica.fraunhofer.de/documents/N-49696.html (Date of access: 02.01.2013).
9. Barelli L, Bidini G, Gallorini F, Ottaviano A. Dynamic analysis of PEMFC-based CHP systems for domestic
application. Applied Energy 2012; 91 : 13–28.
10. Chena YH, Yu CC, et al. Start-up strategies of an experimental fuel processor. Journal of Power Sources
2006, 160 : 1275–1286.
11. Aspen HYSYS Dynamics User Guide, Version Number: V7.1, January 2009.

484
 IONIC CONDUCTIVITY INVESTIGATION OF Ce1-xMxO2-x/2 (M=Sm, Gd, Nd, La) BASED
ELECTROLYTES FOR IT-SOFCs

G. Dönmez, V. Sarıboğa, M.A. F. Öksüzömer*, T.Gürkaynak Altınçekiç

Department of Chemical Engineering, Istanbul University, 34320 ,Istanbul, Turkey
efufu@istanbul.edu.tr

ABSTRACT
Ce1-xMxO2-x/2 (M= Sm, Gd, Nd, La) powders were successfully synthesized by polyol process. The thermal
behavior of powders were examined by thermogravimetry (TG) analysis method. The phase identification and
morphology of powders and pellets were investigated by X-ray diffraction (XRD) and scanning electron microscopy
(SEM). The results of the thermogravimetry/differential thermal analysis (TG/DTA) and XRD indicated that a
single-phase fluorite structure formed at the relatively low calcination temperature of 500 °C. The sintering
behavior and ionic conductivity of pellets prepared from calcined powders were also investigated. The results of
the impedance spectroscopy revealed that the SDC-20 sample exhibited the highest ionic conductivity of 4.29 S
m-1 at 800 °C in air.

1. INTRODUCTION

Solid oxide fuel cells (SOFCs) are attracting widespread attention due to their high conversion effiency and
low pollution. It is widely considered to be a clean power source for current technology. By far, the most
extensively studied electrolyte material is yttria-stabilized zirconia (YSZ). SOFCs with YSZ as oxygen ionic
conductor must operate above 800 °C to get satisfactory power output. However, such high temperatures
often lead to some problems such as solid reactions between the components, thermal degradation and
thermal expansion mismatch. It becomes increasingly important to reduce the operating temperature of the
SOFCs. Therefore, great efforts have been made to reduce the operating temperature. In order to reduce the
operation temperature from 1000 °C to 800 °C or lower, doped ceria has been considered as the solid
electrolyte for intermediate temperature solid oxide fuel cells [1,2]. To synthesize ceria-based electrolytes
with desired properties, various methods such as hydrothermal synthesis [3,4], homogenous precipitation
[5,6], the sol-gel process [7], the glycine-nitrate process [8], combustion [9], the reverse microemulsion
process [10] and a variant of the sol-gel process known as cation complexation [11] have been employed.

Among the several chemical methods to prepare metal particles, a solvothermal route has been considered
as a promising method to obtain monodispersed metal particles. In this procedure, polyalcohols with high
boiling points and ability to solve salts at elevated temperatures are utilized. To prepare metal particles, the
polyol process was used because of its mild reducing properties and simple route. In addition, polyol acts as
a solvent with chelating effect, which prevents agglomeration of particles during reaction. Powders obtained
from this method exhibit homogenous phase composition, narrow particle size distribution and highly
specific surface areas. Due to its simplicity and numerous advantages, the polyol method has received
considerable attention regarding the preparation of highly pure mixed oxides [12,13]. In this work, we aimed
to determine the influence of dopant type, dopant concentration and preparation method on the ionic
conductivity.

485
2. EXPERIMENTS AND CHARACTERIZATION TESTS

The powders of Ce1-xMxO2-x/2 (M= Sm, Gd, Nd, La) at varying dopant concentrations (x=0.05-0.25) were
synthesized by the polyol method. The starting materials were cerium (III) acetate hydrate, gadolinium (III)
acetate hydrate, samarium (III) acetate hydrate, lanthanum (III) acetate hydrate, neodymium (III) acetate
hydrate and triethylene glycol. The stoichiometric amounts of cerium (III) acetate hydrate and metal (III)
acetate hydrate were dissolved in triethylene glycol by keeping the total ion concentration at 0.2 M. The
solution was transferred to a three neck spherical glass flask equipped with a refluxing condenser and a
thermometer and placed on a magnetic stirrer. The solution was heated to 200 °C and kept at this
temperature for chemical synthesis. The resulting precipitate was cooled down to the room temperature,
seperated by centrifugation and washed three times with water and ethanol, respectively. The obtained
solids were dried at 70 °C in an oven.

The materials which have given the best ionic conductivity values in doped ceria series have been prepared
by pechini method to determine the effect of preparation method. In this process, metal acetate hydrates,
citric acid and ethylene glycol were used. The molar ratio of metal cation: citric acid was 1:2 and citric acid:
ethylene glycol was 1:4. Metal acetate powders and citric acid were dissolved in water. After homogenization
of this solution, ethylene glycol was added and the temperature was raised to 80 °C. The solution maintained
under stirring to remove the excess water and convert it to a gel. The thermal decomposition of this material
was carried out at 250 °C. Ash-like material was obtained.

All prepared amorphous solids decomposed to crystalline doped ceria powders by calcination at 500 °C for 4
h. The powder samples were pressed into pellets at 200 MPa with a cold isostatic press (CIP) and then
sintered at 1400 °C in air for 6 h.

The thermal behavior of the powders was studied with SII Exstar 6000 TG/DTA 6300 between 40-900°C with
-1
a constant heating rate of 10°C min under air. The results of the thermogravimetric analysis (TGA) as a
function of temperature were plotted.

X-ray powder diffractometer (XRD; Rigaku D/Max-2200 Ultima) with Cu Kα radiation (0.15418 nm) in an
angular region of 2θ=20-90° was used to identify the crystalline structure, phase purity and lattice
parameters of calcined powders. The average crystallite sizes of samples were calculated from X-ray line
broadening of the reflections of (111) by using the Scherrer equation.

0,9λ
D = β COSθ (1)

The lattice parameters of the samples were calculated by using the relations:

𝑎 = 𝑑√ℎ2 + 𝑘 2 + 𝑙 2 (2)

486
 λ
d = 2 sinθ (3)

where d is the planar spacing, λ the wavelenght of the radiation, θ the diffraction angle, and a is the lattice
parameter.

The ionic conductivities of Ce1-xMxO2-x/2 samples were measured by an AC Impedance analyzer (Solartron
1260 FRA and 1296 interface) in the temperature range of 300-800 °C in air. Arrhenius plots were
constructed and activation energies were calculated according to the following equation:
𝐸𝑎
𝜎 = 𝜎0 𝑒 − 𝑇 (4)

where Ea is the activation energy, T is absolute temperature and σ0 pre-exponential factor.

The densities of the sintered ceramics were measured by the Archimedes method. The theoretical density of
the sampIes were calculated by using the Eq.(5). Densification of samples were determined by dividing the
measured density to the theoretical density. In order to identify the morphology and microstructure of the
sintered samples, the scanning electron microscopy analysis was used.

x
4xMd +4(1−x)MCe +4(2− )MO
𝐷= 2
(5)
NA (a3 )

3. RESULTS AND DISCUSSIONS

The thermal decomposition behavior of Ce1-xMxO2-x/2 powders have shown that these materials which have

been prepared by polyol method lost their weight in three steps. In the first step, a weight loss occured due to

the dehydration and decomposition of acetate groups. The next step is related to the combustion of organic

residues and a minor weight loss is ascribed to the removal of residual triethylene glycol. No weight change

was detected above 500 °C on TG curve. This suggests the formation of crystalline structure. The TG curve

of SDC-15 sample was given in Fig. 1.

487
 40°C
99,6% 193°C
100
96,7%

95

90
TG %

85

229°C
80 500°C 890°C
%77
%75 %74,2
75

70
0 200 400 600 800
Temperature (°C)

Fig. 1: Thermal behavior of Ce0,85Sm0,15O1,92 powder prepared by polyol method.

The XRD patterns of pure ceria and Ce1-xSmxO2-x/2 materials (after calcination and sintering) was given in

Fig.2. According to the results of XRD, crystallization of the powders occured at 500 °C. These results

indicate that the products consist of single phase CeO2 with a cubic fluorite structure (JCPDS Card No: 34-

394). All the samples showed the presence of (111), (200), (220), (311), (222), (400), (331), (420) and (422)

reflection peaks in the range 20-90° of 2θ. No impurity peaks were observed. As seen in Fig. 2, the reflection

peaks became sharper and narrower after sintering at 1400 °C, which indicates that crystallite sizes increase

and crystallinity of powders becomes better.

488
Fig. 2: XRD pattern of a) pure ceria, b) SDC-after calcination at 500 °C and c) SDC-after sintering at 1400 °C.

5,48 GDC
SDC
5,47 NDC
LDC
Lattice parameter(°A)

5,46

5,45

5,44

5,43

5,42

5,41

5,40
0,00 0,05 0,10 0,15 0,20 0,25
Dopant concentration

Fig. 3: Lattice parameter- dopant concentration relation of Ce1-xMxO2-x/2 samples.

In Fig. 3 the lattice parameters of Ce1-xMxO2-x/2 powders were plotted as a function of dopant concentration.

As seen in Fig. 3, the lattice parameter of samples increased as long as the dopant concentration increased.

489
The materials prepared in this study exhibited a linear increase in the lattice parameter of the material, so it

is said that all samples obeyed the Vegard Rule [14,15].

Fig. 4: Impedance spectra of GDC-15 sample at 300°C, 500 °C and 800 °C.

The ionic conductivity measurement was performed in air in the temperature range of 300-800 °C. A
complex plane of real impedance versus imaginary impedance was prepared for each set of data. By curve
fitting a circle to semicircles on these plots, sample resistances were obtained. The total ionic conductivity
values were then calculated from the total resistance, cross-sectional area and thickness by using the
following equation:

𝑙
𝜎 = 𝐴.𝑅 (6)
𝑡𝑜𝑡𝑎𝑙

490
The complex impedance spectra plots, which were constructed for temperatures of 300 °C, 500 °C and
800°C in air, for Ce0,85Gd0,15O1,925 samples were shown in Figure 4. In this figure, two well-defined
semicircular arcs, which denote the grain interior and grain boundary, are discernible in 300 °C. With
increasing temperature to 500 °C, the first semicircle disappears and only the grain boundary and electrode
arcs are visible. When the temperature is increased more, grain interior and grain boundary resistances
become frequency dependent and only one semicircle is visible.

Fig. 5. Arrhenius plots of a) Ce1-xGdxO2-x/2, b) Ce1-xLaxO2-x/2, c) Ce1-xNdxO2-x/2 and d) Ce1-x SmxO2-x/2 samples.

The temperature dependence of the ionic conductivities of Ce1-xGdxO2-x/2, Ce1-xLaxO2-x/2, Ce1-xNdxO2-x/2 and

Ce1-xSmxO2-x/2 pellets were illustrated in Fig. 5. The maximum ionic conductivities and related activation

energy values were tabulated in Table 1. The crystallite sizes, lattice parameters and % relative densities of

these samples were also given.

491
Table 1. Microstructural and electrochemical properties for some of the prepared materials.

Ionic
Activation
Crystallite size Lattice % Relative conductivity,800°C
Material energy
-1
(nm) parameter (Å) Density (S m )
(eV)

-2
CeO2 42,8 5,409 95,8 1,45x10 0,83

SDC-20 9,7 5,432 93,6 4,29 0,72

GDC-15 10,9 5,417 89,2 4,04 0,67

NDC-15 9,2 5,44 93,9 3,35 0,73

LDC-15 7,3 5,459 95,7 2,95 0,74

Polyol method
0,04 Pechini method
Total conductivity (S cm-1)

800°C

0,03

0,02

0,01

0,00
LDC-15 NDC-15 GDC-15 SDC-20

Fig.6. Ionic conductivity graph of samples prepared by polyol and pechini method.

When the ionic conductivity results were examined, it was obtained that the samples with maximum

conductivity values were SDC-20, GDC-15, NDC-15 and LDC-15. To determine the effect of preparation

method, these materials were re-synthesized with pechini method. The same characterization tests were

made. As a result, the materials which were prepared by polyol method gave much higher ionic conductivity

values. The ionic conductivity results of the samples which were prepared by polyol and pechini method was

compared with a graph in Fig. 6. This graph clearly shows that doped ceria powders with adequate ionic

conductivity can be prepared through the polyol method at low temperatures.

492
4. CONCLUSIONS

Doped ceria Ce1-xMxO2-x/2 (M= Sm, Gd, Nd, La), (x= 0,05-0,25) solid solutions with fluorite structure were

prepared by polyol method. The results of X-ray diffraction showed that all powders calcined at 500 °C were

single phase with cubic fluorite structure. No peak which shows the presence of another secondary phase

was detected. It was calculated that the crystallite size of calcined powders was in the range of 6-15 nm for

all samples. After sintering the pellets at 1400 °C for 6 hr, the samples achieved over 90% of their

theoretical densities. Due to the XRD results after sintering, the lattice parameter of samples were calculated

by using the related equations. It was seen that as the dopant concentration increased, the lattice

parameters of samples increased linearly, too. This relation corresponds with Vegard Rule.

The ionic conductivity values of samples increased with the increasing temperature. The electrochemical

performance studies of doped ceria samples showed that among the electrolytes examined, the one doped
-1
with 20 mole % Sm had best ionic conductivity value of 4.29 S m . When compared with the other

electrolytes, the ionic conductivity values decreased in the order of GDC-15, NDC-15 and LDC-15. The

maximum ionic conductivities were observed in 0,15-0,2 composition range. The decrease of ionic

conductivity at higher concentrations was related to decrease of ionic mobility due to the formation of

complex defects.

The activation energies of samples were determined for both low temperature range (300-500 °C) and high

temperature range (500-800 °C). The calculated activation energy values in 500-800 °C (the intended

operating temperature for Intermediate Temperature Solid Oxide Fuel Cells) were in the range of 0,7-1,04

eV. These values are compatible with datas which are present in literature [16].

This paper discusses one of the first studies to use the polyol method to obtain electrolyte materials with

detailed conductivity results. In the next studies, we will try to improve SDC-20 electrolyte material, which

exhibited the best ionic conductivity value according to our study. In addition, we will operate these

electrolytes in fuel cell system to obtain the performance of electrolytes.

493
Acknowledgements:

This study was supported by Istanbul University through Project T.R. of The Turkish Prime Ministry State

Planning Organization (DPT) with Project Number of 2008K121000, and by The Research Fund of Istanbul

University with grant numbers 26803 and 27829.

REFERENCES

[1] Fu, Y.P., 2008, Microwave-induced combustion synthesis and ionic conductivity of Ce0.8(Gd0.2-xSmx)O1.9
ceramics, Ceramics International, 34, 2051–2057.

[2] Sha, X., Lu, Z., Huang, X., Miao, J., Jia, L., Xin, X., Su, W., 2006, Preparation and properties of rare earth
co-doped Ce0.8Sm0.2−xYxO1.9 electrolyte materials for SOFC, Journal of Alloys and Compounds,424,315–321.

[3] [8] M. Hirano, E. Kato, 1996, Hydrothermal synthesis of cerium (IV) oxide, Journal of American Ceramic
Society, 79(3),777-780.

[4] Y.C. Zhou, M.N. Rahman,1993, Hydrothermal synthesis and sintering of ultrafine CeO2 powders, Journal
of Material Research, 8,1680.

[5] N. Uekewa, M. Ueta, Y.J. Wu, K. Kakewa, 2002, Chemistry Letters, 8, 854.

[6] H.I. Chen, H.Y. Chang, 2004, Homogeneous precipitation of cerium dioxide nanoparticles in alcohol/water
mixed solvents, Colloids Surface A, 242, 61.

[7] G.S. Wu, T. Xie, X.Y. Yuan, B.C. Cheng, L.D. Zhang, 2004, An improved sol–gel template synthetic route
to large-scale CeO2 nanowires, Materials Research Bulletin, 39, 1023.

[8] C.R. Xia, M.L. Liu, 2002, Microstructures, conductivities, and electrochemical properties of Ce0.9Gd0.1O2
and GDC-Ni anodes for low-temperature SOFCs, Solid State Ionics, 152–423.

[9] M.M.A. Sekar, S.S. Manoharan, K.C. Patil, 1990, Combustion synthesis of fine-particle ceria, Journal of
Materials Science Letters, 9 ,1205.

[10] J. Chandradass, B. Nam, K.H. Kim, 2009, Fine tuning of gadolinium doped ceria electrolyte
nanoparticles via reverse microemulsion process, Colloids Surface A, 348,130-136.

[11] R.A. Rocha, E.N.S. Muccillo, 2003, Physical and chemical properties of nanosized powders of gadolinia-
doped ceria prepared by the cation complexation technique, Materials Research Bulletin, 38,1979-1986.

[12] A. Gondolini, E. Mercadelli, A. Sanson, S. Albonetti, L. Doubova, S. Boldrini, 2011, Microwave-assisted

synthesis of gadolinia-doped ceria powders for solid oxide fuel cells, Ceramics International, 37, 1423-1426.

[13] T. Karaca, T.G.Altınçekiç, M.F. Öksüzömer, 2010, Synthesis of nanocrystalline samarium doped
CeO2(SDC) powders as a solid electrolyte by using a simple solvothermal route, Ceramics International,
36,1101-1107.

[14] Stephens, I.E.L., Kilner, J.A., 2006, Ionic conductivity of Ce1−xNdxO2−x/2, Solid State Ionics, 177, 669–
676.

[15] Fu, Y.P., Chen, S.H., 2010, Preparation and characterization of neodymium-doped ceria electrolyte
materials for solid oxide fuel cells, Ceramics International, 36, 483–490.

[16] C. Veranitisagul, A. Kaewvilai, W. Wattanathana, N. Koonsaeng, E. Traversa, A. Laobuthee, 2012,

Electrolyte materials for solid oxide fuel cells derived from metal complexes: Gadolinia-doped ceria,
Ceramics International, 38, 2403-2409.

494
 ENHANCED BIOHYDROGEN AND 5-ALA PRODUCTIONS WITH THE OPTIMIZATION OF
CULTURE CONDITIONS

G. Kars*, U. Alparslan
Department of Biology, Faculty of Science, Selçuk University, 42075, Konya, Turkey
Advanced Technology Research and Application Center, Selçuk University, 42075, Konya, Turkey
gkars2004@yahoo.com, u.alparslan@windowslive.com

Abstract
In a previous study, the production of biohydrogen and 5-aminolevulinic acid (5-ALA) by Rhodobacter
sphaeroides O.U.001 using waste barley hydrolysate was investigated within a biorefinery concept. In this
study, the optimization of culture conditions for enhanced biohydrogen and 5-ALA productions was done
using sugar beet molasses. In this context, effects of sugar concentrations and additions of the elements
(molybdenum, iron) and the vitamins (Thiamin, Biotin and Nicotinic acid) on biohydrogen and 5-ALA
productions were examined. The molasses was used as carbon source in the following concentrations 2.76
g/L, 6.89 g/L, 13.78 g/L, 20.66 g/L and 27.50 g/L. Increasing sugar concentration led to an increase in both
biohydrogen and 5-ALA productions in media where elements and vitamins were added. The maximal
biohydrogen and 5-ALA productions were observed in 27.50 g/L sugar-containing medium as 1.42 L H2/L
culture and 23.3 mM respectively. Higher sugar concentrations also resulted in higher cell masses such that
the cells reached the maximal optical density in 27.50 g/L sugar-containing medium (OD660= 8.3). However,
there were little or no biohydrogen and 5-ALA productions when the elements and vitamins were not added
into the media showing that they are absolutely necessary for the biohydrogen and 5-ALA productions. As a
result, the most suitable condition for biohydrogen and 5-ALA productions was obtained when the sugar
concentration was 27.50 g/L with the additions of elements and vitamins.

Keywords: Biohydrogen, Amionolevulinic acid, Rhodobacter sphaeroides, Sugar beet molasses, Biorefinery
concept

1. INTRODUCTION
Energy is of great importance for the continuation of modern human life. And, energy sources are needed to
be diversified in order to eliminate possible energy crisis in the future. One of the alternative approaches is
the hydrogen. It could be produced by biological and non-biological means (Uyar et al., 2012). Among
biological hydrogen production methods, hydrogen production via photofermentation takes special attention.
Purple non-sulfur (PNS) bacteria such as Rhodobacter sphaeroides, Rhodobacter capsulatus,
Rhodospirullum rubrum and Rhodopseudomonas palustris have been used widely for biohydrogen
production (Kars and Gündüz, 2010; Uyar et al., 2012). Biohydrogen production occurs mainly by
nitrogenase enzymes under anaerobic conditions in these bacteria. Moreover, light is required to drive the
electron transfer through the electron transport chain for ATP production. Eventually, ATPs are used by
nitrogenase enzymes for hydrogen production. A wide variety of carbon sources could potentially be used by
these bacteria such as organic acids, various feedstocks and dark fermentation effluent (Kapdan and Kargı,
2006). In a study done by Özgür et al., 13.7 mol H2/mol sucrose was produced in a sequential dark and
photofermentation process corresponding to 57% of the theoretical yield of 24 mol of H2/mol of sucrose
(Özgür et al., 2010). Moreover, genetically engineered strains were used to obtain an improved hydrogen
production. Elimination of uptake hydrogenase, developing ammonium insensitive strains, inhibition of poly-
3-hydroxybutyric acid synthesis and heterologous expression of hydrogen evolving hydrogenases in PNS
bacteria are examples of the genetic approaches (Kars et al., 2008; Kars et al., 2009; Kars and Gündüz,
2010). Despite these efforts, current biohydrogen production efficiencies are still behind practical
applications. Therefore, cost efficient bioprocesses with high hydrogen production efficiencies are needed to
be developed.

In addition to hydrogen production capability of PNS bacteria, they could also produce several high added
value products such as vitamin B12, poly-hydroxy butyrate, and 5-aminolevulinic acid (5-ALA) (Sasaki et al.,
2005 and Yiğit et al., 1999). 5-ALA which takes a role in tetrapyrrole biosynthesis is produced via C4
pathway in Rhodobacter sphaeroides. It is a natural photodynamic compound used as biodegradable
herbicide, insecticide, growth promoter and anti-cancer agent (Sasaki et al., 2002). The ability of PNS
bacteria to produce these value-added products is probably the result of their great metabolic versatility. The
production of more than one product in a single bioprocess which was defined as biorefinery concept may
increase the cost efficiency of the process. Hence, a more economical bioprocess could be obtained. The
use of low cost feedstocks may further increase the efficiency of the process. In conclusion, since the

495
feasibility of biohydrogen production alone is low, the bioprocesses within a biorefinery concept are needed
to be developed. In a previous study, the production of biohydrogen and 5-ALA by Rhodobacter sphaeroides
O.U.001 using waste barley hydrolysate was investigated within a biorefinery concept (Kars and Ceylan,
2013). And, comparable results were obtained. In this study, the optimization of culture conditions for
enhanced biohydrogen and 5-ALA productions was performed using sugar beet molasses. In this context,
effects of sugar concentrations and additions of elements (molybdenum, iron) and vitamins (Thiamin, Biotin
and Nicotinic acid) on biohydrogen and 5-ALA were investigated. As a result, an optimum culture condition
regarding these factors was configured. In this way, a unique and more feasible bioprocess was obtained
using a cost-effective substrate in the context of biorefinery. To our knowledge, this is the first study in which
valorization of sugar beet molasses was performed for the production of biohydrogen and 5-amino levulinic
acid by a photosynthetic bacterium within a biorefinery concept. The results are stimulating for further studies
contributing to the literature in this way. And, industrial scale productions will be more probable after these
improvements.

2. MATERIALS AND METHODS

2.1. Bacterial strain and general culture conditions

Rhodobacter sphaeroides O.U.001 (DSM 5864) was used in this study. It was maintained regularly in Biebl
and Pfennig minimal medium where D,L-malate and L-glutamate (15 mM/2 mM) were used as carbon and
nitrogen sources respectively (Biebl and Pfennig, 1981). Media were solidified using 1.5 % agar when
needed. As a general rule, 10 % fresh inoculations have been done into 55 mL glass bioreactors. The
anaerobic cultures were prepared by argon flushing for a few minutes and incubated at 29 ºC under the
2
illumination of 940 µE/m /s by 100 W incandescent bulbs. The produced gas was accumulated into water
filled graduated tubes and gas composition analysis was done by gas chromatography.

2.2. Biomass and culture conditions for biohydrogen and 5-ALA production
Sugar beet molasses obtained from a sugar factory (Konya Şeker) has been used as carbon source for both
hydrogen and 5-ALA production. Molasses is the by-product of sugar production processes and contains
about 50 % sucrose. And, it can be used as a valuable substrate for microorganisms due to high sugar
content. Since it is the viscous dark brown by-product of the refining of sugar beets, it is first filtrated and
then diluted with water to be made suitable for photofermentation by R. sphaeroides. 2 mM glutamate was
used as nitrogen source in the cultures. In addition to carbon and nitrogen sources, 50 μL of iron solution
(13.2 mM FeSO4), 1 mL of molybdenum solution (0.83 mM Na2MoO4.2H2O) and 2.5 μL of vitamin solution
(Thiamine 500 mg/L, Niacin 500 mg/L and Biotin 15 mg/L) were added into the growth media. The growth
media were buffered to pH 6.9 using phosphate buffer. Five different sugar concentrations were examined in
ten batch bioprocesses (Table 1). Each bioprocess was studied in duplicate.

Table 1. The composition of media for biohydrogen and 5-ALA production.

Sugar Vitamin Glutamat 5-ALA Light
(Fe + Mo) Anaerobic
(g/L) solution (2 mM) (15 mM) (3 klux)
+ +
2.76 + + + +
- -
+ +
6.89 + + + +
- -
+ +
13.78 + + + +
- -
+ +
20.66 + + + +
- -
+ +
27.50 + + + +
- -

After 10 % fresh inoculation, the cultures were flushed with argon gas to make an anaerobic environment.
Then, they were placed under 100 W incandescent bulbs. At certain time intervals, samples were taken to
measure pH and cell density of the cultures. Total hydrogen production was also measured by a water
displacement method. Hydrogen content of the total gas was determined by gas chromatography.

2.3. Measurement of the concentration of 5-ALA in the cultures

496
Levulinic acid addition into the growth media was told to enhance 5-ALA production (Sasaki et al., 1990).
However, it should be added into the cultures at the late logarithmic phase of growth in order not to interfere
with the cell growth and hydrogen production. Therefore, levulinic acid was added into the cultures at around
th
90 hour. The final concentration was 15 mM. After the bioprocess was terminated, cells were separated
from the liquid part of the cultures and the amount of 5-ALA was measured. The extracellular amount of 5-
ALA was determined with the spectrophotometric method (Mauzerall and Granick, 1956). For this, a
standard curve was plotted with the known concentrations of 5-ALA and then concentrations of unknown
samples were calculated by interpolation on the graph.

3. RESULTS AND DISCUSSION

3.1. Growth of Rhodobacter sphaeroides O.U.001

Ten different growth media have been prepared as shown in Table 1 using molasses as carbon source. A
total volume of 50 mL was obtained after 10 % fresh inoculation in 55 mL glass bioreactors. The pH of the
cultures was initially set to 6.9 and was not controlled during the growth. The pH changes were monitored
during the bioprocess by taking samples at certain time intervals and the results were illustrated in Fig. 1A
and 1B. pH of the cultures was initially at around 7. After rising up to 9, the pH values were decreased to 7 in
media which contained elements and vitamins (Fig. 1A). Similar pH values were also observed in the
cultures where additional elements and vitamins were not put into (Fig. 1B). In the case of cell densities,
increasing sugar concentration led to an increase in cell density (Fig. 1C and 1D). However, the increase
was less pronounced at the higher sugar concentrations (13.78 g/L, 20.66 g/L and 27.50 g/L) indicating a
saturation level. The maximal optical cell density was obtained in the 27.50 g/L sugar-containing medium
with element and vitamin additions (OD660= 8.3).
10 10
2.76 g/L 6.89 g/L 2.76 g/L 6.89 g/L
9 13.78 g/L 20.66 g/L 9 13.78 g/L 20.66 g/L
27.5 g/L 27.5 g/L
8 8
7 7
pH

pH

6 6
5 5
4 4
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300
Time (h) Time (h)
(A) (B)

10 10
2.76 g/L 2.76 g/L
6.89 g/L 6.89 g/L
8 13.78 g/L 8 13.78 g/L
20.66 g/L 20.66 g/L
OD 660nm

OD 660nm

6 27.5 g/L 6 27.5 g/L

4 4

2 2

0 0
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300
Time (h) Time (h)

(C) (D)

Figure 1. pH and cell densities of the cultures were monitored through the bioprocesses. In (A) and (C), data
were obtained with the additions of elements and vitamins. In (B) and (D), data were obtained without
element and vitamin additions.

3.2. Biohydrogen production at various sugar concentrations

The produced gas was collected into the water-filled tubes and the amount was measured by a water
displacement method. Increased sugar concentration resulted in a higher hydrogen accumulation (Fig. 2). In
particular, the highest hydrogen production was observed in the 27.50 g/L sugar-containing culture medium

497
with the additions of elements and vitamins (1.42 L H2/L culture). Removing elements and vitamins from the
media severely affected hydrogen production especially at lower concentrations.

80 40
2.76 g/L 2.76 g/L
70 35
Total hydrogenı (mL)

6.89 g/L 6.89 g/L

Total hydrogen (mL)

60 30
13.78 g/L 13.78 g/L
50 20.66 g/L 25 20.66 g/L
40 27.5 g/L 20 27.5 g/L
30 15
20 10
10 5
0 0
0 50 100 150 200 250 300 350 400 450 500 550 0 50 100 150 200 250 300 350 400
(A) Time (h) (B) Time (h)

Figure 2. Hydrogen production at various sugar concentrations. Data in (A) were obtained with the additions
of elements and vitamins. And, data in (B) were obtained without element and vitamin additions.

3.3. 5-aminolevulinic acid production

In general increasing sugar concentrations resulted in a rise in production of 5-ALA. Consequently, the
highest amount was obtained in 27.50 g/L sugar-containing culture medium (23.34 mM). Similar to hydrogen
production, diminishing elements and vitamins severely affected 5-ALA production. The amount of 5-ALA
produced in each culture was given in Table 2.

Table 2. The amount of 5-ALA produced in cultures with different compositions.

The amount of 5-ALA (mM)
Sugar concentration (g/L) With elements and vitamins Without elements and vitamins
2.76 0 0

6.89 0.83 0.83

13.78 1.11 14.75

20.66 5.41 0

27.50 23.34 0

4. CONCLUSIONS
Various sugar concentrations were tested in the presence and absence of elements and vitamins for
biohydrogen and 5-ALA productions. And, the impact of these factors was clearly observed such that
elimination of elements and vitamins severely affected biohydrogen and 5-ALA productions. On the other hand,
increasing sugar concentration had a positive effect. The maximal biohydrogen and 5-ALA productions were
obtained in 27.50 g/L sugar-containing culture medium, 1.42 L H2/L culture and 23.34 mM respectively. These
results significantly contribute to the development of a cost efficient bioprocess within a biorefinery concept.

Acknowledgement
This study has been supported by The Scientific and Technological Research Council of Turkey with the
project number: 111T523 and Selçuk University Scientific Research Fund with the project number:11401112.

498
REFERENCES

Biebl H. and N. Pfennig. 1981. Isolation of member of the family Rhodosprillaceae. In The prokaryotes, vol.
1, edited by M. P. Starr, H. Stolp, H. G. Trüper, A. Balows and H. G. Schlegel, Springer.

I. K. Kapdan and F. Kargi. 2006. Bio-hydrogen production from waste materials. Enzyme Microb Tech
38:569-82.

Kars G, Gündüz U, Rakhely G, Yücel M, Eroğlu İ, Kovacs L. K. 2008. Improved hydrogen production by
hydrogenase deficient mutant strain of Rhodobacter sphaeroides O.U.001. Int J Hydrogen Energy
33(12):3056-60.

Kars G, Gündüz U, Yücel M, Rakhely G, Kovacs K, Eroğlu İ. 2009. Evaluation of hydrogen production by
Rhodobacter sphaeroides O.U.001 and its hupSL deficient mutant using acetate and malate as carbon
sources. Int J Hydrogen Energy 34:2184-90.

Kars G. and U. Gündüz. 2010. Towards a super H2 producer: Improvements in photofermentative

biohydrogen production by genetic manipulations. Int J Hydrogen Energy 35:6646-56.

Kars G. and A. Ceylan. 2013. Biohydrogen and 5-aminolevulinic acid production from waste barley by
Rhodobacter sphaeroides O.U.001 in a biorefinery concept. Int J Hydrogen Energy 38: 5573-9.

Mauzerall D. and S. Granick. 1956. The occurrence and determination of δ-aminolevulinic acid and
porphobilinogen in urine. J Biol Chem 219:435-46.

Özgür E., A. E. Mars, B. Peksel, A. Louwerse, M. Yücel, U. Gündüz, P. A. M. Claassen and İ. Eroğlu. 2010.
Biohydrogen production from beet molasses by sequential dark and photofermentation. Int J Hydrogen
Energy 35:511-7.

Sasaki K., T. Tanaka, Y. Nishizawa and M. Hayashi. 1990. Production of a herbicide, 5-aminolevulinic acid,
by Rhodobacter sphaeroides using the effluent waste from an anaerobic digestor. Appl Microbiol Biot
32:727-31. Sasaki K., M. Watanabe, T. Tanaka, T. Tanaka. 2002. Biosynthesis, biotechnological production
and applications of 5-aminolevulinic acid. Appl Microbiol Biot 58:23-9.

Sasaki K., M. Watanabe, Y. Suda, A. Ishizuka and N. Noparatnaraporn. 2005. Applications of photosynthetic
bacteria for medical fields. J Biosci Bioeng 10(5):481-8.

Uyar B., G. Kars, İ. Eroğlu, M. Yücel and U. Gündüz. 2012. Hydrogen Production via Photofermentation. In
State of the art and progress in production of biohydrogen, edited by N. Azbar and D. B. Levin. Bentham
Science Publishers.

Yiğit D. Ö., U. Gündüz, L. Türker, M. Yücel and İ. Eroğlu. 1999. Identification of by-products in hydrogen
producing bacteria; Rhodobacter sphaeroides O.U. 001 grown in the waste water of a sugar refinery. J
Biotechnol 70:125-31.

499
OPTIMIZATION OF THE OPERATION CONDITIONS IN A DIRECT BOROHYDRIDE FUEL
CELL WITH CARBON SUPPORTED Au ANODE

F. G. Boyaci San, O. Okur, Ç. Iyigün Karadag, I.Işik Gulsac, E. Okumus

TÜBİTAK Marmara Research Center Energy Institute P.O Box 21, 41470 Gebze/Kocaeli/Türkiye
Fatmagul.boyaci@tubitak.gov.tr

ABSTRACT
In this study, response surface methodology (RSM) based on a five variable central composite design
(CCD) is employed for the optimization of the direct borohydride fuel cell (DBFC) operation conditions,
where the variables are the catalyst loading, cell temperature, borohydride concentration, flow rates of fuel
and oxidant. The main effects, quadratic effects and interactions of the five variables on the power density
of fuel cell are investigated by the analysis of variance. The carbon supported Au as the anode catalyst is
used to control the hydrolysis of sodium borohydride under operating conditions. The results showed that
the catalyst loading is the most significant factor on the power density. Under the operation conditions of
0.52 mg/cm2 catalyst loading, 80°C fuel cell temperature, 3.28 L/min fuel and 0.27 L/min oxidant flow rate
and 1.5 M borohydride concentration, the maximum power density of 37.15 mW/cm2 is obtained.

Key words: direct sodium borohydride, fuel cell, oxidation, operation condition, response surface
methodology, optimization

INTRODUCTION
Direct borohydride fuel cells (DBFC) are considered to be the most efficient and clean power
generation devices for portable and mobile applications because of their high energy densities and
theoretical working potentials. Usually, a DBFC employs an alkaline solution of sodium borohydride as
fuel and oxygen or air as oxidant. Performance of a DBFC single cell depends on a large variety of
factors such as materials for anode, cathode, fluid flow field, and operational parameters such as
anolyte composition, flow rates of fuel and oxidant, humidity of oxidant, temperature, etc. (Nagle et al.,
2009; Celik et al.,2009; Ma et al., 2010; Celik et al., 2010;).

The oxidation of sodium borohydride is based on a complete eight electron reaction process and
provides a high theoretical specific capacity (5.67 Ah/g, based on NaBH4) (Kim et al., 2004).
− −
BH4 + 8OH → BO2 + 6H2O+8e E = −1.24 V vs SHE
- - 0
(1)

However, the actual number of electrons released is typically less than the theoretical value, due to
the parallel unwanted reaction of hydrolysis, which competes with the oxidation:
- -
BH4 + H2O → BO2 + 4H2 (2)
−
Unlike Pt, electrode materials such as Au and Ag showed little or no activity towards the BH4
hydrolysis reaction (Lam et al., 2009; Pei et al. 2010). As a consequence, the coulombic numbers near
−
8 were reported for the electrooxidation reaction of BH4 (Amendola et al., 1999). But these electrode
materials usually demonstrated slow electrode kinetics and low power density. Gyenge et al.
suggested that alloying Ag with Ir and Pt could effectively improve the electrode kinetics (Gyenge et
al., 2006).

One way to improve the cell performance was reported as choosing a metal with high electocatalytic
activity [8]. Kim et al. assembled the DBFCs using carbon supported Pt or unsupported Pt catalyst. It
2
was reported that the maximum power density could only reach to 44.2 mW/cm (Kim et al., 2004).
Gyenge et al. reported that Pt-Ni and Pt-Ir exhibited higher electrocatalytic activity than Pt giving a cell
2
voltage of 0.53V at 100 mA/cm (Gyenge et al., 2006). However, the maximum power density around
2
50 mW/cm was still lower than that of the polymer electrolyte membrane fuel cell. Liu et al. studied
the influence of hydrogen evolution on the cell performance. They found that the reduction of the
effective reaction area caused by hydrogen evolution during the DBFC operation was a more critical
problem (Liu et al., 2008). They suggested that the addition of Nafion in anode and using a composite
of Ni, metal hydride and Pd/C as the anode catalyst improved the cell and stack performance of the
DBFC by suppressing the hydrogen evolution (Liu et al., 2006; Li et al., 2008).

500
Suppression of hydrogen evolution during operation was one of the solutions to improve the cell
performance. Through the anolyte flow observation in a transparent cell, Park and Liu confirmed that
hydrogen bubbles blocked channels and thus hindered the anolyte flow (Park et al., 2006; Liu et al.,
2008). They investigated the effect of the anode diffusion layer on the cell performance and stability of
the DBFC. It was found that the thinner the anode diffusion layer, the higher was the performance. Liu
et al. tried to increase anode porosity by pasting anode ink into a piece of Ni foam. It was found the
2
cell performance was improved from 50 to 180 mW/cm (Liu et al., 2008). Kim et al. tried to improve
the stack performance by decreasing the hydrogen evolution through suitable anode design and the
flow field improvement (Kim et al., 2008).

The borohydride cross over was reported as another problem influencing cathode performance (Liu et
al., 2007). It was found that the applied current during the cell operation would influence the fuel cross-
over. The Pt cathodes were found to have better performance stability than the Ag cathodes. The cell
stability within hours was found to be greatly affected by the mass transport properties of different
components. The porosities of the anode and cathode supporting materials, the membrane
pretreatment, and borohydride concentration had significant influence on the performance stability (Liu
et al., 2008).

To the best of our knowledge, no study was reported on the optimization of the DBFC operation
conditions by a statistical approach. In this work, the carbon supported Au as the anode was used to
control the hydrolysis of sodium borohydride under operating conditions. Response surface
methodology (RSM) based on a five variable central composite design (CCD) was employed for the
optimization of DBFC operation conditions The five variables considered were the catalyst loading, cell
temperature, borohydride concentration, flow rates of fuel and oxygen as oxidant. The main effects,
quadratic effects and interactions of the five variables on the power density of fuel cell were
investigated by the analysis of variance. The results showed that, the catalyst loading was the most
significant factor on the power density. The quadratic model was chosen to realize the response of
borohydride concentration and interaction of catalyst loading with borohydride concentration.
Predicted values from the regression equations were found to be in good agreement with observed
values. According to the results, the regression equations could be used to predict and optimize the
2
performance of DBFC. Under the operation conditions of 0.52 mg/cm catalyst loading, 80°C fuel cell
temperature, 3.28 L/min fuel and 0.27 L/min oxidant flow rate and 1.5 M borohydride concentration,
2
the maximum power density of 37.15 mW/cm is obtained.

EXPERIMENTAL
Materials and chemicals
The chemical reagents, including 2-propanol, NaOH, NaBH4, and 5 wt. % Nafion solution (Sigma-
Aldrich), Nafion 117 membrane (DuPont, USA), 20 wt. % Pt/C on Vulcan XC-72 carbon support (BASF
Fuel Cell, Inc.) are all purchased. Homemade 10 wt. % Au/C electrocatalyst was used in the electrode
preparation.

Experimental design
A CCD with five factors at five levels was performed to ascribe the effect of factors on response
surface in the region of investigation. Investigated factors which were considered to have the greatest
influence on fuel cell performance in literature were cell temperature, borohydride concentrations,
catalyst loading, anode and cathode feeding rate. The CCD was applied using Design Expert 6.0 (trial
k
version). The total number of experiments with five factors was 50 (= 2 + 2k + 6), where k is the
number of factors. During our study we used ½ fraction design instead of full factorial design with 26
experiments. The first six columns of Table 1 show run number and experimental conditions of the
runs arranged by CCD. The fuel cell performance of the process was evaluated by analyzing the
response.

Table 1. Experimental data

Run Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Response
A:Cell B:Anode C:Cathode D:Borohydride E:Catalyst Power

501
 temperature flow flow rate concentration loading density
2 2
(C) rate(ml/min) (L/min) (M) (mg/cm ) (mW/cm )
1 -1 1 -1 -1 -1 14.13
2 1 -1 -1 1 1 6.34
3 0 0 0 -1 0 7.05
4 -1 -1 1 1 1 1.41
5 -1 1 1 1 -1 23.32
6 0 0 0 0 -1 26.84
7 -1 1 1 -1 1 0.7
8 0 -1 -1 -0 0 7.05
9 1 1 1 -1 -1 24.72
10 -1 -1 -1 -1 1 0.7
11 1 -1 -1 -1 0 21.9
12 -1 1 1 -1 -1 12.71
13 1 0 0 0 0 14.1
14 -1 1 -1 1 1 0.7
15 -1 -1 -1 1 -1 27.42
16 0 0 0 0 1 2.12
17 1 1 -1 1 -1 36.74
18 0 0 1 0 0 7.76
19 1 -1 1 1 -1 38.16
20 -1 0 0 0 0 2.83
21 1 1 1 1 1 7.02
22 1 1 -1 -1 1 2.11
23 1 -1 1 -1 1 2.82
24 0 0 0 1 0 11.27
25 0 -1 0 0 0 7.05
26 0 1 0 0 0 7.05

Fuel cell performance tests

The anode electrocatalyst ink was prepared by mixing 2- propanol with 5 wt. % Nafion solution and 10
wt. % Au/C. Then, the ink was coated on a carbon cloth (Electrochem, Inc.), yielding a metal loading
2
mass on the electrode of 0.52 mg/cm . The Nafion 117 membrane was cleaned by boiling in 3 wt. %
H2O2 and 3 wt. % H2SO4 for 1 h, followed by boiling in ultrapure water for 2 h. The cleaned membrane
was activated in 2 M NaOH solution for 1 h prior to use. The anode/membrane/cathode unit was
compressed between two graphite blocks with pin type flow fields. Silicon gaskets were assembled
2
between the electrode and the graphite block. The active area of the fuel cell was 25 cm . Cell
2
performance was tested against a 1 mg/cm Pt/C cathode coated on carbon paper. Peristaltic pumps
were used to feed the fresh anolyte (an aqueous solution of NaBH4 and 4M NaOH) at selected flow
rate. The flow rate of oxygen as an oxidant at the cathode chamber was 0.2 L/min. The oxidant was
humidified by passing through a bubbler at 65°C. Cell performance data were obtained using an
electrochemical fuel cell test system (Electrochem 400 W) and a computer controlled E-load system
(ECL 150).

RESULTS AND DISCUSSION

A comparative investigation of DBFC polarization behavior was carried out relating to the cell
temperature, borohydride concentration, flow rates of fuel and oxidant. The higher and lower level
values were used in a statistical analysis to determine significant parameters that are given in Table 2.

Table 2. Factors and levels of input factors for experimental design

Input Factors Low Mean High
o
Cell temperature, C 50 65 80
Borohydride
0.5 1.0 1.5
concentration, M
Anode flow rate,
/ 1.0 3.0 5.0
ml min

502
 Input Factors Low Mean High
Cathode flow rate,
/ 0.1 0.2 0.4
L min
Catalyst loading,
2 0.52 0.67 0.82
mg/cm

Polarization graph for different runs are given in Fig. 1. According to Fig. 1 the highest performance
o
was obtained at Run 19 (80 C, 1.5 M borohydride concentration, 1 ml/min anode, 0.4 L/min flow rates
2
and 0.52mg/cm catalyst loading).

45
40
35
30
mWcm-2

25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Run number

Fig.1. Run number versus power density

The data in Table 1 are run through the RSM to construct an empirical model for the representation of
power density in the terms of independent parameters. Based on the regression analysis at 95% of the
confidence interval, the lack of fit error and p-values of parameter estimations are found to be
significant. The value of p < 0.0001 indicates that there is only a 0.01% chance that a “model F-value”
this large could occur due to noise (Silva and Rouboa, 2012). The quadratic model is used to fit the
observed data by the least squares analysis.

Some of the model terms were found to be significant. The insignificant model terms can be removed
and may result in an improved model. Obtained from Analysis of variance (ANOVA) result that
temperature, borohydride concentration and catalyst loading have substantial influences on cell
performance. By selecting the backward elimination procedure to automatically reduce the terms that
are not significant, the resulting ANOVA tables for the reduced quadratic model is shown in Table 3.

Table 3. ANOVA results

Sum of Mean F
Source Squares DF Square Value p Value
Model 3137 6 522.88 212.65 < 0.0001
A 272.14 1 272.14 110.68 < 0.0001
D 238.64 1 238.64 97.05 < 0.0001
E 2267.34 1 2267.34 922.09 < 0.0001
E2 190.26 1 190.26 77.38 < 0.0001
AE 53.14 1 53.14 21.61 0.0002
DE 115.78 1 115.78 47.08 < 0.0001
Residual 46.72 19 2.46
Cor Total 3184.02 25

2
Std. Dev. 1.57 R 0,9853

503
 2
Mean 12.08 Adj R 0,9807
2
C.V. 12.98 Pred R 0,9713
PRESS 91.24 Adeq Precision 46.096
A: Cell temperature, B: Anode flow rate, D:Borohydride concentration, E:Catalyst loading
2
Results show that the models are still significant. The predicted R is in reasonable agreement with the
2 2
adjusted R . The adjusted R value is particularly useful when comparing models with different number
of terms. Adequate precision compares the range of the predicted values at the design points to the
average prediction error. Ratios greater than 4 indicate adequate model discrimination. The following
equations are the final empirical models for the power density.

Power Density =+91.64873 + 0.80192 * Cell Temperature + 31.31289 * Borohydride

Concentration - 335.36728 * Catalyst Loading + 260.49383 * Catalyst Loading2 -
0.81000 * Cell Temperature * Catalyst Loading - 35.86667 * Borohydride
Concentration * Catalyst Loading (3)

The Fig. 2 verifies that the predicted data of the response from the empirical models are in agreement
with obtained ones in the range of the operating variables. The F value, model mean square divided
by the error mean square, is considerably high meaning that this model predicts fit the experimental
2
data. Furthermore, the high value of adjusted R indicated that the model fits the observed data well.

Fig. 2. Predicted data versus actual data

In Fig. 3, it is obtained that by increasing catalyst loading of electrodes, the power density decreases
linearly up to a certain value (0.7) and then it becomes constant. By increasing catalyst loading, the
thickness of catalyst layer on gas diffusion layer (GDL) surface increases, hence, the active sides of
catalyst particles are plugged by the other catalyst particles in upper layer. So, the reaction rate
decreases and power density becomes constant. At low catalyst loading, by the reaction rate and
power density increase with the increase in temperature, as shown in the same figure, since the
activities of catalysts and borohydride increase at high cell temperature. However, at high catalyst
loading, although the cell temperature increases the reaction rate does not change too much, because
-
of the mass transfer limitation. At low catalyst loading, the activation of BH4 ion is more effective by
temperature and this effect cannot be observed significantly at high loading due to the blocked active
sides.

504
Fig. 3. The effect of cell temperature and catalyst loading on power density

The negative effect of catalyst loading on power density is also observed in Fig. 4 as in Fig. 3. As
catalyst loading of electrodes increases, the thickness of catalyst layer increases and therefore the
mass transfer limitation of ions to active sites also increase. On the other hand, as borohydride
concentration increases at low loading, the power density increases. Since, the number of borohydride
ions per one active side of catalyst is increased, the reaction rate increases. However, at high catalyst
loading, although borohydride concentration increases, the reaction rate is not affected because of
plugged active sites and high mass transfer limitations. So, if the catalyst loading in electrodes
increases, the membrane electrode assembly preparation cost also increases and the cell
performance decreases.

Fig. 4. The effect of catalyst loading and borohydride concentration on power density

Optimization
Two cases were determined for optimization study. The first case is used to maximize the power
density. As shown in Table 4; while temperature, anode and cathode feed flow rate, borohydride
concentration and catalyst loading were kept constant in the range, the power density was tried to
maximize and its importance was accepted as maximum. By using the program, five optimum points
were obtained (Table 5).

505
Table 4. Optimization study (first case)
Operation conditions Value Importance

Cell temperature in range +++

Anode flow rate in range +++

Cathode flow rate in range +++

Borohydride concentration in range +++

Catalyst loading in range +++

Power density maximum +++++

2
The maximum power density is 37.15 mW/cm for 5 points. The operating temperature was obtained
o
at 80 C for 4 points. The main difference between 1 and 4 points were temperature, the flow rates of
o
anode and cathode. In order to obtain the same power density value at 40 C; the anode flow rate
should be increased from 3.45 to 4.97 and the cathode feed flow rate should be decreased from 0.38
to 0.2. The schematic representation of first optimum point can be seen from Fig. 5.

Table 5. Five optimum points

Optimization Cell Anode flow Cathode Borohydride Catalyst Power
points temperature rate flow rate concentration loading density
o -2
( C) (mL/min) (L/min) (M in 4 M (mg (mWcm )
2
NaOH) Au/cm )
1 80.00 3.45 0.38 1.50 0.52 37.1464
2 80.00 2.90 0.34 1.50 0.52 37.1464
3 80.00 1.86 0.30 1.50 0.52 37.1463
4 40.00 4.97 0.20 1.50 0.52 37.1463
5 80.00 4.67 0.21 1.50 0.52 37.1463

Fig. 5. The schematic representation of first optimum case

Second optimum case was obtained by minimizing operation conditions and by maximizing power
density. The importance of operation conditions was given in Table 6.

506
Table 6. Optimization study (second case)
Operation conditions Value Importance
Cell temperature Minimum +++++
Anode flow rate Minimum +++++
Cathode flow rate Minimum +++++
Borohydride concentration Minimum +++++
Catalyst loading Minimum +++++
Power density Maximum +++++

2
In Table 7, the maximum power density was obtained as 14 mW/cm by considering minimum
operating conditions (cell temperature: 50 °C, anode flow rate: 1.0 mL/min, cathode flow rate: 0.1
2
L/min., borohydride concentration: 0.57 M in 4 M NaOH and catalyst loading: 0.52 mg Au/cm ).
However, the power density for second case was 3 times lower than that of first case. The schematic
representation of second optimum case was given in Fig. 6.

Table 7. Maximum power density by considering minimum operating conditions

Optimization Cell Anode flow Cathode Borohydride Catalyst Power
points Temperature rate flow rate concentration loading density
o 2
( C) (mL/min) (L/min) (M in 4 M (mg (mW/cm )
2
NaOH) Au/cm )
1 50.02 1.00 0.10 0.57 0.52 13.92
2 51.19 1.00 0.10 0.53 0.52 13.90
3 50.10 1.00 0.10 0.61 0.52 14.55
4 50.00 1.00 0.10 0.50 0.52 13.06
5 50.65 1.00 0.10 0.58 0.52 14.29

Fig. 6. The schematic representation of second optimization case

CONCLUSION
DBFC operation conditions were optimized with central composite method of the Design Expert
software. The effects of the chosen parameters on the power density were investigated. According to
the results, the most efficient parameter on the power density was determined as the catalyst loading.
The other important parameters were temperature and borohydride concentration. The decrease in the

507
performance with increasing catalyst loading can be explained as the blocking of the active surface
area at high catalyst loading and the mass transfer limitation. The maximum power density of 37.15
2 2
mW/cm was obtained at the operation conditions of 0.52 mg/cm catalyst loading, 80°C fuel cell
temperature, 3.28 L/min fuel and 0.27 L/min oxidant flow rates and 1.5 M borohydride concentration.
The achieved formula can be adapted to all fuel cell studies including Au catalyst, N117 membrane
and ELAT GDL.

REFERENCES
Ma, J., Choudhury N. A. and Y. Sahai. 2010. A comprehensive review of direct borohydride fuel cells.
Renewable and Sustainable Energy Reviews 4:183–199

Celik, C., Boyaci San F.G. and HI, Sarac. 2008. Effects of operation conditions on direct
borohydride fuel cell performance. J. Power Sources 185(1):197–201.
Celik, C., Boyaci San F.G, and HI Sarac.2010. Influences of sodium borohydride concentration on
direct borohydride fuel cell performance. J. Power Sources 195(9):2599–2603.

Nagle, L.C. and Rohan, J.F. 2011. Nanoporous gold anode catalyst for direct borohydride fuel cell. Int
J Hydrogen Energy 2011, 36:10319-10326.

Kim, J.H., Kim, H.S., Kang, Y.M., Song, M.S., Rajendran, S.,. Han, S.C, Jung, D.H. and J.Y. Lee.
Carbonsupported and unsupported Pt anodes for direct borohydride liquid fuel cells.2004.
Electrochem. Soc. 151:A1039–A1043.

Lam, V. W. S. Alfantazi A. and E. L. Gyenge. 2009. The effect of catalyst support on the performance
of PtRu in direct borohydride fuel cell anodes. Journal of Applied Electrochemistry 39:1763-1770.
Pe,i F. , Wang, Y., Wang, X., He, P., Chen, Q., Wnag, X., Wnag, H., Yi, L., Guo, J. 2010 Performance
of supported Au–Co alloy as the anode catalyst of direct borohydride-hydrogen peroxide fuel cell. Int J
Hydrogen Energy 35: 8136-8142.

Amendola, S.C., Onnerud P., Kelly, M.T., Petillo, P.J., Sharp-Goldman, S.L., and M. Binder. 1999.
Anovel high power density borohydride-air cell. J Power Sources 84:130–3.

Gyenge E., Atwan M. andD. Northwood. 2006. Electrocatalysis of borohydride oxidation

on colloidal Pt and Pt-alloys (Pt-Ir, Pt-Ni, and Pt-Au) and application for direct borohydride fuel cell
anodes.J. Electrochem. Soc. 153: A150–A158.

Liu B.H., Li Z.P., Zhu J.K. and S. Suda. 2008. . Influences of hydrogen evolution on the cell and stack
performances of the direct borohydride fuel cell. J. Power Sources 183:151–156.

Li Z.P., Liu B.H., Zhu J.K. and S.Suda . 2006. Depression of hydrogen evolution during
operation of a direct borohydride fuel cell. J. Power Sources 163:555–559.

Park K.T., Jung, U.H., Jeong, S.U. and S.H.Kim. 2006. Influence of anode diffusion layer
properties on performance of direct borohydride fuel cell. J Power Sources 162:192–197.

Liu B.H., Li Z.P.and S. Suda. 2008. A study on performance stability of the passive direct
borohydride fuel cell J. Power Sources 185: 1257–1261.

Kim C., Kim K.J. and M.Y. Ha., 2008. Performance enhancement of a direct borohydride fuel cell in
practical running conditions. J. Power Sources 180: 114–121.Liu B.H. and S., Suda . 2007. Influences
of fuel crossover on cathode performance in a micro borohydride fuel cell. J. Power Sources 164:100–
104.

508
 DESIGN AND ASSESSMENT OF A NET ZERO ENERGY HOUSE

A. Hassoun and I. Dincer

Faculty of Engineering and Applied Science
University of Ontario Institute of Technology
2000 Simcoe Street North, Oshawa, Ontario, L1H 7K4, Canada
Emails: Hassoun Anwar@uoit.ca, Ibrahim.Dincer@uoit.ca

ABSTRACT
This work presents an attempt to solve the shortage of power generation in Lebanon as a renewable
energy-based case study and supply a neighborhood with electricity, hot water, and seasonal heating and
cooling. The study starts by choosing the right orientation and construction material of the house, in order,
to minimize energy and water usage and reduce outside noise transmission, by optimizing the thicknesses
of acoustic and thermal insulation materials used to achieve the desired requirements. An exergy analysis
is conducted to assess various system components and the overall system efficiency through multi
parametric variables. A comprehensive economic evaluation is carried out where several techniques are
utilized and optimization of several factors that achieve, maximizing of renewable energy usage by reducing
exergy destruction and minimizing cost. The ultimate goal of the study is to produce a sustainable net zero
energy house with very low level of CO2 gas emissions.

Keywords:Zero energy, photovoltaic, wind turbine, absorption chiller electricity, energy, exergy, efficiency,
exergoeconomic analysis, optimization, Pareto front.

1. INTRODUCTION
Energy shortages have been increasing due to population growth and increased industrial activities in
especially developing countries. This makes a sustainable supply of energy critical to world development.
After the maximum rate of global oil extraction is reached, a decline in the supply occurs. This will cause a
significant increase in oil prices and their byproducts which will cause worldwide socio-economic
implications. Delaying the time of peak oil can be achieved by resorting to other alternatives for producing
energy such as, nuclear energy, renewable energy, unconventional oil resources like oil shale and oil sands
and by producing hydrogen fuel from the electrolysis of water. Moreover, fossil fuels are a major contributor
to pollution and global warming. Hence reducing the usage of petroleum and most fossil fuels is as
important as achieving a sustainable energy supply. So, switching to sustainable energy options, such as
renewable energy, nuclear energy, etc. becomes an ultimate goal and helps reduce carbon dioxide
emissions.

This present study aims to achieve the above mentioned goal for Lebanon, in addition to the fact that
Lebanon has big problems in electricity supply, because it has been stricken by civil war, regional wars and
bad governments. On average, throughout the year, a house is supplied by no more than eight hours of
electricity daily. This forces the population to buy electricity from non-governmental individuals who use
generators to provide electricity at very high rates. Adding to all that shortages in water supply during the
summer season, even though Lebanon is rich with water, but due to the above reasons and the past
occupations which prevented the Lebanese from using their water resources or develop projects to provide
water to the country. These factors combined, make the need to generate electricity in Lebanon by
renewable energy a must. In this regard, we select an area in South Lebanon where nearby river water is
being wasted, but the water needs treatment due to negligence, to promote a community which would
benefit from the natural resources and become self sufficient if the proposed design is implemented. A net
zero energy house where production of electricity is equal to consumption or greater will be targeted in this
study.

2. PROBLEM STATEMENT
In this case study, a house occupied by a family of six, with a total living area of 225𝑚2 located in South
Lebanon near the Litani Riveris considered, as to be powered by maximum renewable energy with least
green house gases emission and low noise pollution. An overall heat transfer coefficient for walls not
𝑤 𝑤
exceeding 0.1 2 and for ceiling 0.075 2 are to be used, while for acoustic insulation the ability to
𝑚 ℃ 𝑚 ℃
attenuate a sound pressure level from 76 db outside the house to 30 db inside the house.

509
3.SOLUTION STRATEGY
Since Lebanon has a moderate climate, where peek summer temperatures are at 32.8°C and winter
temperatures are around 5.6°C in average [1].The following approach is followed to reduce the energy
requirements and noise pollution:
• Constructing a good building envelope by optimizing the sizes of thermal and acoustic insulation used.
• Choosing the right orientation of the house.
• Using energy-efficient appliances and light fixtures.
• Using efficient water fixtures.
• Using high-efficiency mechanical systems.
• Maximizing the use of solar powered cooling and heating techniques.
• Producing space and water heating by solar thermal systems, thermal storage techniques and heat
pumps.
• Producing water by systems powered by solar energy.
• Offsetting electrical use by an integrated system with grid-connected, photovoltaic, wind turbine, etc.

4. FIELDS OF STUDY
The concept of net zero energy house is described by The U.S. Department of Energy (DOE) Building
Technologies Program:“A net zero-energy building (ZEB) is a residential or commercial building with greatly
reduced energy needs through efficiency gains such that the balance of energy needs can be supplied with
renewable technologies”[e.g., 2].What is important in this study is to maximize electrical power from
renewable sources that would be equal to the house consumption and preferably more, to offset any
governmental taxes and price variations, in order to cost the owner no additional charges and to minimize
the government grid supply of electricity for sustaining a low carbon emission house, because, the
government supply might be through power stations running on fossil fuels.

4.1 Acoustic insulation of the proposed house is determined through experimental techniques.
An experimental study is in this regard carried out using Roxul fiber glass insulation and rigid expanded
polystyrene foam to determine the SPL transmitted through the acoustic material, in order to determine the
thickness to be used in the wall construction. In addition another experiment is conducted using the
impedance tube to determine the coefficients of reflection of the two acoustic materials.The driving force for
any heat transfer is the temperature difference while for acoustics it is the propagation of sound waves.The
sound is a pressure wave that travels through an elastic medium with certain speed.
4.1.1 Experimental Set up
The experiment is conducted at the acoustics and noise control lab at the UOIT campus where SPL and
reflection coefficients are measured for the following acoustical materials by two different techniques:
4.1.1.1 Impedance tube: A speaker is placed at the face of one end of a tube and is sealed properly by
using two wooden flanges to insure that there is no leakage of transmitted sound, while at the other end the
tested material is fixed by two flanges to secure the test in place and to make sure that no noise is leaking.
Two Pressure microphones are installed at the surface of the tube as shown in the following figure.The
experiment was conducted to calculate the reflection coefficients using: a 100x15 cmPVC pipe arranged as
shown in Figure1where S = 2 cm.L= 15 cm being the distance from the face of the acoustic material to the
first microphone.

Figure 1.Impedance tube experiment setup.

510
In this method, a broadband stationary random acoustic wave in a tube is mathematically decomposed into
its incident and reflected components using a simple transfer-function relation between the acoustic
pressures at two locations on the tube wall. The wave decomposition leads to the determination of the
complex reflection coefficient from which the complex acoustic impedance and the sound absorption
coefficient of a material and the transmission loss of a silencer element can be determined [3]. A
methodology is used as follows, based on the work done by [4]:
𝑍 𝐻12 𝑠𝑖𝑛(𝑘𝑙)−𝑠𝑖𝑛 (𝑘(𝑙−𝑠)
= 𝐽[ (1)
𝜌𝑐 𝑐𝑜𝑠(𝑘(𝑙−𝑠)−𝐻12 𝑐𝑜𝑠(𝑘𝑙))

𝑍
= 𝑥 + 𝑖𝑟(2)
𝜌𝑐

1
[𝑅𝑒 (𝐻12 ) 𝑠𝑖𝑛 (𝑘 (2𝑙−𝑠) − 2 {𝑠𝑖𝑛 (𝑘 (2𝑙−𝑠)) +|𝐻12 |2 𝑠𝑖𝑛 (2𝑘𝑙)}]
x = 𝜌𝑐 𝐻𝑑
(3)

[−𝐼𝑚 (𝐻12 )𝑠𝑖𝑛 (𝑘𝑠)]

r = 𝜌𝑐 𝐻𝑑
(4)

where𝜌𝑐 is the characteristic impudence of air.

𝐻𝑑 = 𝑐𝑜𝑠 2 [𝑘(𝑙 − 𝑠)] + |𝐻12 |2 𝑐𝑜𝑠 2 (𝑘𝑙)(5)

where𝐼𝑚 , 𝑅𝑒 represent the real and imaginary parts, and l is the distance from the speaker to the first
microphone.H12 Is the transfer signal between the two microphone signals.
𝐻12 (𝑓)−𝐻𝑖 (𝑓)
𝑅1 (𝑓) = (6)
𝐻𝑟 (𝑓)−𝐻12 (𝑓)

𝑅 = 𝑅1 𝑒 𝑗2𝑘𝑙 (7)

whereR is the reflection coefficient and 𝜶 is the absorption coefficient.

𝜶 = 1-|𝑅1 |2 (8)

4.1.1.2 Enclosure Experiment

The enclosure is made up of the acoustic material .The enclosure is well sealed to avoid any sound
leakage and a free field microphone is installed inside the enclosure while a generated noise is built on the
outside using a speaker .The microphone signal is connected to the analog instrument and then transferred
to the lab view software where the FFT voltage signal is recorded at different frequencies and then
transformed into pressure signal by multiplying with the voltage/pressure sensitivity factor of the
microphone. After which the SPL is calculated and the attenuation by each material are evaluated as
follows:
𝑃𝑟𝑚𝑠
SPL=20 𝑙𝑜𝑔10 � 𝑃0
� (9)
−6
𝑃0 =20 .10 (10)

𝑃𝑟𝑚𝑠 = Value obtained from 𝐹𝐹𝑇(11)

4.2 Photovoltaic systemsare used in this study produce electricity directly from sun light and are referred
to as solar energy systems. Two layers of a semi-conducting material are combined to produce electricity
when exposed to sun light. It works when one layer has a depleted number of electrons that gets excited by
absorbance of sun light photons and cause electrons to jump from one layer to the other causing electric
charge. The photovoltaic material is silicon cut into very thin wafers and undergoes contamination by
doping to create electron imbalance. These wafers are joined together to form a solar cell and many cells
are connected in series and laminated between clear glazing and an encapsulating substrate.

The efficiency of PV/T system used in this study depends on many factors, and it is described in detail by
[5]:

511
 ℎ𝑐 (𝑇𝑐𝑒𝑙𝑙 − 𝑇𝑎 )𝐴𝑠𝑃𝑣𝑇 +𝐸𝑙𝑒𝑐𝑔𝑒𝑛𝑒𝑟
𝜂= (12)
𝑄̇𝑠

𝑇
ℎ𝑐 (𝑇𝑐𝑒𝑙𝑙 − 𝑇𝑎 )𝐴𝑠𝑃𝑉𝑇 (1− 𝑇𝑎𝑚𝑏 )+𝐸𝑙𝑒𝑐𝑔𝑒𝑛𝑒𝑟
Ψ= 𝑇
𝑠𝑢𝑛
(13)
𝑄̇𝑠 (1− 𝑎𝑚𝑏 )
𝑇𝑠𝑢𝑛

where all temperatures in exergy and energy calculations are absolute and must be in K.

4.3 Wind turbineis considered in this study to produce electricity by converting wind kinetic energy into
mechanical energy. At the end of 2009, 159.2 GW [5] of wind-powered generators were produced
worldwidewhich is about 2% ofthe total electricity produced. Several countries have achieved relatively high
levels of wind power penetration.

The efficiency of the wind power generators is best explained by [6]:

̇
𝑃𝑡𝑢𝑟𝑏 3
= 0.5𝜌𝑎𝑖𝑟 𝐴𝑠𝑤𝑒𝑝𝑡 𝑉𝑎𝑖𝑟 𝐶 (14)
𝑖𝑛 𝑃𝑂𝑊

̇
𝑃𝑡𝑢𝑟𝑏_𝐸𝑙𝑒𝑐 ̇
= 𝑃𝑡𝑢𝑟𝑏 𝜂 𝑇𝑅𝐴𝑁𝑆 𝜂𝐶𝐴𝑃𝐴𝐶 (15)

𝑃̇𝑇𝑢𝑟𝑏,𝑒𝑙𝑒𝑐
𝜂turb = (16)
𝑃̇𝑇𝑢𝑟𝑏_𝑁𝑜𝑚

𝑇 𝑃 𝑄 ̇
𝜓𝑡𝑢𝑟𝑏 = 𝑃̇𝑇𝑢𝑟𝑏,𝑒𝑙𝑒𝑐 /(𝑃̇𝑇𝑢𝑟𝑏,𝑁𝑜𝑚 + 𝑚̇ 𝑇𝑢𝑟𝑏 Cp(𝑇𝑏𝑡 -𝑇𝑎𝑡 )+𝑚̇ 𝑇𝑢𝑟𝑏 𝑇𝑜 (Cp Ln ( 𝑎𝑡 )-R Ln ( 𝑎𝑡))- 𝑇𝑢𝑟𝑏,𝑙𝑜𝑠𝑠)(17)
𝑇𝑏𝑡 𝑃𝑏𝑡 (𝑇𝑎𝑚𝑏 )

4.4 Batteries and electrical systems are considered and used in this study. Batteries store DC electrical
energy in chemical form for later use. The energy is used at night and during periods of emergency .The
DC is inverted to AC before it is supplied to the house.

4.5Absorption Chilleris considered and used in this study operates as a mechanical vapor compression
heat pump except that the compressor is replaced by an absorbent fluid circuit. An absorption chiller uses
an aqueous solution of LiBr where the absorbent has a great affinity for the refrigerant. The refrigerant is
absorbed by the absorbent solution in the absorber, releasing heat of solution. The weak solution is then
pumped to the generator. In the generator, hot water is supplied to drive off the refrigerant from the
absorbent solution. The refrigerant is condensed giving up its latent heat in the condenser. The strong
solution is then returned to the absorber. The refrigerant from the condenser is pumped to the evaporator
pressure, where air from the house causes it to evaporate. The refrigerant vapor then passes to the
absorber to close the cycle. The overall system COP is defined by [6]:
𝑄̇𝑒𝑣𝑎
𝜂𝑡ℎ,𝑎𝑏𝑠,𝑠𝑦𝑠 = 𝐶𝑂𝑃𝑎𝑏𝑠,𝑠𝑦𝑠 = (18)
𝑊̇𝑃𝑚𝑝,1 +𝑄̇𝑔𝑒𝑛

𝑇
𝑄̇𝑒𝑣𝑎 (𝑇 0 −1)
𝑎,𝑖𝑛
𝜓𝑥,𝑎𝑏𝑠,𝑠𝑦𝑠 = 𝑇 (19)
𝑊̇𝑃𝑚𝑝 +𝑄̇𝑔𝑒𝑛 �1− 0 �
𝑇𝑔𝑒𝑛

𝑄̇𝑒𝑣𝑎 +𝑃̇𝑒𝑙𝑒𝑐,𝑃𝑉𝑇 +𝑃̇𝑇𝑢𝑟𝑏,𝑒𝑙𝑒𝑐 +ℎ𝑐 𝐴𝑃𝑉𝑇 (𝑇𝑐𝑒𝑙𝑙 −𝑇𝑎𝑚𝑏 )

𝜂𝐺𝐿_𝑆𝑌𝑆 = (20)
𝑄̇𝑆 (𝐴𝑠𝑜𝑙𝑎𝑟+𝐴𝑃𝑉𝑇 )+𝑊̇𝑃𝑚𝑝1 +𝑊̇𝑃𝑚𝑝,𝑠 +0.5𝜌𝑎𝑖𝑟𝐴𝑆𝑤𝑒𝑝𝑡 𝑉𝑎𝑣𝑔 𝑉𝑎_𝑖𝑛2

𝑇 𝑇
𝑄̇𝑒𝑣𝑎 �𝑇 0 −1�+𝑃̇𝑒𝑙𝑒𝑐 𝑃𝑉𝑇 +𝑃̇𝑒𝑙𝑒𝑐𝑇𝑢𝑟𝑏 +ℎ𝑐 𝐴𝑃𝑉𝑇 (𝑇𝑐𝑒𝑙𝑙 −𝑇𝑎𝑚𝑏 )�1−𝑇 0 �
𝑎,𝑖𝑛 𝑐𝑒𝑙𝑙
𝜓𝐺𝐿_𝑆𝑌𝑆 = 𝑇 (21)
𝑄̇𝑆 (𝐴𝑠 +𝐴𝑃𝑉𝑇 )�1− 0 �+𝑊̇𝑃𝑚𝑝1 +𝑊̇𝑃𝑚𝑝,𝑠 +0.5𝜌𝑎 𝐴𝑆𝑤𝑒𝑝𝑡 𝑉𝑎𝑣𝑔 𝑉𝑎,𝑖𝑛2
𝑇𝑆𝑢𝑛

5. HEATING LOAD CALCULATIONS

These calculations are carried out using hourly analysis program software for the proposed house.The
software simulates hour-by-hour operation of all heating and air conditioning systems in the house and
operation of all plant equipment and non HVACs systems including lighting and appliances. It uses results

512
of simulations to calculate total annual energy use and energy costs and generates tabular and graphical
reports of hourly, daily, monthly and annual data [7].

6. EXERGOECONOMIC ANALYSIS
The following methodology of exergoeconomic analysis is employed to evaluate the energetic and
exergetic costs, based on the studies by Refs. [6, 8]:

𝐶𝑖̇ = ∑𝑖 𝑐𝑖 𝑚̇ 𝚤 [(ℎ𝑖 − ℎ0 ) − 𝑇0 (𝑠𝑖 − 𝑠0 )](22)

𝐶𝑒̇ = ∑𝑒 𝑐𝑒 𝑚̇ 𝑒 [(ℎ𝑒 − ℎ0 ) − 𝑇0 (𝑠𝑒 − 𝑠0 )](23)

̇ = �∑𝑖 𝑐𝑤,𝑖 𝑊̇𝑖 − ∑𝑒 𝑐𝑤,𝑒 𝑊̇𝑒 �(24)

𝐶𝑤
̇
𝑄 (𝑇 −𝑇 ) 𝑄 (𝑇 ̇
−𝑇 )
𝐶𝑞̇ = ∑𝑖 𝑐𝑞,𝑖 𝑖 𝑠𝑜𝑢𝑟𝑐𝑒 0 − ∑𝑒 𝑐𝑞,𝑒 𝑒 𝑠𝑜𝑢𝑟𝑐𝑒 0 (25)
𝑇𝑠𝑜𝑢𝑟𝑐𝑒 𝑇𝑠𝑜𝑢𝑟𝑐𝑒

(1+𝑖𝑟)𝑛 1
̇
𝐶𝑐𝑖𝑛𝑣 = 𝐼𝐶 ( )(26)
(1+𝑖𝑟)𝑛−1 𝑁

(1+𝑖𝑟)𝑛 1
𝐶𝑜̇ & 𝑚 =𝐶𝐹𝑌𝑀
̇ � �(27)
(1+𝑖𝑟)𝑛 −1 𝑁

where 𝑖𝑟 = interest rate, 𝑛 = # of life cycle years, 𝑁 = # of hours of operation/year,𝐼𝐶 = Initial cost,

FYM = First year maintenance cost.

𝐶𝑖̇ + 𝐶𝑞̇ + 𝐶𝑐𝑖𝑛𝑣

̇ +𝐶𝑜̇ & 𝑚 = 𝐶𝑒̇ + 𝐶𝑤
̇ (28)

̇
𝐶𝑒𝑥_𝑑𝑒𝑠 = �∑𝑘 𝑚̇ 𝑘 𝑐𝑜𝑠𝑡𝑆𝐹𝑃𝑘 𝑒𝑥𝑘 �(29)

̇ = Ċ ex_des + Ċ cinv +Ċ o & 𝑚 = Ċ e (30)

CTC

7. MULTI OBJECTIVE OPTIMIZATION

A multi-objective optimization approach is used to minimize the total cost and maximize the global
exergetical efficiency. This multi objective optimization uses a genetic algorithm that deals with the
minimization of a vector of objectives F(x) that can be the subject of a number of constraints or bounds and
when maximization is needed the F(x) is multiplied by a minus sign [9-12]:

𝐺𝑖 (𝑥) = 0, 𝑖 = 1, … … … … . . , 𝑘𝑒 ; 𝐺𝑖 (𝑥) ≤ 0, 𝑖 = 𝑘𝑒 + 1, … 𝑘; 𝑙 ≤ 𝑥 ≤ 𝑢(31)

8. RESULTS AND DISCUSSION

The design of a zero energy house was established by following the procedure outlined in section 3 and all
the system components needed to provide the house with electricity, hot water and seasonal air-
conditioning are shown in Figure 2.First, the Thermal and acoustic sizes of the material chosen for this
study is determined and the heat load calculations is carried out next. After establishing the total electrical
KW needed for the house, the optimum power system, based on the system minimum net present value
cost, minimum Co2 emissions and maximum renewable energy fraction, is selected. After which an energy,
exergy analysis is conducted to assess the system performance along with an exergoeconomic analysis to
the global system. Finally, the selected system is optimized for minimum global cost and maximum
exergetic efficiencies and assessment for zero energy classification is investigated.Here, in the system hot
water from the solar collector powers the generator of the absorption chiller while ambient air cools off the
condenser and absorber. Cooled air is being provided to the house at stream 14 after being re-circulated at
stream 13 to the evaporator where it loses heat and moisture as explained in section 4.5, while hot water is
supplied to the house at stream 18 through the photovoltaic cum thermal system. Electricity is generated by
the photovoltaic cells, wind turbine and diesel generator.

513
Figure 2.A schematic diagram of the present system.

8.1Selecting the thermal and acoustic material sizes.

After conducting the impedance tube and enclosure experiments a Matlab code was generated to produce
the results as shown in Figures 3 and 4.

140
1"+3"
120
2"+3"
100
Amb+1KHz
Decibels

80

60

40

20

0
1
115
229
343
457
571
685
799
913
1027
1141
1255
1369
1483
1597
1711
1825
1939
2053
2167
2281
2395

Figure3.Sound pressure Level for ambient noise+1000 Hz sinusoidal wave vs. a combination of (1" + 3",
2”+3”) materialmeasured at different frequencies (Hz).

An examination of Figure 3 indicates that the sound pressure level (SPL) of the ambient noise plus the
1000Hz sinusoidal noise is 75 db, while the SPL of the tested specimen ofone inch rigid expanded
polystyrene foamand 3 inchRoxul fiber glass insulation is 35 db, which means a reduction of 40 db was
caused by using an enclosure of the 1”+3” combination. The second experiment, using the same ambient
plus 1000 Hz noise level, but this time a 2” + 3” combination were used for the enclosure on all sides and
the measured SPL as seen in figure 3 was 15 db. Therefore the 1” rigid expanded polystyrene foam has the

514
power to attenuate 20 db and is equal to that of the 3” fiber glass insulation. These results were verified by
the impedance tube experiment which shows that the two specimens as in figure 4 have the same
reflection coefficient. The thermal and acoustic insulation sizes that satisfy the problem statement were
determined by a code which optimized the thicknesses to be at, 4.1″& 2.3″ for the wall and 4.72”& 2.3” for
the roof, structures, etc.

Figure 4. Reflection coefficients of the 1" and 3" materials vs. the theoretical of open Impedance tube
measured at various frequencies.

8.2 Calculating the seasonal heat loads.

The building wall and roof composition were finalized by the previous step and the rest of the design criteria
were designed to enable saving in heat transfer from and to the building, to reduce electricity
requirements.Special care was given to the glass to wall ratio not to exceed 8% and to have an air tight
building.Summer and winter loads were calculated using HAP software for design temperatures as shown
in Table 1. The building heat transfer coefficients were set low to minimize the seasonal loads and the
results obtained by the simulation software were 4601W for summer cooling and 1703 W for winter heating.

Table 1 Design temperature values for summer cooling and winter heating
Air temperature Cooling Heating
Outside dry bulb (℃) 32.8 5.6
Outside wet bulb (℃) 25.6 1.8
Inside dry bulb (℃) 23 21.1
Inside relative humidity (%) 50 50

8.3Selection of the power system

After determining the heat loads of the house, the total peek electrical usage was calculated at 5.7 kW and
the average electrical daily demand at102 kWh/d which was the basis for our modeling and optimization.
Two schemes were configured for our analysis, one was connecting to the grid while the other not. The
basis for our optimization is to minimize the net present value of the power system and minimize the
CO2emissions. Around 18100 simulations were conducted and the following on grid and off grid systems
were selected as shown in table2 and their monthly average electrical production in Figures 5 and 6.

The net present cost of the on grid and off grid systems are $41,775 and $80,146 respectively as seen in
Figures 7 and 8. The emissions generatedby the selected optimized systems are given in Table 3 where
the grid connected system produces carbon dioxide emissions about 10 times more than the off grid in
kg/year. This is a major concern that will be considered in system evaluation further down the study

515
Table 2.Optimized systems architecture for on and off grid connections.
On grid connection Off grid connection
PV Array 16 kW PV Array 2 kW
Wind turbine 1 Generic 10kW Wind turbine 1 Generic 10kW
Generator 1 4 kW Grid 1,000 kW
Battery 50 Hoppecke 24 OPzS 3000 Battery 2 Hoppecke 24 OPzS 3000
Inverter 7 kW Inverter 10 kW
Rectifier 7 kW Rectifier 10 kW

Figure 5.Monthly electrical productions for on grid connection.

Figure 6.Monthly electrical productions for off grid connection.

The electrical production for the selected systems components and the renewable energy fractionsare
determined and listed in Table 4.

Table 3 Pollutants emitted by the selected optimized on grid and off grid systems.
Off grid connection On grid connection
Pollutant Emissions (kg/yr) Pollutant Emissions (kg/yr)
Carbon dioxide 922 Carbon dioxide 8,920
Carbon monoxide 2.28 Carbon monoxide 0
Unburned hydrocarbons 0.252 Unburned hydrocarbons 0
Particulate matter 0.172 Particulate matter 0
Sulfur dioxide 1.85 Sulfur dioxide 38.7
Nitrogen oxides 20.3 Nitrogen oxides 18.9

516
 Figure 7. Net present cost for on grid connection.

Figure8. Net present cost for off grid connection.

Table 4 Electrical production for offgrid and ongrid system components

Off grid connection On grid connection
Component kWh/yr Fraction Component kWh/yr Fraction

PV array 29,961 57% PV array 3,745 8%

Wind turbine 21,969 41% Wind turbine 21,969 47%

Generator 1 1,024 2% Grid purchases 20,738 45%

Renewable - 0.972 Renewable - 0.527

8.4 Exergy analysis results

A sensitivity analysis on the system components was carried out by varying independent parameters. EES
software was used to calculate all the system properties and develop the parametric studies. The effect of
keeping the dead state temperature 𝑇𝑜 = constantand 𝑇𝑒𝑣𝑎 = 7, 𝑇𝑔𝑒𝑛, = 80 , 𝑇𝑎,𝑖𝑛 = 20(˚C) are shown for the
overall COP for the absorption chiller and the global system efficiencies in figures 9&10.The COP varied
from0.53 to0.28 at around 3:00 pm due to increase in generator heat as seen in figure 11 while the
exergetical COP varied slightly around 0.15.The global efficiency varied between 38% and 45% due to

517
solar and velocity changes while the global system exergy varied between 6&10%. These results clearly
indicate that the exergetic analysis represents a better sensitivity analysis of the system performance than
energy.

0,6
COP COP_II
0,5

0,4
COP

0,3

0,2

0,1

0
298 298 298 298 298 298 298 298 298 298

Reference Temperature (˚C)

Figure 9.First and second law coefficients of performance of the absorption chiller

50%
η ψ
45%
40%
35%
Effieciency

30%
25%
20%
15%
10%
5%
0%
298 298 298 298 298 298 298 298 298 298
Reference Temperature (˚C)

Figure 10.Energy and exergyefficiencies of the global system

Figures 12 and 13 where all parameters were fixed except for reference temperature, we notice that the
exergetical COP of the absorption system increased from 0.16 to about 0.22 at maximum 𝑇0 because the
𝑇 𝑇
factor 𝑄̇𝑔𝑒𝑛 �1 − 0 � decreases more than 𝑄̇𝑒𝑣𝑎 � 0 − 1�. While the global exergy efficiency has the factor
𝑇𝑔𝑒𝑛 𝑇𝑒𝑣
𝑇0
�1 − � in the denominator which is not affected by the range of changesin𝑇0 since 𝑇𝑆𝑢𝑛 is large and the
𝑇𝑆𝑢𝑛
𝑇0
numerator was affected slightly by� − 1�.Thus causing an unnoticeable decrease.
𝑇𝑒𝑣

518
 27,00
Q͘_gen Q͘_eva
24,00
21,00
Heat Rate (k W)

18,00
15,00
12,00
9,00
6,00
3,00
0,00
9 10 11 12 13 14 15 16 17 18
Hour of the day
Figure 11. Generator and evaporator thermal heat at different hour of the day and constant temperatures

0,6

0,5
COP COP_II
0,4
COP

0,3

0,2

0,1

0
298 299 300 301 302 303 304 305 306 307
Reference Temperature (˚C)

Figure 12. First and second lawcoefficients of performance of the absorption system at variable dead state
temperature and𝑇𝑒𝑣𝑎 = 7, 𝑇𝑔𝑒𝑛, = 80 , 𝑇𝑎,𝑖𝑛 = 20(˚C)
50%
η ψ

40%
Effieciency

30%

20%

10%

0%
298 299 300 301 302 303 304 305 306 307
Reference Temperature (˚C)
Figure 13.Energy and exergy efficiencies of the global system at variable dead state temperature and𝑇𝑒𝑣𝑎 =
7, 𝑇𝑔𝑒𝑛, = 80 , 𝑇𝑎,𝑖𝑛 = 20(˚C).

519
Comparing Figures 14 and 15 where all parameters were fixed except for temperature of air entering the
evaporator, the COP of absorption chiller increased dramatically from 0.3 to 0.9 with temperature increase,
because, 𝑄̇𝑒𝑣𝑎 increased. Also the global efficiency increased slightly, while the exergetical COP of the
absorption system increased with increased air temperature from 0.1 to 0.24. The global system exergy
was increased slightly with increased air temperature while the energy efficiency dropped slightly at higher
temperature with comparison to Figure10.
1,00
COP COP_II
0,80

0,60
COP

0,40

0,20

0,00
26 25 24 23 22 21 20 19 18 17
Air Temperature (˚C)

Figure 14.First and second law coefficients of the absorption system at variable air on evaporator
temperature and 𝑇𝑒𝑣𝑎 = 7, 𝑇𝑔𝑒𝑛, = 80 , 𝑇𝑜 = 25(˚C)

60%
η ψ
50%

40%
Efficiency

30%

20%

10%

0%
26 25 24 23 22 21 20 19 18 17
Air Temperature(˚C)

Figure 15. Energy and exergy efficiencies of the global system at variable air on evaporator temperature
and 𝑇𝑒𝑣𝑎 = 7, 𝑇𝑔𝑒𝑛, = 80 , 𝑇𝑜 = 25(˚C)

In Figures16 and 17, the temperature of air entering the evaporator and the evaporator temperature are
varying 𝑤ℎ𝑖𝑙𝑒 𝑇𝑔𝑒𝑛 = 80 , 𝑇𝑜 = 25 (˚C) remains constant. We notice that the COP of absorption chiller
increased but less than that of figure 14 due to evaporator temperature increase which countered some of
the increase caused by air temperature. Also, the global efficiency decreased from that of Figure 15 due to
the same reasoning, while the exergetical COP of the absorption system decreased with increased air and
evaporator temperature due to additional decrease in 𝑄𝑒𝑣𝑎 ,caused by increase in evaporator temperature .
Global system exergy decreased unnoticeably due to increase in evaporator temperature in comparison to
Figure 15.

520
 0,8
0,7
COP COP_II
0,6
0,5
COP

0,4
0,3
0,2
0,1
0
Ta 26 25 24 23 22 21 20 19 18
T_ev 13 12 11 10 9 8 7 6 5

Evaporator/Air Temperature (˚C)

Figure 16.First and second law coefficients of absorption system at variable 𝑇𝑒𝑣𝑎 and 𝑇𝑎 (˚C).

50%
η ψ
45%
40%
35%
Efficiency

30%
25%
20%
15%
10%
5%
0%
Ta 26 25 24 23 22 21 20 19 18
T_ev 13 12 11 10 9 8 7 6 5

Evaporator/Air Temperature(˚C)
Figure 17. Energy and exergy efficiencies of the global system at variable 𝑇𝑒𝑣𝑎 and 𝑇𝑎 (˚C).

3%
3% abs
6%
con
44%
solar
37% eva
Gen
4% hex
PVT
turbine
1% 2%

Figure 18. Percentage of total exergy destructed for system components.

521
The percentage and values of exergy destructed for system components are presented in figures 18&19 for
variable T0 . The absorber followed by the generator has the highest exergy destruction rate because of
chemical exergy losses caused by heat of solution due to absorption and desorption respectively.

120 abs

100 con

80 solar
Energy Rate (kW)

60 eva

40 Gen
hex
20
Gen PVT
0
To 25 26 abs turbine
27 28 29
30 31 32
33

Figure 19.The values of exergy destructed for system components at variable reference temperature (℃).

8.5 Exergo-economic analysis and optimizationresults

The exergo-economic analysis was conducted to calculate the total cost of the system equipments,
operation, maintenance and exergy destructed. Optimization using multi-objective technique was carried
out for minimum total system cost and maximum Global system exergy efficiency.The Pareto Front is
plotted in Figure 16 where the global exergy efficiency ranged from 19.5% to 20.5% at a total cost range of
0.785 to 0.82 $/hr.

0,82
Pareto Front Curve fitting
0,815

0,81
Total cost $/hr

0,805

0,8

0,795
y = 2E-06x5 - 6E-05x4 + 0,0007x3 - 0,0034x2 + 0,0072x +
0,79 0,7937
R² = 0,9618
0,785

Global exergy efficiency

Figure 20. Optimal solution curves for the global system exergy efficiency vs. the total cost.

8. CONCLUSIONS
The system with no grid connection was chosen, even though the net present value of the grid connected is
half that of the off grid. The reason was the large undesired emissions. A sustainable environment and the
objective of this study outweigh the cost.

522
The global system exergy efficiency was improved from the original design, by the optimization, due to
variation of 24 independent parameters through an exhaustive simulation, from 11% to a maximum of 21%
at a total system net present cost increase of 16%.The Pareto Front fromthe simulations represents all
possible combinations for the system objectives at the studied parameters. It is worth noting that if one
objective is enhanced the other is degraded. Even though the global exergetical efficiency was low due to
the influence of solar and wind losses the system is recommended because it serves the objective well and
cost of solar energy is free.

Exergy analysis of the system efficiencies clearly showed a lower performance of the system components
in comparison to energy.

The maximum renewable energy fraction of the selected system was 0.972. In other words, we have
achieved a near zero sustainable energy house.

Future study will concentrate on adding fresh water to the house and utilizing thermal ground storage with
hydro and ammonia turbine generators.

NOMENCLATURE
Cp specific heat (kJ/kg-k)

Ċ stream cost rate ($/s)

ṡ Entropy (kJ/kg-k)

ḣ enthalpy (kJ/kg)

Ṗ Power

PO reference pressure (kpa)

p pressure (kpa)

To reference temperature (°C)

T temperature (°C)

ṁ mass flow rate (kg/s)

Q̇ heat rate (kW)

ẇ power rate (kW)

Subscripts

gen generator

abs absorber

eva evaporator

con condenser

hex heat exchanger

in Inlet

i Inlet stream

523
e Exit stream

PV/T Photovoltaic/Thermal

turbine Wind turbine

Ve Wind velocity

Ex Exergy

Cpow Power coefficient

ηtrans Efficiency of transferring Dc to Ac current

in Inlet

ASwept Area swept by turbine blades (m2

s Solar

W Water

cell Photovoltaic cell

h_c Heat transfer coefficient

ir Interest rate

N Yearly hours of operation

R Universal constant(kJ/kg-k)

n life cycle in years

c cost of stream ($/GJ)

Re real part

Im imaginary part

R reflection coefficient

Turb Turbine

Capc Capacitance

Trans Transmission

at After turbine

bt Before turbine

Pmp Pump

k Stream number

gener Generated

SFPK Stream component fueling the product

524
GL Global

Greek Letters

𝝍 Exergy efficiency

η efficiency

ε heat exchanger effectiveness

λ Turbine blade tip speed (m/s)

x Concentration of the desiccant by mass (%)

° degrees

𝝎 Turbine rotational speed(rpm)

𝛼 absorption coefficient

REFERENCES

[1] ASHRAE, Handbook of Fundamentals, American Society of Heating, Refrigerating and Air-Conditioning
Engineers, Inc., Atlanta, GA, Chapter 28 (1993).

[2] P. Torcellini, S. Pless, and M. Deru, D.Crawley,“Zero Energy Buildings,a Critical Look at the Definition”,
ACEEE, Pacific Grove, California, USA(2006).

[3] S.S.Jung, Y.T.Kim, Y.B.Lee, “Measurement of Sound Transmission Loss byUsing Impedance Tube”,
Journal of the Korean Physical Society, 53, No. 2: pp 596-600(2008).

[4] J.Y.Chung and D.A.Blaser ‘Transfer Function Method of Measuring In Duct Acoustics Properties”,
Journal of Acoustical Society of America, 68, No.3 (1980).

[5]Wind Atlases of the World .http://www.windatlas.dk/Wold/About.html.

[6] I. Dincer, M. Rosen, “Exergy: Energy, Environment and Sustainable Development” Elsevier, Second
Edition (2007).

[7] Carrier Corporation “Hourly Analysis Program System Design Load Software”.Version:4.2-a (1999-
2004).

[8] A.Bejan,G. Tsatsaronis, M. Moran, “ThermalDesign and Optimization”John Wiley & Sons, (1996).

[9] Y.Censor. "Pareto Optimality in Multiobjective Problems," Applied Mathematical Optimization, 4: pp 41–
59(1977).

[10] Da.Cunha, E. Polak, "Constrained Minimization under Vector-Valued Criteria in Finite Dimensional
Spaces," Journal ofMathematical Analysisand Applications, 19: pp 103–124(1967).

[11] Deb. Kalyanmoy, "Multi-Objective Optimization using Evolutionary Algorithms," John Wiley & Sons,
(2001).

[12] L.A. Zadeh,"Optimality and Nonscalar-Valued Performance Criteria, "IEEE Transactions Automatic
Control, AC-8, No.1 (1963).

525
 NUMERICAL INVESTIGATION OF THE HEATING DISTRIBUTION EFFECT ON THE BOILING FLOW IN
THE BUBBLE PUMPS
1 2 2 1
Raoudha Garma , Youssef stiriba ,Mahmoud bourouis ,Ahmed Bellagi
1
: U.R .Thermique et thermodynamique des procédés industriels,
Ecole Nationale d’ingénieur de Monastir, E.N.I.M, Av. Ibn Jazzar, 5060 monastir, Tunisie
2
: Rovira i Virgili University Tarragona Spain

garma_raoudha@yahoo.fr
a.bellagi@enim.rnu.tn

ABSTRACT
In this paper, numerical simulations of the heating repartition effect on the boiling flow in vertical tube was
conducted with the commercial CFD (Computational Fluid Dynamics) package FLUENT. The Eulerian
multiphase flow framework model was used to model the phase’s interaction. User-Defined Functions
(UDFs) are provided to compute the wall heat transfer and to calculate inter-phase heat and mass transfer.
Total and partial heating are treated to approach to two different configurations of the bubble pump which
considered as one of the most important components of the diffusion absorption machine.
In the simulation results we present preliminary calculations with pure water as working fluid to compare void
fraction distribution, liquid-vapor velocities arrangement and temperature profiles.

INTRODUCTION
Based on multiple good features such as the absence of moving part which cause vibration and noise and
specially the possibility to operate with waste heat and solar energy, the diffusion absorption refrigerator
(DAR) invented by Platen and Munters [1] in the1920s has been experimentally and numerically investigated
over the years. In these absorption systems, the generator is considered as one of the most important
components and improving its performance will automatically contribute to the whole absorption system
performance improvement.
Detailed thermodynamic models for three DARs generator configuration and bubble pump were developed
and analyzed by computer simulation by A. Zohar et al. [2] to study their effect on the system performance. It
was found that the configuration that integrated both the generator and the bubble pump is of great interest.
A mathematical model for the forced convective boiling of refrigerant-absorbent mixtures in vertical tubular
generator is developed by B. Pasupathy et al. [3]. The generator is made up of two stainless steel coaxial
tubes of 1m length. A parametric analysis has also been performed to study the effect of various parameters
on the performance of the generator. Two-fluid model was employed by Zheshu Ma et al. [4] to describe the
two-phase flow and heat transfer processes in a two-phase closed thermosyphon. It was found that
numerically predicted flow patterns and distribution of parameters under different conditions show a good
agreement in the most part with experimental results. The some model was used by A. Bemhmiden et al. [5,
6] to investigate the influence of heat input on the uniformly heated bubble pump for different operating
conditions. The optimum heat flux was correlated as a function of the tube diameter and mass flow rate, and
the minimum heat flux required for pumping was correlated as a function of the tube diameter. The same
bubble pump configuration (uniformly heated tube) was numerically investigated by R. Garma et al. [7] using
the commercial CFD (Computational Fluid Dynamics) package FLUENT. It was found that the onset boiling
point is reduced from 0.43 to 0.0016 m and the void fraction at tube’s outlet is increased from 0.01 to 0.7
2
when increasing the wall heat input from 25 to 150 kW/m .
An experimental investigation of an air-cooled diffusion-absorption machine operating with a binary light
hydrocarbon mixture (C4H10/C9H20) as working fluids and helium as pressure equalizing inert gas is
presented by Ben Ezzine et al. [8]. A new concept of generator consisting in a separated boiler and bubble
pump instead of the usual combined generator is tested. The bubble pump is a single vertical tube (height:
800 mm, diameter: 8/10 mm) to which the heat input is restricted to a small zone in the bottom. The
experimental results show that the bubble pump exiting temperature as well as those of the major
components of the machine but the absorber is very sensitive to the heat power inputs to the bubble pump.
For bubble pump heat inputs from 170 to 350 W, the driving temperature varies in the range of 120-150 °C.
Experimental study and the theoretical thermodynamic simulation of the same absorption refrigerator
prototype with methylamine-water-helium were carried out by Mazzouz Souha et al. [9]. The test results
showed that cold at -5°C is effectively produced by the machine for heating temperature of about 105°C and
an ambient air of 22°C. The comparison of the experimental data with the theoretical model showed a good
agreement. Jakob et al. [10] reported that the indirectly heated generator with its bubble pump is the main
new feature of a solar heat driven ammonia-water diffusion-absorption cooling machine and that all the
prototypes constructed performed well.

526
DARs WORKING PRINCIPAL
To explain the working principal of the diffusion absorption refrigerator, we refer to the simplified schematic of
a DAR depicted in figure 1. The machine consists of a generator including a separated bubble pump and a
boiler, a rectifier, a condenser, an evaporator, a gas heat exchanger (GHX), an absorber and a solution heat
exchanger (SHX). The working fluids mixture is composed of ammonia as refrigerant, water as absorbent
and helium as pressure equalizing inert gas.
The strong solution coming from the receiver tank flows to the generator via a solution heat exchanger. The
single-phase solution is heated at the bottom of the bubble pump causing some ammonia to evaporate and
then small vapour bubbles begin to appear on the tube wall. As the solution is further heated continuously by
the heat flux Qpb, the vapour bubbles grow and merge forming gas plugs that push liquid slugs in the pump-
tube up. In this way, the two-phase solution is lifted up to the gas–liquid separator. When the two-phase solution
is further heated by the additional heat flux QB provided to the boiler, the solution gets further concentrated and
supplement quantity of vapor is generated.The resulting ammonia poor solution returns back to the absorber
through the SHX. The vapour leaving the bubble pump contains a small quantity of the solvent water which is
removed by condensation in the rectifier. The purified ammonia vapour is then condensed in the air cooled
condenser and then introduced in the evaporator. As this evaporator is charged with helium, the partial
pressure of the liquid refrigerant decreases and it evaporates at low temperature producing cold. While
proceeding to the evaporator exit, the ammonia continues to evaporate and its partial pressure rises. The
resulting ammonia-helium gas mixture becomes heavier as the ammonia still evaporating and so it drops
down in the receiver think located at the bottom of the air cooled absorber. There, the weak solution arriving
from the absorber’s top absorbs the refrigerant and as a result the vapour ammonia-helium mixture becomes
more and more lighter and it rises up to the top. The rich solution resulting at the bottom flows down toward
the generator.

Fig.1. Schematic representation of a DAR

CFD MODELLING AND RESOLUTION

The commercial CFD code FLUENT was used as a frame to perform the simulation. The interfacial forces
models and the wall boiling model described previously were implemented in the code through User-Defined
Functions (UDFs). The Eulerian multiphase flow framework model was used to model the phase’s
interaction. GAMBIT software was used to build and mesh a two-dimensional computational domain (2-D
geometry) with 5x500 uniform rectangular cells. Only half of the pipe is considered due to the symmetry at
the pipe axis. Unsteady state calculations with a time step of 0.1 s were fixed for all cases. SIMPLE algorithm

527
was carried out for the calculations of the pressure velocity-coupling and the first order upwind calculation
scheme was performed for the discretization of momentum, energy and volume fraction equations.
The saturation temperature is fixed to 425.15 K corresponding to the operating pressure (pressure outlet 15
bars). Fully-developed profile of velocity and sub-cooling temperature (Tsub =5K) is applied at the inlet. The
gravity acceleration is 9.8 m/s2. Wall thickness is fixed to 2mm.

BUBBLE PUMP HEATING CONFIGURATION

Total and partial heating are treated to demonstrate the heating repartition effect on the boiling flow in the
bubble pumps. In the two cases the bubble pump consists of a single vertical tube of 1 m height and 10 mm
outer diameter. The sub-cooled water (420.15 K) enters the system at the bottom, and then boils due to the
heat fluxes supplied from the pipe wall. The same amount of heat flux, Qbp = 2.827 kW, is applied to the
stainless-steel made tube wall as follow:

• In the first case we have a uniformly heated bubble pump as demonstrated in Fig.2 and,
• In the second case the heat flux is supplied to 30 cm of the tube length (≈ 1/3 of the total tube’s
length) as indicated in fig.3.

Fig.2. First bubble pump configuration: Uniformly heated bubble pump

Fig.3. Second bubble pump configuration: Partially heated bubble pump

528
RESULTS AND DISCUSSION
In the simulation results we present preliminary calculations with pure water as working fluid to compare void
fraction distribution, liquid-vapor velocities arrangement and temperature profiles corresponding to heat input
of Qbp = 2.827 kW

0,7

0,6
Heated
Adiabatic zone
zone
0,5

0,4
Void fraction
0,3
Partial heating
0,2 Total heating

0,1

0,0
0,0 0,2 0,4 0,6 0,8 1,0
Tube's length (m)

Fig.4.a. Axial void fraction evolution for total and partial heated tube

Fig.4.a. illustrates the axial void fraction evolution for total and partial heating cases. One can see that the
boiling onset in the partially heated tube is clearly lower than that for total heated one and this can be
explained by the higher heat density in the bottom of the tube. One can also observe that for uniformly
heated bubble pump the void fraction increases slightly over the whole length of the tube, attains a value of
0,1 when L=30 cm ( L is the tube´s length) and a maximum of 0,55 at the tube outlet. This behavior is
completely different from that of the partially heated pump. In fact, for this case the void fraction rises sharply,
reaches a value in the order of 0,6 when L=30cm and then it remains approximately constant over the
remaining length of the tube.

The bubble pump, consisting on a vertical heated tube, may present more than one flow regime. Identifying
these flow regimes can be determined by referring to the limit values of the void fractions. The critical void
fractions corresponding to the bubbly-slug, slug-churn and churn-annular transition are respectively 0.3, 0.55
and 0.8 [11, 12]. Fig. 4.b shows the limits of flow regimes from the curve of the void fraction evolution. One
can the effect of the heat distribution on the flow régime repartition along the tube.

529
 1,0
0,9 Annular
0,8
0,7 Churn
0,6
Void fraction
0,5
0,4 Slug
0,3
0,2
0,1 Bubbly Partial heating
Total heating
0,0
0,0 0,2 0,4 0,6 0,8 1,0
Tube's length (m)
Fig.4.b. Flow pattern prediction

Radial temperature profiles at different tube locations for total and partial heating are respectively depicted in
Figure 5.a and fig.5.b. One can observe for the first case (fig.5.a) that the liquid temperature profiles show
the same behavior for the various locations. Actuality, the liquid at the center of the tube is subcooled even at
the outlet (z = 1 m) and it is superheated near to the wall. We can see also that the subcooling level is
decreasing throughout the tube’s length. Referring to fig.5.b, we can remark the same temperature evolution
As the total heated tube until L=0.3 m (heated zone) except that the superheat near to the wall is clearly
higher in this case. Whereas for L upper than 30 cm (the riser) the temperature remain constant (adiabatic
zone)

Fig. 5.a: Radial liquid temperature profile for total heating

530
 Fig. 5.b: Radial liquid temperature profile for partial heating

Liquid and vapor velocities arrangement for total and partial heating are respectively illustrated in Figure 6.a
and fig.6.b. In the first case, one can remark that the liquid velocity increases slightly, it goes from 1.2 to 1.3
m/s from the entrance to 0.4 m, then increases sharply to the rest of the tube, it reached 2.9 m/s at the exit.
The steam follows almost the same profile as the liquid. In the second case, the liquid-vapor velocity follows
the same behavior of the void fraction. In fact, it increases rapidly in the heated area then it increases slightly
(almost constant) on the rest of the tube.

3,2
2,8
2,4
Axial velocity (m/s)

2,0
1,6
1,2 Liquid phase
Vapor phase
0,8
0,4
0,0
0,0 0,2 0,4 0,6 0,8 1,0
Tube's lenght
Fig. 6.a: Liquid and vapor velocities arrangement for total heating

531
 3,5

3,0
Heated zone Adiabatic zone
Axial velocity (m/s) 2,5

2,0

1,5

1,0

0,5 Liquid phase

Vapor phase
0,0
0,0 0,2 0,4 0,6 0,8 1,0
Tube's lenght (m)

Fig. 6.b: Liquid and vapor velocities arrangement for partial heating

CONCLUSIONS
numerical simulations of the heating distribution effect on the boiling flow of water in vertical tube was carried
out with the commercial CFD (Computational Fluid Dynamics) package FLUENT. User defined Functions
(UDFs) are employed to model the boiling phenomena. Void fraction distributions, liquid-vapor velocities
arrangement and radial temperature profiles are calculated and discussed. It was found that the the void
fraction is higher when heating partially the wall. Flow regimes repartitions are identified referring to the void
fraction variation along the tube. The liquid temperature in the middle part of the tube is still sub-cooled along
the pipe and it is superheated near to the wall so bubbles migrate from the heated wall to the center of the
tube and it still constant along the riser when the tube is partially heated.

REFERENCES
[1] Von Platen BC, Munters CG. 1928. US Patent 1: 685-764.
[2] A. Zohar, M. Jelinek, A. Levy, I. Borde. 2008. The influence of the generator and bubble pump
configuration on the performance of diffusion absorption refrigeration (DAR) system. International journal of
refrigeration 31: 962–969
[3] Balamurugan PASUPATHY, Mani ANNAMALAI. 2010. Numerical Studies on Vertical Tubular Generator in
Vapour Absorption Refrigeration System. International Refrigeration and Air Conditioning Conference at
Purdue, July 12-15, 2010
[4] Zheshu Ma, Ali Turan, and Shengmin Guo. 2009. Practical Numerical Simulations of Two-Phase Flow
and Heat Transfer Phenomena in a Thermosyphon for Design and Development. G. Allen et al. (Eds.): ICCS
2009, Part I, LNCS 5544, pp. 665–674, 2009.© Springer-Verlag Berlin Heidelberg 2009
[5] A. Benhmidene, B. Chaouachi, M. Bourouis, and S. Gabsi. 2011. Effect of Operating Conditions on the
Performance of the Bubble Pump of Absorption-Diffusion Refrigeration Cycles. THERMAL SCIENCE, Vol.
15, No. 3, pp. 793-806
[6] A. Benhmidene, B. Chaouachi, M. Bourouis, and S. Gabsi. 2011. Modelling of heat flux received by a
bubble pump of absorption-diffusion refrigeration cycles. Heat Mass Transfer. 47: 1341–1347
[7] R. Garma, Youssef stiriba, M. Bourouis, and A. Bellagi. 2012. Numerical Investigation of Nucleate Boiling
Flow in Vertical Tubes, 2ème Colloque International sur l’Énergie CIE 12,Tozeur (Tunisie) Le 26-27-28 mars
2012
[8] N. Ben Ezzine, R. Garma, M. Bourouis, A. Bellagi. 2010. Experimental Studies on Bubble Pump Operated
Diffusion Absorption Machine Based on Light Hydrocarbons for Solar Cooling, Renewable Energy, 35
(2010), 2, pp. 464-470
[9] S. Mazouz, N. Ben Ezzine, R. Garma, M. Bourouis, A. Bellagi. 2008. Experimental Investigation and
Theoretical Model of a Diffusion Solar Absorption Machine International Sorption Heat Pump Conference,
23-26 September, 2008, Seoul, KOREA

532
[10] U. Jakob, U. Eicker, D. Schneider, A.H. Taki, and M.J, Cook. 2008. Simulation and experimental
investigation into diffusion absorption cooling machines for air-conditioning application. Applied Thermal
Engineering. 28: 1138–1150
[11] K. Mishima, M. Ishii, 1984, Flow regime transition criteria for upward two-phase flow in vertical tubes,
International Journal of Heat and Mass Transfer. Vol. 27, No. 5, pp.723-737.
[12] N. Brauner, D.A. Barnea, 1986, Slug-to-churn transition in upward gas-liquid flow, Chemical Engineering
Sciences, 41, 159-163

533
OPTIMIZED ANALYSIS OF COLD ENERGY UTILIZATION FOR COLD STORAGE PROJECT OF
XINGTAN LNG SATELLITE STATION

Wendong XU, Zhonghao HUANG, Shuanshi FAN

South China University of Technology, School of Chemistry and Chemical Engineering, Key Laboratory of Ministry of
Education on Enhanced Heat Transfer and Energy Conservation
Wushan road No. 381, Tianhe District, Guangzhou City, 510640, Guangdong province, China
E-mail: hzhprince@163.com

ABSTRACT
The project of cold energy utilization for cold storage of Xingtan LNG satellite station is the first cold energy
utilization demonstration project of LNG satellite station in China with 2～4×104 m3/d gasification rate of LNG and
10～15 t/d supply of liquid ammonia in a temperature range of -25～-38℃. Its innovation lies in the point of adopting
two ammonia refrigeration circulation systems to recover LNG cold energy for cold storage at the same time.
Although this project had operated smoothly for more than two years after commissioning, there still existed some
problems. Such as refrigerant frozen blocking problem and undersupply of cold energy which caused by
underperforms coordination between the two ammonia refrigeration circulation systems. According to the
equipment running state and automatic control system data, this article presented optimized analyses of the whole
process, including using a redesigned ammonia refrigerant flow channel and achieving dynamic balance of
ammonia refrigeration to relieve the refrigerant frozen blocking phenomenon, improving production process and
optimizing process parameters to ameliorate the condition of cold supply shortage problem. Combined with above
research, many new methods were proposed which can effectively solve above problems in the utilization of LNG
cold energy so as to increase the LNG cold energy utilization ratio and enhance the safety and stability of project
operation. In addition, it provides a valuable reference for design, development and construction of other analogous
LNG cold energy utilization projects in China.

KEYWORDS: LNG; Cold Energy; Cold Storage; Optimized Analysis

INTRODUCTION
With the energy security and environmental pollution problem becoming more and more serious, liquefied
natural gas (LNG) industry and energy-saving projects is developing rapidly in China. LNG need to be heated and
gasified before it supplying to downstream users with a cold energy release of 830 kJ/kg[1]. If the price of electricity
is calculated as 0.85 yuan/(kW•h), LNG of -162 ℃ contains cold energy value of about 510 yuan/t. So making full
use of LNG cold energy can not only save enormous amounts of electricity for refrigeration, but also be helpful to
reduce the costs of LNG gasification and expand natural gas downstream market[2].
There are many theoretical studies on the LNG cold energy utilization technology at present, mainly including
the cold energy used for power generation, air separation, cryogenic freezing (cold storage and cold water), etc[3].
So far, the LNG cold energy utilization has successfully realized the application for power generation, air separation
and cryogenic freezing project in the abroad. But in domestic, the LNG cold energy utilization projects are still in the
demonstration research stage and the development of technology is relatively backward. In order to improve this
actuality, we developed a technical solution with independent intellectual property rights about LNG cold energy
utilization on the basis of fully absorbing domestic and foreign experience, and successfully applied it to Xingtan

534
LNG satellite station in Shunde district, Foshan city. After having operated smoothly for more than 2 years, the cold
energy utilization project of Xingtan LNG satellite station reflected its well technical feasibility and practical
operability. The project can recycle LNG cold energy effectively and creating economic benefits at the meantime.
However, some problems still exist in the actual operation of the project reflecting some shortcomings of the
technical design.
Therefore, this paper analyzed the LNG cold energy recycling system of Xingtan LNG satellite station and put
forward some constructive improvement and optimization solutions. It can not only perfect the overall scheme and
solve the existing problems in the current project, but also provide a very valuable reference to the design,
development and construction for similar projects of LNG cold energy utilization in our country. All in all, it is of great
significance for promoting the work process of energy-saving and emission reduction in China.

1. THE PROJECT OVERVIEW OF LNG COLD ENERGY UTILIZATION FOR COLD STORAGE IN XINGTAN
LNG SATELLITE STATION

1.1. Project background and gas station situation

Xingtan LNG satellite station which affiliated to Hong Kong China gas co. LTD. is a supplementary gas
regulating station for the insufficient daily average and peak flow in Shunde district, Foshan city. The current
amount of LNG gasification is 2 ~ 10 × 104 Nm3/d with a maximum supply capacity of 24×104 Nm3/d, the gasification
pressure is 0.4 ~ 0.7 MPa and the pipeline transmission pressure is 0.3 MPa.
There are six cylindrical LNG storage tanks with 100 m3 cubage in the station, including four tanks with storage
pressure of 0.35 MPa and two tanks with storage pressure of 0.55 MPa. The station also has eight air-heated
evaporators with a single gasification capacity of 2000 Nm3/h. During daily operation, a mass of mist condensed
around the evaporators, which caused a certain degree of cold pollution on the station environment.
Adjacent to the station, there is a cold storage with the total capacity of 3000 tons. The cold storage which is
divided into a freezer of - 30 ℃and a refrigerator of 15 ℃is mainly engaged in aquatic products process and storage.
The cold storage has three ammonia compression refrigeration units those should be turned on selectively
according to the quantity of aquatic products process and the process requirements. Generally, two refrigerating
installations will be turned on and run 15 ~ 16 h per day to supply cold quantity of 4890 ~ 5216 kW · h in summer,
while one will be turned on and run 18 ~ 20 h per day to supply cold quantity about 3100 kW • h in winter, and the
Electricity fee is about 120×104 yuan all the year round[4].

1.2. Technological process description

In the light of the characteristics of Xingtan LNG satellite station and the needs of downstream cold users, the
scheme of LNG cold energy recovery for cold storage has been chosen for the station cold energy utilization project.
The technical solution mainly includes three parts: LNG gasification system, low pressure ammonia refrigeration
circulation and ammonia electric compression refrigeration circulation. The process flow diagram is shown in Fig.1.

535
 0.3~0.35 Mpa
NG City Gas Pipeline
T1 LNG 25 ℃
（-162℃；0.4~0.6 MPa）
c NG
d V-1 25 ℃
H1

E1 NG
V-2 H2
-25~ -30℃

Pressure Sensor

V-4
T2 Liquid Ammonia
（-40℃，0.1 MPa）
Ammonia
a
（-20℃，0.1 MPa）
Temperature Sensor b
Cold Storage
V-3
-40℃;0.15 MPa V-5

0.4~0.6 MPa

E2
126℃;1.4 MPa

C1

Cooling Water（30℃） 40℃

V-6 Liquid Ammonia（35℃）

V-7

T3

Fig.1. The process flow diagram of LNG cold energy utilization technology for cold storage

LNG out from the storage tank (T1) will be divided into two streams, one stream flow into the air-heated
evaporator (H1) directly and evaporate into gas of 25 ℃, then it will pass through the pressure regulating valve and
be regulated to 0.3～0.35 MPa. The other stream will flow into heat exchanger (E1) first and exchange heat with
ammonia gas from cold storage of - 20 ℃, 0.10 MPa, then the LNG stream become into gas of - 25~ - 30 ℃ and flow
into the air-heated evaporator (H2) to evaporate to 25 ℃. After mixing with the former stream and being regulated
pressure, the gas will enter the urban gas pipe network in the end. In the meanwhile, Ammonia gas will be
condensed into liquid ammonia of - 40 ℃ and flow into the liquid ammonia tank (T2), then the pump pressurize the
liquid ammonia to 0.3～0.5 MPa. The stream will flow via flow-meter to pressure regulating valve, dropping to 0.15
MPa, and finally flow into the cold storage for cooling supply[5].
When the cold quantity demand of cold storage in a larger amount or the gasification quantity of LNG in a short
supply, the electric compression refrigeration circulation of ammonia will be turned on and the two ammonia
refrigeration circulation systems will operate together. In the electric compression ammonia refrigeration circulation
system, ammonia gas out from cold storage will transform into liquid ammonia of 35 ℃ through twice compression
and heat exchange with cold water in heat exchanger (E2). Then after being regulated to 0.15 MPa by the pressure
regulating valve with the temperature dropping to - 40 ℃, the liquid ammonia stream will mix with the stream out
from low pressure ammonia refrigeration circulation and finally flows into the cold storage for cooling supply.

1.3. Project operation situation

The project was completed in May 2010 and began to process debugging in June. One mouths later, the
technology process basically achieved the design requirements with the LNG gasification rate of 2 ~ 10 × 104 Nm3/d,

536
provided liquid ammonia temperature of - 25 ~ - 38 ℃ and Liquid ammonia flow rate of 10 ~ 15 tons/day. After three
months of commissioning, the project has been put into production officially. So far the system has been running for
more than 2 years, no major accidents occurred and equipments were all in good condition and working properly
during that period. What’s more, the system obviously alleviated the cold-fog phenomenon in the station and
continued supplying reclaimed cold energy of LNG to cold storage for its normal operation.
The project construction was invested by Xingtan LNG satellite station, and the cold storage users bear the
cost of cold supply by taking liquid ammonia flow rate as the settlement measurement. The annual sales income of
the project is 850000 yuan/year and the average annual after-tax profit is 300700 yuan/year, so it shows a good
economy.
In addition, as the first domestic engineering example of LNG cold energy used for cold storage, the project
passed the recognition of scientific and technological achievements by experts of gas professional committee,
Guangdong institute of petroleum, and won the Hang Seng pearl river delta environmental protection awards issued
by Hong Kong industrial association. The success of the project has attracted extensive attention among the LNG
industry.

1.4. Analysis of the existing problems in the system

1.4.1. Refrigerant frozen blocking

From the operation report recorded by the operation department of Xingtan LNG cold energy utilization project
group, we have known that during the actual running, the system often appeared the problem that part of the liquid
ammonia solidified and blocked pipes in heat exchanger. Refer to the designed process, the inlet of LNG is at the
bottom of the heat exchanger and the outlet of NG is at the top, while the inlet of ammonia is at the top left of the
heat exchanger and the outlet of liquid ammonia is at the bottom right so as to facilitate their counterflow heat
exchange. Because of the freezing point of ammonia is -77.73 ℃ and the inlet temperature of LNG which would
exchange heat with ammonia is - 162 ℃, especially as the inlet of LNG located closely to the outlet of liquid
ammonia, that will cause liquid ammonia partial undercooling by LNG in the process of heat exchange and result in
liquid ammonia freeze between the outlet at the bottom of heat exchanger and tube plates. The frozen layer would
thicken with the extension of time and eventually make the blockage of liquid ammonia pipeline, leading to
equipment damage. The problem of refrigerant frozen blocking seriously influenced the thermal efficiency in heat
transfer system, hindered the normal operation of the system as well[6].

1.4.2. Cold energy undersupply

From the design perspective of the whole system，the cold storage refrigeration system would not affect the
overall LNG gasification quantity and the control operation of gas pressure, thus the system has the advantages of
more simple operation, less investment and so on. But as the project only has one heat exchange system, the
refrigerant exchange heat with LNG directly and recycle cold energy to supply for cold storage cooling capacity. The
system can only adjust the flux of LNG flow based on the actual cold capacity requirement of cold storage, and it
can't adapt to the volatility of gasification quantity of the station.
Moreover, although the electric compression ammonia refrigeration system as a backup control system can
replenish cold energy to cold storage with the method of electric compression refrigeration when the cooling supply
is insufficient, the coordination and operation of the two ammonia refrigeration systems was not that effectively in
the actual operation. In addition, the cold storage continuously was expanding year by year. By the end of 2012, the
capacity of cold storage has been increased to 8000 tons, which made the problem of cold energy undersupply

537
increasingly outstanding.

2. THE DESIGN OPTIMIZATION OF SYSTEM PROCESS

2.1. The design optimization of heat exchange system

Aiming at the problem of refrigerant frozen blocking, the following optimization measures in several aspects
are put forward:
(1) In the heat exchange system, LNG of - 162 ℃ will first flow through the heat exchanger and pass cold
energy to ammonia gas, then enter the evaporator and be heated up again. Under the circumstances, there is a big
heat transfer temperature difference between the two heat transfer medium in the heat exchanger, and the
temperature of LNG is far below the freezing point of refrigerant. So it is very likely to lead to the refrigerant frozen
blocking phenomenon. The key to solve this problem is the optimal selection of refrigerant and redesign of pipelines.
Through the comparison and analysis we found that choosing the refrigerant which can resistant to low temperature
much more, such as butane or R404A, can effectually avoid the solidification of refrigerant. What else, the
horizontal heat exchanger should be adopted in the design of heat exchange system, and strictly monitoring should
be taken for the fluid pipeline to prevent the infiltration of impurities and water in the process of equipment
operation.
(2) Through the practical observation and operation experience, we also found that when the flow rate of
refrigerant (ammonia) was less than 30% of design value, the flow velocity of refrigerant kept slow and the unit heat
load increased; When the flow rate of refrigerant was more than 140% of design value, the pressure drop of the
heat exchanger would increase more than 0.05 MPa. These are most possible to cause the refrigerant freeze. So
during the running of the equipment, it is necessary to make appropriate regulation of the refrigerant flow rate and
keep the fluctuation of flow rate in the allowed range for avoiding refrigerant frozen blocking.
(3) In order to deal with the refrigerant frozen blocking problem occurred in the actual operation process, a
thaw line should be setup in the heat exchange system. When the problem happens, LNG pipe line will be shut off
first and then introduce part of natural gas of 25 ℃ from the original gasification process into the LNG flow channel
through this thaw line to purging melted the frozen refrigerant until the refrigerant circulation restore clear.

2.2. The design optimization of refrigerant circulation

Aiming at the problem of cold energy undersupply, the following optimization measures in several aspects are
put forward:
(1) Because of a lack of the intermediate refrigerant circulation, the whole system lacks the function of cold
storage and peak shaving. In the current designed system, refrigerant storage tank is the sole equipment that can
temporarily store cold energy, so we need to modify it. The transformation method is tank capacity expansion:
increasing the volume of tank from 3 m3 to 15 m3. Using the new designed storage tanks of large capacity can make
it carry more refrigerant and ensure that there is plenty of refrigerants to transfer heat with LNG and provide cold
energy to the cold storage continuously. What’s more, increasing the capacity of refrigerant tank may also avoid the
liquid ammonia pump cavitation caused by the flash of internal liquid ammonia in storage tank to a certain extent.
(2) In the long run, the better solution is adding a set of refrigerant circulation devices in the system and using
the intermediate refrigerant cycle for the purpose of cold storage and peak shaving[7]. The redesign process is
shown in Fig.2.

538
 P-41 0.3~0.35 Mpa
NG City Gas Pipeline
T1 LNG 25 ℃
（-162℃；0.7MPa）
c
d H1

E1 NG
P-47 0~ 5℃ H2
Pressure Sensor
-120℃ 20℃；0.5MPa
E3

Circulating Water 25℃；0.15MPa -10℃；0.15MPa

T2

Butane E2 T3
-120℃；0.5MPa -26℃ -26℃；0.5MPa

P-64 P3
P1

V-13
Liquid Ammonia
T4
（-45℃，0.15 MPa）

Ammonia
Temperature Sensor （-20℃，0.15 MPa）
Cold Storage

V-3 0.15 MPa

0.4~0.6 MPa

P2

Fig.2. The optimized process flow diagram of LNG cold energy utilization technology for cold storage

Firstly, LNG of - 162 ℃, 0.7 MPa, 162 kg/h will exchange heat with the refrigerant (butane) of 20 ℃, 0.5 MPa,
3000 kg/h in heat exchanger (E1). LNG will be warmed up to 2 ℃after gasification and flow into the city gas pipeline
network after pressure regulating; Butane will be cooled down to -120 ℃ and enter the low temperature liquid phase
refrigerant tank (T2). Then, Butane will be pressurized to 0.5 MPa by the pump (P1) and exchange heat with
ammonia of - 20 ℃, 0.15 MPa, 380 kg/h out from cold storage in heat exchanger (E2). Butane will be warmed up to -
26 ℃ and enter the liquid refrigerant tank (T3); Ammonia gas will be cooled into liquid ammonia of - 45 ℃ and enter
the liquid ammonia tank (T4). Finally, Butane will be pressurized to 0.5 MPa by the pump (P2) and exchange heat
with circulating water of 25 ℃，0.15 MPa，5000 Kg/h in heat exchanger (E3). Butane will be warmed up to 20 ℃and
flow back to heat exchanger (E1) to complete the intermediate refrigerant circulation; Circulating water will be
cooled down to 10 ℃. Meanwhile, liquid ammonia of - 45 ℃ will be pressurized to 0.2 MPa by the pump (P2) and
enter the cold storage for cooling supply.
According to the analysis computation by Aspen program, above scheme can ensure the system operate well
with the fluctuation of LNG gasification quantity and still meet the dynamic needs of cold storage by appropriate
adjusting. The addition of intermediate refrigerant circulation will directly increase the total cost and the initial
investment of the project and make the operation more complex. But for the growing amount of LNG gasification
quantity and cold energy demand of cold users, this design can not only solve the imbalance problem of cold
energy supply and demand, but also make the adjustment and distribution of cold energy utilization more
reasonable.

3. THINKING AND SUGGESTIONS ON THE DEVELOPMENT OF COLD ENERGY UTILIZATION PROJECT IN

LNG SATELLITE STATION
As for the development and construction of current domestic LNG cold energy utilization projects, both the
depth and the scale of cold energy utilization still have a very big development space. Through the case study of the

539
first LNG satellite station cold energy utilization project in China: cold energy utilization for cold storage in Xingtan
LNG satellite station, we raised some thinking and suggestions on the development of current domestic cold energy
utilization project of LNG satellite station below.

3.1. Choosing appropriate refrigerant to adapt to different occasions

In the technical process of LNG cold energy recycling, the selection of refrigerant is one of the key
technologies. The desired refrigerant not only needs to meet the requirements of some specific physical and
chemical properties, but also should be good economical and environmental friendly. At present, the domestic and
foreign experts have done a lot of researches about the circulating refrigerant, and proposed many available
refrigerants which can be applied to LNG cold energy utilization technology such as ammonia, glycol aqueous
solution and R410a[8], etc. But for the specific application occasions and design requirement, we should considerate
comprehensively containing the function of refrigerant circulation system, the equipment and operating cost, the
control temperature and so on before choosing the optimal refrigerant. For example, R404a or butane can be a
good choice used as the refrigerant in the low temperature refrigerant circulation system, while ammonia or
ethylene glycol aqueous solution can be a good choice used as the refrigerant in high temperature refrigerant
circulation system.

3.2. Setting the intermediate refrigerant heat circulation to realize cold storage and peak shaving.
LNG satellite stations have the characteristics of wide distribution, small but wildly fluctuant flow rate and
indefinite operation time[9]. So if want to fully recover and utilize the cold energy of LNG, it needs to setup a set of
device for cold storage and peak shaving in design of LNG cold energy utilization projects. That is using an
intermediate refrigerant circulation to store excess cold energy in the gas supply peak and provide the stored cold
energy to cold users when LNG gasification quantity decrease in gas supply low ebb. In this way, the entire system
can run smoothly and meet the demand of cold users.
The refrigerant circulation system is mainly composed of four parts: low temperature storage tank (containing
phase change material), high temperature storage tank, heat exchanger for refrigerant and gas, heat exchanger for
low temperature refrigerant and high temperature refrigerant. The refrigerant circulation system can adjust the
supply and demand of the recovery and utilization of LNG cold energy. Especially for the long-term development of
LNG cold energy utilization project, the utilization of cold energy from LNG gasification and the energy matching in
different temperature levels with other cold users all depend on the operation of multistage refrigerant circulation to
achieve. Therefore, setting intermediate refrigerant circulation in the cold energy utilization system is not only the
requirement of cold storage and peak shaving of LNG cold energy, but also the needs of the development of LNG
cold energy utilization technology.

3.3. Planning comprehensive utilization scheme to promote the cascade utilization of LNG cold energy
As we know, according to the principle of refrigeration, the lower the requirement of the process temperature is,
the more power consumed by a conventional refrigeration way takes. Within a certain low temperature region, every
1 K decrease of the evaporation temperature will increase the energy consumption by 10%. So at that point, the
energy saving effect of LNG cold energy utilization will be more obvious and the utilization ratio of cold exergy will
be higher[10]. Thus it can be seen, if we only use LNG cold energy for cold storage refrigeration that means only the
high temperature part of LNG cold energy has been recycled and the utilization of the low temperature part still has
great development space, which makes the low utilization ratio of cold exergy.
Consequently, we need to design a comprehensive utilization of LNG cold energy scheme in consideration of

540
the characteristics of LNG satellite stations. First, establish the refrigerant circulation system to recover LNG cold
energy and store it appropriately. Then, provide the stored cold energy for the cold users based on the principle of
"temperature counterpart and cascade utilization" by matching with their cold needs in different temperature
sections. The low temperature section of cold energy can be used for air separation or cryogenic crushing, and the
high temperature section of cold energy can be used for cold storage or cold water air conditioning[11]. Through this
way LNG cold energy can be made full use of and the cold exergy utilization ratio will be improved significantly.

CONCLUSIONS
(1) The cold energy utilization for cold storage project of Xingtan LNG satellite station, which was designed by
Shunde, Hong Kong China gas co. LTD. and South China University of Technology, is the first LNG cold energy
utilization project of LNG satellite station in China. In this project, the cold energy generated during LNG gasification
would be recycled by ammonia and applied to cooling the cold storage so as to save electricity for refrigeration. The
annual profit of the project is about 300,700 yuan. After over two years’ practical running, the system has shown a
pretty good result of cold energy utilization. There is no doubt that the project has great benefits of technical,
economic and social.
(2) According to the practical operation condition of the cold energy utilization for cold storage project of
Xingtan LNG satellite station and the analysis of the equipment running state and automatic control system data,
some corresponding optimization measures were proposed: redesigning the heat exchange system to ease the
problem of refrigerant frozen blocking, adding an intermediate refrigerant circulation system to solve the situation of
insufficient cold supply by re-matching the cold energy with the downstream cold users, which will ultimately make
the device running smoothly and orderly.
(3) Concerned with the characteristics of the cold energy utilization project of LNG satellite station, we put
forward some suggestions on long-term development of LNG cold energy utilization: choosing the proper
refrigerant and setting multistage refrigerant circulation system to establish the integrated technology of LNG cold
energy comprehensive utilization. The integrated technology will recycle LNG cold energy in different temperature
ranges more effectively, match the cold-requiring systems more reasonably and gain maximum economic benefits.
This paper would provide a valuable reference to the similar LNG cold energy utilization projects in China.

ACKNOWLEDGEMENT
Here we express sincerely acknowledge to Hong Kong China gas co. LTD. of Shunde District, Foshan City for
facilitating this paper in research section by providing parameters of gas station and the system running condition.

REFERENCES
[1] Gang ZHU and Anzhong GU. 1999. The utilization of liquefied natural gas (LNG) cold energy. Energy
Engineering 3: 1- 2.
[2] Jing LI, Zhihong LI and Ben HUA. 2005. The current situation and development prospect of LNG cold energy
utilization. Natural Gas Industry 05:103-105+14.
[3] Qiang WANG, Yanzhong LI and Zhaochang ZHANG. 2002. The recovery and utilization of liquefied natural gas
(LNG) cold energy. Cryogenic Engineering 128(4): 38- 42.
[4] Zongbin LIU, Huiping ZHENG, Wei SHANG and Wendong XU. 2010. The project development of cold energy
utilization in LNG satellite station. Gas & Heat 09:1-5.
[5] Zongbin LIU, Jianwei HUANG and Wendong XU. 2010. The technology development and demonstration of
cold energy utilization for cold storage in LNG satellite station. City Gas 09:8-11.

541
[6] Jiansheng TAO. 2011. The technology design of LNG cold energy used for cold storage. Guangdong Chemical
Industry 06:179-180.
[7] Shulong ZHANG, Lin JIAO, Xiujuan LI, Yanzhe NIE and Changgui DUAN. 2007. The research of the cold
storage system using LNG gasification cold energy of LNG station. Gas & Heat 12:15-17.
[8] Linlin DU, Dongxiao LUO and Wendong XU. 2007. The research of LNG cold energy utilization technology for
cold storage in south China. Natural Gas Industry 06:115-117+160.
[9] Wendong XU, Haijun BIAN and Shuanshi FAN. 2009. The cold energy utilization technology of LNG station
satellite. Natural Gas Industry 29(5): 112-14.
[10] Xianwen TANG, Cong WANG and Zeliang YANG. 2012. The exergy analysis of Guangdong LNG cold energy
and its application in cold storage. Energy and Energy Conservation 05:12-13+29.
[11] Haijun BIAN. 2011. The research and process analysis of liquefied natural gas (LNG) cold energy utilization
technology. South China University of Technology.

542
 A COMPUTER BASED ANALYSIS OF SINGLE-STAGE
VAPOR ABSORPTION REFRIGERATION SYSTEM
1,* 1 1 2
O. Kaynakli , N. Arslanoglu , R. Yamankaradeniz , and N. Yamankaradeniz
1
Uludag University, Faculty of Engineering-Architecture, Department of Mechanical Engineering, 16059, Bursa, Turkey
2
Uludag University, Vocational School of Technical Sciences, Air Conditioning and Refrigeration Technology Program,
16059, Bursa, Turkey
email: kaynakli@uludag.edu.tr, narslanoglu@uludag.edu.tr, recep@uludag.edu.tr, nyk@uludag.edu.tr
*
Corresponding Author: (O. Kaynakli) Tel: + 90 224 2941984, Fax: + 90 224 2941903

ABSTRACT
This study focuses on the thermodynamic analysis of the single-stage Vapor Absorption Refrigeration (VAR)
systems using NH3-H2O and H2O-LiBr solutions. A mathematical model was developed and a computer
program written in visual programming language Delphi was prepared. Students of engineering faculties and
researchers can easily use this user friendly, practical and attractive computer program. This program allows
analyzing in detail the effect of main system temperatures, the solution pump efficiency and the effectiveness
of the heat exchangers on system performance and working conditions. This study explains this computer
program in detail using the flow chart, figures and tables. Furthermore, some of the results were presented in
tabular and graphical form and discussed.

1. INTRODUCTION
A Vapor Absorption Refrigeration (VAR) System is similar to a Vapor Compression Refrigeration (VCR)
System. In both systems the required refrigeration is provided by refrigerants vaporizing in the evaporator.
However, in the VAR System, a physico-chemical process replaces the mechanical process of the VCR
system and heat rather than a mechanical and electrical energy is used. The advantages of this system lie in
the possibility of utilizing of waste energy from industrial plants as well as of using solar energy.

The VAR cycle uses a refrigerant-absorbent solution rather than pure refrigerant as the working fluid. The
absorbent acts as a secondary fluid to absorb the primary fluid which is the refrigerant in its vapor phase.
The VAR cycle widely uses ammonia-water (NH3-H2O) solution with ammonia as the refrigerant and water-
lithium bromide (H2O-LiBr) solution with water as the refrigerant. The ammonia-water VAR system has been
used in large capacity industrial applications requiring low temperatures for process cooling. The water-
lithium bromide vapor absorption system is used to produce hot water for comfort heating, process heating
and domestic purposes, as well as for cooling. This system can also be used to deliver heat at a temperature
higher than that of the driving heat source, but it is predominantly for air-conditioning applications. The main
problem with the water-lithium bromide pair is the possibility of solid formation. Since the refrigerant turns to
ice at 0°C (32°F), the pair cannot be used for low temperature refrigeration. Furthermore, lithium bromide
crystallizes at moderate concentrations. When the absorber is air cooled, these concentrations tend to be
reached; thus, the pair is usually limited to applications in which the absorber is water cooled. It is possible
that the use of a combination of salts as the absorbent will reduce the crystallizing tendency enough to
permit air cooling.

A detailed thermodynamic analysis of H2O-LiBr absorption refrigeration cycle was performed in Kaynakli and
Kilic (2007). The influences of operating temperature and effectivenesses of solution and refrigerant heat
exchangers on the thermal loads of components, coefficients of performance (COPc, COP) and efficiency
ratio were investigated. Sun (1998) investigated alternative solutions of NH3-H2O solution and after having
carried out a thermodynamic analysis, he investigated the solution performance when mixing ammonia with
different absorbents. A similar study was carried out by Karamangil et al. (2010). They investigated the VAR
system performance using commonly encountered solution pairs in the literature. Keçeciler et al. (2000)
carried out experiments on the VAR using H2O-LiBr driven by geothermal energy. Mostafavi and Agnew
(1996) investigated the performance of the VAR system using H2O-LiBr depending on the ambient
temperature. Saravanan and Maiya (1998) analysed the system using water as refrigerant coupled with
different absorbents. Horuz (1998) carried out an investigation to analyze the VAR systems using NH3-H2O
and H2O-LiBr solutions and compared them in respect of the coefficient of performance (COP), the cooling
capacity and the maximum and minimum system pressures. Xu and Goswami (1999) presented a new
method for calculating the thermodynamic properties of ammonia-water mixtures using Gibbs free energies
and empirical equations for bubble and dew point temperature.

543
 The simulation of the VAR systems using NH3-H2O and H2O-LiBr solutions and the calculation of the COP
for different working conditions are quite complex and difficult. There is a possibility to make a mistake when
calculating the thermodynamic properties of refrigerants and solutions for different working conditions using
calculator. For that reason, it will be good to have a user-friendly computer program, calculating every data,
one can need to analyze the system.

This study explains the detailed thermodynamic analysis procedure of the single-stage VAR systems using
NH3-H2O and H2O-LiBr solutions. A mathematical model was developed and the thermodynamic properties
of refrigerants (water and ammonia) and solutions (NH3-H2O and H2O-LiBr) were calculated by using related
equations of state. Next, a computer program written in visual programming language Delphi was prepared
to calculate the thermodynamic properties of each point (the inlet and outlet of each component), the heat
capacity of each component, the main system pressures, concentrations, the circulation ratio and the COP of
the VAR system. This computer program is explained in detail using the flow chart, figures and tables and
what can be done with this computer program is also explained and some of the results are presented in
tabular and graphical form and analyzed.

2. SYSTEM ANALYSIS
The schematic illustration of the single-stage VAR cycle is presented in Figure 1. As Figure 1 illustrates, the
fundamental VAR cycle contains four main components; a generator, an absorber, a condenser and an
evaporator. The condenser and the evaporator function in the same manner as they do in the Vapor
Compression Refrigeration System. The refrigerant-absorbent solution passing through the solution pump is
referred to as strong solution, being relatively rich in refrigerant. The solution returning from the generator to
the absorber contains only a little refrigerant and is therefore referred to as weak solution.

Figure 1. The schematic illustration of the single-stage VAR cycle

There are two heat exchangers such as; the solution heat exchanger (heat exchanger 1) and the refrigerant
heat exchanger (heat exchanger 2). The more heat the strong solution recovers from the weak solution
leaving the generator, the less it needs from the external sources in the generator (see the heat exchanger 1
in Figure 1). The saturated liquid from the condenser gets cold and the saturated vapor from the evaporator
gets warm in the heat exchanger 2 and this causes an increase on the cooling capacity and so the COP.
Mass balance at the generator can be written as follows;
for NH3-H2O

m SS = m WS + m NH 3 (total mass balance) (1)

m SS X SS = m WS X WS + m NH 3 (ammonia balance) (2)

544
and for H2O-LiBr

m WS = m SS + m H 2O (total mass balance) (3)

m WS X WS = m SS X SS (LiBr balance) (4)

From Eqs. 1- 4, the strong and weak solution mass flow rates can be determined as follows,
for NH3-H2O

1 − X WS
m SS = m NH 3 (5)
X SS − X WS
1 − X SS
m WS = m NH 3 (6)
X SS − X WS

and for H2O-LiBr

X WS
m SS = m H 2O (7)
X SS − X WS
X SS
m WS = m H 2O (8)
X SS − X WS

The heat capacities of the main components of the VAR system can be calculated by using circulation ratio
(CR),

m SS
CR = ;
m R
q E = h5 − h4 (9)
qC = h1 − h2 (10)

for NH3-H2O

qG = h1 − CRh9 + (CR − 1)h10 (11)

q A = h6 + CRh12 − (CR − 1)h7 (12)

and for H2O-LiBr

qG = h1 + CRh10 − (CR + 1)h9 (13)

q A = h6 + CRh12 − (CR + 1)h7 (14)

Heat exchanger 1
Eqs. 15, 16 and 17 can be determined using the energy balance at the Heat exchanger 1 [1];

T11 = ε HEX ,1T8 + (1-ε HEX,1 )T10 (15)

h9 = h8 + (h10 − h11 )(CR − 1) / CR (for NH3-H2O) (16)
h9 = h8 + (h10 − h11 )CR / (CR + 1) (for H2O-LiBr) (17)

and the heat capacity of the Heat exchanger 1 (solution heat exchanger) can be calculated per unit
refrigerant mass flow rate,

545
 q HEX ,1 = (h10 − h11 )(CR − 1) = (h9 − h8 )CR (for NH3-H2O) (18)
q HEX ,1 = (h10 − h11 )CR = (h9 − h8 )(CR + 1) (for H2O-LiBr) (19)

Heat exchanger 2
The saturated liquid from the condenser and the saturated vapor from the evaporator enter the Heat
exchanger 2 which is situated at the refrigeration side. The saturated vapor takes heat and leaves the heat
exchanger as superheated vapor and the saturated liquid gives heat and leaves as sub-cooled liquid. If the
energy balance is written for Heat exchanger 2;

h6 = h5 + ε HEX , 2 (h6 (T 2 ) − h5 ) (20)

h3 = h2 − ε HEX , 2 (h6 (T 2 ) − h5 ) (21)

where, h6(T2), is the enthalpy of superheated vapor at the condenser temperature. The heat capacity of the
Heat exchanger 2 can be calculated per unit refrigerant mass flow rate,

q HEX , 2 = (h2 − h3 ) = (h6 − h5 ) (for both NH3-H2O and H2O-LiBr) (22)

When the changes in potential and kinetic energies of a liquid are negligible, the isentropic efficiency of a
pump is defined as;

ws h8,s − h7
ηP = = (23)
wa h8,a − h7

where h8,a and h8,s are the enthalpy values at the exit state for actual and isentropic process, respectively.
The enthalpy at the solution pump outlet and pump power consumption per unit refrigerant mass flow rate
can be determined as follows;

h8,a = h7 + ( pC − p E )v / η P (24)
wP = CR ( pC − p E )v / η P (for NH3-H2O) (25)
wP = (CR + 1)( pC − p E )v / η P (for H2O-LiBr) (26)

The measure of performance of refrigerators is expressed in terms of coefficient of performance (COP),

defined as,

qE
COP = (27)
qG + wP

3. COMPUTER PROGRAM
The flow chart of the computer program is presented in Figure 2. The computer program was prepared in
three languages; Turkish, English and German. The user first selects the language and then the solution.
The main system parameters can also be entered in this introduction page which is shown in Figure 3.

After the system working temperatures, the heat exchanger effectiveness and the solution pump efficiency
are entered, if the user hits the “calculate” button, the results page appears. The results page (see Figure 4)
has two sections which are, “summary” and “detail” sections. The “summary” section includes the heat
capacities of main system components, the solution pump power consumption, the COP, the circulation
ratio, the system working pressures and solution concentrations. The “detail” section (see Figure 5) includes
the thermodynamic properties of each point (T, P, X and h).

546
At this stage the user can go back to inputs section (introduction page) by hitting the “inputs” button or
forward to outputs section by hitting the “outputs” button. The “outputs” page is presented in Figure 6. As
can be seen from Figure 6, the “outputs” page shows all parameters related to the simulation including input
and output parameters. The reason for this is to present the effect of one certain parameter by listing all
results. The user can verify one certain parameter (for example the generator temperature) and list all the
results to observe the effect of this parameter on the system performance.

Solution selection
{NH3 – H2O or H2O-LiBr}

Input of main parameters

{TG, TC, TA, TE, εHEX,1, εHEX,2, ηP}

The thermodynamic properties calculation of

refrigerant and solution

The calculation of main system parameters for

performance analysis
{qG, qC, qA, qE, qHEX,1, qHEX,2,wP,
CR, COP etc.}

The output of main system parameters for

performance analysis

The output of thermodynamic properties of

each point

Stop

Figure 2. The flow chart of the computer program

547
The “save” button of the outputs page can be used to save the results as a text file (.txt) which can be
analyzed by some other computer software such as Excel, Lotus etc. The “delete” button of the outputs page
can be used to delete the selected line.

Figure 3. The introduction page of the computer program

Figure 4. The results/summary page

548
Figure 5. The results/detail page

Figure 6. The outputs page

4. DISCUSSIONS AND RESULTS

Table 1 and Table 2 present the computer simulation results for NH3-H2O and H2O-LiBr solutions
respectively. In this simulation, it is assumed that, TG = 90 °C, TC = TA=40 °C, TE = 5 °C, εHEX,1 = 0.75 and ηP
= 0.85. As can be seen from Table 1 and Table 2, the system working pressures are higher at the VAR

549
system using NH3-H2O solution, since ammonia is the refrigerant. This is due to fact that, the saturation
pressure against the temperature is much bigger in ammonia compare to water.
Table 3 presents the heat capacities of the main components. As can be seen from Table 3, compare
to other components the generator heat capacity is the biggest and the solution pump power consumption is
negligible. The solution pump power consumption is bigger in the system using NH3-H2O solution, because
of the bigger pressure difference. Table 3 also indicates that the cooling capacity and COP of the system
using H2O-LiBr solution is bigger compare to the VAR system using NH3-H2O solution for the same working
conditions.

Table 1. The computer simulation results for NH3-H2O

Points T (°C) P (kPa) X (%) h (kJ/kg)
1 90.00 1553.82 100.00 1631.44
2 40.00 1553.82 100.00 367.10
3 40.00 1553.82 100.00 367.10
4 5.00 515.72 100.00 367.10
5 5.00 515.72 100.00 1448.60
6 5.00 515.72 100.00 1448.60
7 40.00 515.72 50.39 -75.16
8 40.33 1553.82 50.39 -73.65
9 74.03 1553.82 50.39 86.40
10 90.00 1553.82 45.00 160.61
11 53.17 1553.82 45.00 -16.83
12 53.17 515.72 45.00 -16.83

Table 2. The computer simulation results for H2O-LiBr

Points T (°C) P (kPa) X (%) h (kJ/kg)
1 90.00 7.38 0.00 2662.34
2 40.00 7.38 0.00 167.60
3 40.00 7.38 0.00 167.60
4 5.00 0.87 0.00 167.60
5 5.00 0.87 0.00 2508.50
6 5.00 0.87 0.00 2508.50
7 40.00 0.87 57.84 105.04
8 40.00 7.38 57.84 105.05
9 72.80 7.38 57.84 170.40
10 90.00 7.38 62.10 221.22
11 52.51 7.38 62.10 151.07
12 52.51 0.87 62.10 151.07

Table 3. The heat capacities of main components

Heat capacities (kJ/kg) NH3-H2O H2O-LiBr
Generator (qG) 2227.89 3182.86
Condenser (qC) 1264.34 2494.75
Evaporator (qE) 1081.50 2340.91
Absorber (qA) 2060.50 3029.07
Pump (wP) 15.45 0.07
Heat Exchanger-1 (qHEX,1) 1632.74 953.70
COP 0.48 0.74

The COP and CR against the generator temperature of the VAR systems using NH3-H2O and H2O-LiBr
solutions are presented in Figures 7 and 8, respectively. The main parameters were assumed to be
TC=TA=40°C, TE=8°C, εHEX,1=0.50, ηP=0.85. In both systems, if the generator temperature increases, the
COP increases as well, on the other hand the CR decreases. The COP increases up to TG = 95 °C and after
this point keeps its value around 0.49 and 0.74 for NH3-H2O and H2O-LiBr solutions respectively. Below TG =
85 °C, the COP decreases considerably, so below TG = 85 °C, the VAR system is not efficient. The system

550
using H2O-LiBr solution presents always better performance, when the generator temperature increases, the
difference is significant. However, there is a crystallization problem in the system using H2O-LiBr solution.

Figure 7. The COP against the generator temperature (TC=TA=40°C, TE=8°C, εHEX,1=0.50, ηP=0.85)

Figure 8. The CR against the generator temperature (TC=TA=40°C, TE=8°C, εHEX,1=0.50, ηP=0.85)

Figure 9 shows the COP against the effectiveness of the heat exchanger 1. Figure 10 presents the outlet
temperatures against the effectiveness of the heat exchanger 1 of the VAR system using NH3-H2O solution.
The main parameters were assumed to be TG=90°C, TC=TA=40°C, TE=8°C, εHEX,2=0, ηP=0.85. If the
effectiveness of the heat exchanger 1 equals to 0 (εHEX,1 = 0), the COP values are 0.35 and 0.63 for the VAR
systems using NH3-H2O and H2O-LiBr solutions respectively. When the effectiveness increases, the heat
input to the generator decreases and this causes an increase on the COP. If the effectiveness of the heat
exchanger 1 equals to 1 (εHEX,1 = 1), the COP of the VAR systems using NH3-H2O goes up to 0.65 which
means 84% increase.

551
Figure 9. The COP against the effectiveness of the heat exchanger 1(TG=90°C, TC=TA=40°C, TE=8°C, ηP=0.85)

Figure 10. The outlet temperatures against the effectiveness of the heat exchanger 1 of the VAR system using NH3-H2O solution
(TG=90°C, TC=TA=40°C, TE=8°C, ηP=0.85)

5. CONCLUSIONS
This study focuses on the thermodynamic analysis of the single-stage VAR systems using NH3-H2O and
H2O-LiBr solutions in detail. After having set up the required formulation, there is a possibility to make a
mistake in theoretical solution using a calculator for each condition. For that reason, a mathematical model
was developed and a user friendly, practical, easy and attractive computer program written in Delphi was
prepared. Using this computer program, anyone can analyze the VAR systems using NH3-H2O and H2O-LiBr
solutions and the effect of the main system temperatures such as the generator, the evaporator, the
absorber and the condenser temperatures, the solution pump efficiency and the effectiveness of the heat
exchangers on system performance and working conditions. This computer program can be used in
Refrigeration Techniques courses and refrigeration applications and also can be developed to cover the two-
stage VAR systems.

NOMENCLATURE
COP coefficient of performance
CR circulation ratio
h enthalpy (kJ/kg)
m mass flow rate (kg/s)
p pressure (kPa)
q heat transfer (kJ/kg)
T temperature (°C)
VAR vapor absorption refrigeration

552
VCR vapor compression refrigeration
w pump work (kJ/kg)
X concentration
ε effectiveness
η efficiency
3
v specific volume (m /kg)

Subscripts
a actual process
A absorber
C condenser
E evaporator
G generator
HEX heat exchanger
H2O water
NH3 ammonia
P pump
WS weak solution
R refrigerant
SS strong solution
s isentropic process
1..12 points

REFERENCES
Horuz I. 1998. A comparison between ammonia-water and water-lithium bromide solutions in vapor
absorption refrigeration systems. Int. Comm. Heat and Mass Transfer 25(5): 711-721.

Karamangil M.I., Coskun S., Kaynakli O., Yamankaradeniz N. 2010. A simulation study of performance
evaluation of single-stage absorption refrigeration system using conventional working fluids and alternatives.
Renewable and Sustainable Energy Reviews 14 (7): 1969-1978.

Kaynakli O., Kilic M. 2007. Theoretical study on the effect of operating conditions on performance of
absorption refrigeration system. Energy Conversion and Management 48(2): 599-607.

Keçeciler A., Acar I.H., Doğan A. 2000. Thermodynamic analysis of the absorption refrigeration system with
geothermal energy: an experimental study. Energy Conversion and Management 41(1): 37-48.

Mostafavi M., Agnew B. 1996. The impact of ambient temparature on lithium bromide-water absorption
machine performance. Applied Thermal Engineering 16(6): 515-522.

Saravanan R., Maiya M.P. 1998. Thermodynamic comparison of water based working fluid combinations for
a vapour absorption refrigeration system. Applied Thermal Engineering 18(7): 553-568.

Sun D.W. 1998. Comparison of the performance of NH3-H2O, NH3-LiNO3 and NH3-NaSCN absorption
refrigeration systems. Energy Conversion 39 (5/6): 357-368.

Xu F., Goswami D.Y. 1999. Thermodynamic properties of ammonia-water mixtures for power-cycle
applications. Energy 24(6): 525-536.

553
 PERFORMANCE EVALUATION OF A FLOWING-ELECTROLYTE DIRECT METHANOL
FUEL CELL THROUGH NUMERICAL MODELING
C. O. Colpan
Mechanical Engineering Department, DokuzEylul University
Buca, Izmir, Turkey
ozgur.colpan@deu.edu.tr

ABSTRACT
A model of a flowing-electrolyte direct methanol fuel cell has been developed to predict the performance
of the cell under various operating conditions. Governing equations includingthe proton and electron
transport, continuity, momentum, species transport, and the auxiliary equations are coupled to determine
the output parameters. The results show thathigher fluid velocities at the channel inlets are needed to
obtain higher power densities. It is also found that the electrical efficiency of the FE-DMFC can be further
increased if the methanol leaving the flowing electrolyte channel is recirculated into the methanol storage
tank.

Keywords: Methanol, direct methanol fuel cell, DMFC, sulfuric acid, flowing electrolyte, modeling

1. INTRODUCTION

Direct methanol fuel cell (DMFC) is one of the most promising technologies that can be used in portable
power applications. Some of the advantages of the DMFC are related to using methanol as fuel: easy to
store, low cost, and high energy density. However, the cost of some of the components such as catalysts
is still a major issue for its successful commercialization. In addition to its expensive components, main
technical challenges with DMFC include the low power density and electrical efficiency due to the
methanol crossover problem and the low rate of methanol oxidation kinetics on the anode [1, 2]. The
Flowing-Electrolyte Direct Methanol Fuel Cell (FE-DMFC) is a novel DMFC design which provides
performance improvement by eliminating the methanol crossover from the anode to the cathode[3]. FE-
DMFC design basically reduces the cathodicoverpotential as it hinders the amount of methanol reaching
the cathode catalyst layer. Schematic of a FE-DMFC is shown in Figure 1.

Fig. 1.Schematic of a 2D cross-section of the FE-DMFC [4]

The operation principle of FE-DMFC can be explained as follows. A liquid mixture consisting of methanol
and water is pumped to the fuel cell through the fuel channel. Some amount of these species diffuses

554
through the anode backing layer, which is typically carbon cloth; and then they reach the anode catalyst
layer having Pt and Ru, where the protons and electrons are generated. The reaction occurring at the
ACL is shown in Eq. (1) [4].

𝐶𝐻3 𝑂𝐻 + 𝐻2 𝑂 → 𝐶𝑂2 + 6𝐻 + + 6𝑒 − (1)

The protons generated in the ACL reach the cathode catalyst layer having Pt after they transport through
®
Nafion anode and cathode membranes and flowing electrolyte channel consisting of diluted sulfuric acid
solution. The electrons are conducted to the external load through the ABL and the anode interconnect.
These electrons reach the CCL after they pass through the cathode interconnect and the cathode backing
layer. As the electrons and protons reach the CCL, they react with oxygen pumped to the cell through the
air channel (AC) and transported through the CBL. The electrochemical reaction occurring at the CCL is
shown in Eq. (2) [4].

1.5𝑂2 + 6𝐻 + + 6𝑒 − → 3𝐻2 𝑂(2)

The methanol crossing over the membranes and FEC, if any, also reaches the CCL. Here, the methanol
reacts with oxygen as shown in Eq. (3) [4].

𝐶𝐻3 𝑂𝐻 + 1.5𝑂2 → 𝐶𝑂2 + 2 𝐻2 𝑂(3)

There have been many papers published on DMFC modeling. Some of the recent papers have included
the multi-dimensional effects to predict the distributions of the output parameters (e.g. [5-8]), the two-
phase transport taking into account the effects of the gaseous species (e.g. [7-9]), and the water transport
phenomena (e.g. [10-12]). However, there have been only a few papers published in FE-DMFC modeling.
Kjeang et al. [13, 14] studied the methanol crossover reduction in the flowing electrolyte channel with
different operating parameters. Colpan et al. [15] developed a one-dimensional FE-DMFC model to study
the effects of flowing electrolyte channel thickness, concentration of methanol at the inlet, volumetric flow
rate of the flowing electrolyte, and the recirculation of the methanol at the flowing electrolyte channel
outlet. Their study showed that the maximum power densities of the DMFC and FE-DMFC are almost
same; but the electrical efficiency of the recirculated FE-DMFC could be 57% more than that of the
DMFC. Ouellette et al. [16] developed an one-dimensional two-phase model to predict the effects of CO2
formation on the performance of the FE-DMFC. Their study showed that the presence of CO2 results in a
reduction in the methanol crossover as well as the active area of the fuel cell.

2. MODELING

The main assumptions used in the modeling are as follows [4].

• The changes of the protonic and electronic conductivities within a layer are neglected.
• The formations of the CO2 bubbles and water vapour are neglected; i.e. two phase effects are not
taken into account.
• Membranes are fully hydrated.
• Methanol is fully consumed at the interface of the cathode membrane and the cathode catalyst layer.
• The effects due to the channel curvature are not taken into account.
• The flow in the electrolyte channel is considered as fully developed laminar flow.
• The fuel cell is isothermal.
• The fuel cell operates at the steady state condition.

555
Table 2.Governing Equations used to Model the FE-DMFC [17]

Conservation Equation Expression

Mass ∂ (ρu )
= S gen
∂x
Momentum ∂P
= Su
∂x
Species ∂  ∂Ck 
 − Deff , k + uCk  = S gen , k
∂x  ∂x 
Charge ∂  ∂ φ 
−  σ p p  = S c , p
∂x  ∂x 

The average current density of the cell can be calculated using Eq. (4).

1
𝑖𝑐𝑒𝑙𝑙 = � 𝑗𝑎 𝑑𝑥𝑑𝑦 (4)
𝑙𝑐𝑒𝑙𝑙
𝐴𝐶𝐿

The power density of the cell, and the electrical efficiency of the cell can be found using Eqs. (5) and (6),
respectively.

𝑊̇𝑐𝑒𝑙𝑙
"
= 𝑖𝑐𝑒𝑙𝑙 𝑉𝑐𝑒𝑙𝑙 (5)
"
𝑊̇𝑐𝑒𝑙𝑙
𝜂𝑒𝑙 = 𝑓𝑐,𝑖𝑛 (6)
𝑁̇" 𝑀𝑒𝑂𝐻 𝐿𝐻𝑉
������

If the flowing electrolyte is recirculated, the electrical efficiency can be shown as follows.
"
𝑊̇𝑐𝑒𝑙𝑙
𝜂𝑒𝑙 = 𝑓𝑐,𝑖𝑛 𝑓𝑒𝑐,𝑜𝑢𝑡 (7)
�𝑁̇" 𝑀𝑒𝑂𝐻 −𝑁̇ " 𝑀𝑒𝑂𝐻 �𝐿𝐻𝑉
������

3. RESULTS AND DISCUSSION

The results of the DMFC model developed are compared to the experimental data published by Ge and
Liu [18]. The comparison of the experimental data and the results of the simulation of the model is shown
in Figure 2. As can be seen from this figure, there is a good agreement between the experimental and
modeling data. The minor discrepancy at the low current density conditions can be attributed to the
simplified equations used for electrochemical relations, occurring at the anode and cathode catalyst
layers.

The effect of the fluid velocity at the inlet of the fuel channel on the cell voltage and power density is
shown in Figure 3. This figure shows that the trend of the polarization curves does not change
significantly for different stoichiometric coefficients. However, the limiting current density increases as this
coefficient increase; thus lower fuel channel inlet velocities cause a more limited operating range for the
fuel cell. This figure also shows that higher maximum power densities could be achieved with higher
-2 -
stoichiometric coefficients. For example, the maximum power densities are found as 600 W m , 670 W m
2 -2 -2
, 720 W m , and 730 W m when the stoichiometric flow coefficients at the fuel channel inlet are taken as
1, 2, 3 and 4, respectively.

556
Fig. 2.Validation of the DMFC model (Modified from [4])

Fig. 3.Cell voltage and power density of the FE-DMFC for different fuel channel inlet velocities (Modified
from [4])

557
The fluid velocity at the fuel channel inlet has a significant effect on the electrical efficiency of the cell
significantly, as can be seen in Figure 4. The ratio of the total amount of the methanol utilized to the
methanol entering the fuel cell increases as the fluid velocity decreases; which provides an increase in
the electrical efficiency of the cell. Thus, lower stoichiometric coefficients at the fuel channel inlet are
desirable to obtain higher electrical efficiencies. But, it should be kept in mind that lower maximum power
densities are also achieved at these coefficients. Hence, since power density and electrical efficiency
cannot be maximized at the same time, the choice of the fluid velocity at the fuel channel inlet mainly
depends on the purpose of the device and economical factors such as the fuel cost and the purchase and
operation costs of the fuel cell.

The electrical efficiency of the FE-DMFC can be further increased if the methanol leaving the FEC can be
recovered and recirculated back to the methanol storage tank. As less amount of methanol enters the fuel
cell at a given time due to the addition of this recirculation loop, the electrical efficiency of a recirculated
FE-DMFC at a given stoichiometric coefficient becomes higher than that of a FE-DMFC, which can be
seen in Figure 4. The maximum electrical efficiency for the recirculated FE-DMFC is found to be 56%
-2
when the stoichiometric coefficient is taken as 1 and the current density is equal to 1700 A m .

Fig. 4.Electrical efficiency of the FE-DMFC and recirculated FE-DMFC for different fuel channel inlet
velocities (Modified from [4])

4. CONCLUSIONS

A multi-physics model was developed to simulate the performance of the cell and to study the effects of
some of the key operating parameters. The results showed that there is a good agreement between the
results of the model and the experimental data.As the stoichiometric flow coefficient at the fuel channel
inlet increases, higher power densities could be achieved; however less methanol is utilized and the
electrical efficiency of the cell decreases.The electrical efficiency of the FE-DMFC could be further
increased if the methanol leaving the FEC is recirculated to the methanol storage tank.

558
REFERENCES

1. Colpan, C.O., Dincer, I., Hamdullahpur, F. 2008. Portable fuel cells – Fundamentals, technologies
and applications. pp. 87-101. Mini-Micro Fuel Cells: Fundamentals and Applications. Eds. Kakac, S.,
Pramuanjaroenkij, A., Vasiliev, L. NATO Science for Peace and Security Series. Springer.
Netherlands.
2. Wang, C.Y. 2008. Principles of direct methanol fuel cells for portable and micro power. pp. 235-242.
Mini-Micro Fuel Cells: Fundamentals and Applications. Eds. Kakac, S., Pramuanjaroenkij, A.,
Vasiliev, L. NATO Science for Peace and Security Series. Springer. Netherlands.
3. Kordesch K., Hacker, V., Bachhiesl, U. 2001. Direct methanol-air fuel cells with membranes plus
circulating electrolyte. Journal of Power Sources. 96. pp. 200-203.
4. Colpan,C.O.,, Fung, A., Hamdullahpur, F. 2012. 2D modeling of a flowing-electrolyte direct methanol
fuel cell. Journal of Power Sources. 209. pp. 301-311.
5. Ge, J., Liu, H. 2006. A three-dimensional mathematical model for liquid-fed direct methanol fuel cells.
Journal of Power Sources. 160. pp. 413-421.
6. Liu, W., Wang, C-Y. 2007. Three-dimensional simulations of liquid feed direct methanol fuel cells.
Journal of the Electrochemical Society. 154(3). pp. B352-B361.
7. Yang, W.W., Zhao, T.S. 2007. A two-dimensional, two-phase mass transport model for liquid-feed
DMFCs. ElectrochimicaActa. 52. pp. 6125–6140.
8. Ge, J., Liu, H. 2007. A three-dimensional two-phase flow model for a liquid-fed direct methanol fuel
cell. Journal of Power Sources. 163. pp. 907–915.
9. Murgia, G., Pisani, L., Shukla, A.K., Scott. K. 2003. A numerical model of a liquid-feed solid polymer
electrolyte DMFC and its experimental validation. Journal of the Electrochemical Society. 150(9). pp.
A1231-A1245.
10. Liu, W., Wang, C-Y. 2007. Modeling water transport in liquid feed direct methanol fuel cells. Journal
of Power Sources. 164. pp. 189–195.
11. Liu, F., Lu, G., Wang, C-Y. 2006. Low crossover of methanol and water through thin membranes in
direct methanol fuel cells. Journal of the Electrochemical Society. 153(3). pp. A543-A553.
12. Lu, G.Q., Liu, F.Q., Wang, C-Y. 2005. Electrochemical and Solid-State Letters. 8(1). pp. A1-A4.
13. Kjeang, E., Goldak, J., Golriz, M.R., Gu, J., James, D., Kordesch, K. 2005. Modeling methanol
crossover by diffusion and electro-osmosis in a flowing electrolyte direct methanol fuel cell. Fuel
Cells. 4. pp. 486-498.
14. Kjeang, E., Goldak, J., Golriz, M.R., Gu, J., James, D., Kordesch, K. 2006. A parametric study of
methanol crossover in a flowing electrolyte-direct methanol fuel cell. Journal of Power Sources. 153.
pp. 89-99.
15. Colpan, C.O., Cruickshank, C.A., Matida, E., Hamdullahpur, F. 2011. 1D modeling of a flowing
electrolyte-direct methanol fuel cell. Journal of Power Sources. 196. pp. 3572–3582.
16. Ouellette, D., Colpan, C.O., Matida, E., Cruickshank, C.A. 2011. Modeling of carbon dioxide formation
inflowing electrolyte-direct methanol fuel cells. Proceedings of the ASME 2011 9th Fuel Cell Science,
Engineering and Technology Conference (FuelCell2011). August 7-10, 2011. Washington, DC, USA.
17. Ouellette, D., Colpan, C.O., Matida, E., Cruickshank, C.A., Fung, A., Hamdullahpur, F. 2012.
Multiphysics Modeling and Testing of a Flowing Electrolyte-Direct Methanol Fuel Cell. Proceedings of
the International Conference on Hydrogen Production-2012 Seoul (Republic of Korea). June 24-27,
2012.
18. Ge, J., Liu, H. 2005. Experimental studies of a direct methanol fuel cell. Journal of Power Sources.
142. pp. 56-69.

559
 INVESTIGATION OF HUMIDITY EFFECTS ON THE THERMAL COMFORT AND HEAT
BALANCE OF THE BODY
1,* 1 2 1
O. Kaynakli , M. Mutlu , I. Atmaca and M. Kilic
1
Uludag University, Faculty of Engineering-Architecture, Mechanical Engineering Depart., Bursa, TURKEY
2
Akdeniz University, Faculty of Engineering, Mechanical Engineering Depart., Antalya, TURKEY
email : kaynakli@uludag.edu.tr, mustafamutlu@uludag.edu.tr, atmaca@akdeniz.edu.tr, mkilic@uludag.edu.tr

ABSTRACT
Humidity, one of the most confusing of all climatic parameters in assessing the indoor climate, affects
comfort in a number of ways both directly and indirectly, and the avenues by which humidity affects comfort
are not completely known. The aim of this study is to comprehend the effects of humidity on thermal
interaction between the human body and its environment, and thermal sensation. In this paper, the effect of
humidity on heat and water balance of human body, and in turn on body temperatures and thermal sensation
are investigated. A mathematical model of heat and mass interaction between the human body and the
surrounding environment has been established, and the effect of air humidity has been examined under
various relative humidity levels by means of using the empirical relations that express thermoregulatory
control mechanisms. In the numerical model, human body has been separated to 16 sedentary segments
and possible local discomforts have been taken into consideration. Using the model, changes in the sensible
and latent heat losses, body temperatures, skin wettedness and thermal comfort indices have been
calculated and results have been discussed explicitly.

1. INTRODUCTION
The living body constantly produces heat and this must be transferred to the environment. Heat balance is
the balance between the rate of heat production and the rate of heat loss. Body temperature is regulated by
the blood flowing from the core to the skin where the heat is dissipated by radiation, convection and
evaporation to the environment. Thermal comfort is generally associated with a neutral or near neutral whole
body thermal sensation. Thermal sensation depends on body temperature, which in turn depends on thermal
balance and the effects of environmental factors, as well as personal factors (Fanger 1972, Gagge et al.
1971). Air temperature and humidity are two important factors affecting thermal comfort. Other factors
include: radiant temperature, air velocity, occupant activity levels, and clothing.

Humidity is the most confusing of all climatic parameters in assessing the indoor climate. The humidity of the
ambient air has a wide range of effects on the energy and water balance of the body as well as on elasticity
of the materials, air quality perception, build-up of electrostatic charge and the formation of mould. The
humidity of ambient air has an influence on three mechanisms of water loss from the human body. These
are: the diffusion of water vapor through the skin, the evaporation of sweat from the skin surface, and the
humidification of the respired air. In all cases, the water is lost in gaseous form, which also results in an
energy loss from the body from which the relevant evaporation heat is taken (Höppe and Martinac 1998,
Teodosiu et al. 2003).

Relative humidity is the ratio of the water vapor content of air to the maximum possible water vapor content
of air at the same temperature and air pressure. For people who perform very light or sedentary activities,
the influence of humidity on personal comfort is not great. Higher humidity will make a person feel warmer
particularly in a ventilated area where the air speed is low. Elevated relative humidity reduces the body’s
ability to lose heat through perspiration and evaporation. When this happens, individuals may perceive
temperatures to be higher than they actually are. On the other hand, low relative humidity can cause
discomfort due to drying of the eyes, nose, throat, mucous membranes and skin (Berglund 1998, Toftum et
al. 1999).

Humidity affects human comfort in heat balance and in other many ways, such as healthy, thermal
sensation, skin wettedness, tactile sensation of fabrics and perception of air quality. Liviana et al. (1988),
which studied humidity effects on human healthy, found that low humidity can contribute to eye irritation. Eye
discomfort increased with time in low humidity environments (Tdp < 2°C). Teodosiu et al. (2003) described a
numerical model to assess the thermal comfort taking into account the indoor air moisture and its transport
by the air flow within an enclosure.

560
Thermal neutrality is a necessary but not sufficient condition for thermal comfort (Olesen et al. 1988). At the
thermal neutrality, humidity may be a cause of discomfort for two reasons: an uncomfortably high level of
skin humidity or insufficient cooling of the mucous membranes in upper respiratory tract by inhalation of
humid or warm air. Toftum et al. (1998a) dealt with discomfort caused by skin humidity. In the companion
paper of that, Toftum et al. (1998b), they have presented the results of the experiments on humidity and
temperature of inhaled air on discomfort. de Dear et al. (1989) studied the impact of humidity on thermal
comfort during step changes between 20% and 80% relative humidity. They found that thermodynamics of
moisture absorption and desorption in clothing fibers during humidity transients affected the heat balance
and thermal sensation of subjects. Berglund and Cain (1989) showed that air freshness and acceptability
perceptions decreased with increasing temperature and humidity. In particular, relative humidities above
65% were associated with unacceptable air quality judgments. Fang et al. (1996) reached to similar results
from air contaminated with emissions for common building materials. The air quality was perceived to be
more acceptable with decreasing temperature and humidity. Fountain et al. (1999) presented a short
summary of previous studies on the influence of humidity on thermal comfort. Guan et al. (2003) summarized
the current advances in thermal comfort modeling for both building and vehicle HVAC applications that have
occurred in the literature. Fiala et al. (1999) modeled the physiological mechanisms within the body, thermal
exchanges between the body and its environment, and gave the detailed numerical model in their study.
Each body segment was considered as five body layers (core, muscle, fat, inner skin and outer skin). In their
model, local variations of surfaces convection, directional radiation exchange, evaporation and moisture
collection at the skin, and the nonuniformity of the clothing ensembles were considered. Further study is
warranted to better understand and quantify how humidity affects building occupants and HVAC system
users.

This paper inclines to relative humidity, one of the most complex parameter of the independent thermal
comfort variables. As mentioned above, relative humidity affects comfort in a number of ways both directly
and indirectly. In this study, investigation of the effects of low, intermediate and high relative humidity values
on the energy balance and the thermal comfort indices is aimed. Human body is divided into 16 sedentary
segments and simulation of the heat-mass transfer between the body segments and their environment is
performed. By using the present model, variations of the core, skin and mean body temperatures, sensible
and latent heat losses from the body, skin wettedness with the relative humidity are obtained for both whole
body and body segments at different ambient temperatures. The effects of the same conditions on the
thermal comfort indices, TSENS and DISC, are presented graphical forms.

2. MATHEMATICAL MODEL
The model used in this study combines the Gagge et al. (1971) and Olesen et al. (1988) models and based
on the same approach used in the study by McCullough et al. (1989). The fundamental equations of the
model are presented in this section. Further details of the developed model can be found in references
(Kaynakli and Kilic 2005, Kaynakli et al. 2005).

A two-compartment transient energy balance model developed by Gagge et al. (1971) represents the body
as two concentric cylinders the inner cylinder represents the body core (skeleton, muscle, internal organs)
and the other cylinder represents the skin layer. This model, by considering instantaneous heat storage of
the core and the skin compartment, assumes that temperatures of these compartments change with time.
Thermal interaction of the human body with the environment for the core and the skin compartments can be
written as follows (Gagge et al. 1971, ASHRAE 1989):

S cr = M − W − Qres − Qcr , sk (1)

S sk = Qcr ,sk − (Qcd + Qcv + Qrd + Qe ,sk ) (2)

where M is the rate of metabolic heat production, W is the rate of external work done by the muscles, Qres
is the total rate of heat loss through respiration, Qcr,sk is the rate of heat exchange between the core and skin,
Qe,sk is the total rate of evaporative heat loss from skin. The total rate of sensible heat loss from skin can be
divided to three parts namely conductive (Qcd), convective (Qcv) and radiative (Qrd) heat loss rates. Since the
conductive heat loss is a small fraction of the total heat loss from the body, generally it is neglected in the
comfort calculations (Kaynakli et al. 2003, Burch et al. 1991). Heat storage rate of the core and skin causes
instantaneous temperature changes in each compartment. This effect may be expressed with following
equations:

561
 dt cr / dθ = S cr Ab / ((1 − α )mc p ,b ) (3)
dt sk / dθ = S sk Ab / (αmc p ,b ) (4)

where α is the fraction of body mass concentrated in skin compartment, m is the mass of body, cp,b is the
specific heat of body, tcr and tsk are the core and skin temperature of the body, respectively.

Heat is generated in the body by metabolism and can be lost to the environment by conduction, convection,
radiation, evaporation of moisture from the skin and respiration. Convective and radiative heat loss from
clothed body to environment is calculated as:

(t sk − t o )
Qcv + Qrd = (5)
1
Rcl +
(hcv + hrd ) f cl
where Rcl is thermal resistance of clothing, fcl is the ratio of clothed to nude body area, hcv and hrd are the
convective and radiative heat transfer coefficients, respectively. Operative temperature (to) can be defined as
the average of the mean radiant and ambient air temperatures, weighted by their respective heat transfer
coefficients. The radiative heat transfer coefficient (hrd) was assumed to be 4.7 W/m °C for typical clothing
2

systems and necessary equation for calculating the convective heat transfer coefficient for seated person
with moving air was taken from ASHRAE (1989).

In naturally ventilated rooms with closed windows air velocity rarely exceeds 0.1 m/s (Höppe and Martinac
1998). For that reason, in the present study average air velocity over the body is assumed to be 0.1 m/s. The
maximum total latent heat loss from the skin (Qe,max) is given by

(p − pa )
Qe ,max =
sk , s
(6)
Rcl 1
+
icl LR hcv f cl LR

where, psk,s is the saturated water vapor partial pressure at the skin temperature and pa is the water vapor
partial pressure in the ambient air, icl is permeation efficiency of the clothing and LR is the Lewis Ratio which
is the ratio of the evaporative heat transfer coefficient to the convective heat transfer coefficient. McCullough
et al. (1989) have been found an average value of icl =0.34 for common indoor clothing and LR equals
approximately 16.5 °C/kPa at typical indoor conditions (ASHRAE 1989).

Total skin wettedness (w) including wettedness due to regulatory sweating (wrsw) and to diffusion through the
skin (wdif) can be written as:

w = wrsw + wdif (7)

where wrsw and wdif are given by

m rsw h fg
wrsw = (8)
Qe,max
wdif = 0.06(1 − wrsw ) (9)

where m  rsw is the rate of sweat production, hfg is the heat of vaporization of water. m rsw is given by control
signal equation. Evaporative heat loss from the skin (Qe,sk) is a combination of the evaporation of sweat
secreted due to thermoregulatory control mechanisms (Qe,rsw) and the natural diffusion of water through the
skin (Qe,dif):

562
 Qe,rsw = wrswQe,max (10)
Qe,dif = wdif Qe,max (11)
Qe,sk = (Qe,rsw + Qe,dif ) = w Qe,max (12)

During respiration, the body losses both sensible and latent heat by convection and evaporation of heat on
water vapor from the respiratory tract to the inhaled air. Sensible (Qs,res) and latent (Qe,res) heat losses due to
respiration are:

[ ]
Qs ,res + Qe ,res = m res c p ,a (t ex − t a ) + h fg (Wex − Wa ) / Ab (13)

where m res is the mass flow rate of air inhaled, tex and ta are the exhaled air and the ambient air
temperatures, respectively. Wex and Wa are the exhaled air and the ambient air humidity ratio, respectively.
6
The heat of vaporization (hfg) is 2.43 x 10 J/kg. Humidity ratio of the exhaled and ambient air can be
expressed:

Wex = 0.0277 + 0.000065Ta + 0.2Wa (14)

Ra pa pa
Wa = = 0.622 (15)
Rv ( p atm − p a ) ( p atm − p a )
where, Ra and Rv gas constants of air and vapor, respectively, and pa is the water vapor partial pressure in
the ambient air.

The body is divided into 16 segments which are uniformly clothed. In the model, dry and evaporative
resistances are calculated for each body segments which are treated as concentric cylinders. The total
thermal resistance (Rt) and the total evaporative resistance (Re,t) for each segments are as follows
(McCullough et al. 1989):

r (i,0) nl  r (i,0) r (i,0) 

Rt (i ) = Ra (i ) + ∑  Ral (i, j ) + R f (i, j )
r (i, j ) 
(16)
r (i, nl ) j =1  r (i, j − 1)
r (i,0) nl  r (i,0) r (i,0) 
Re,t (i ) = Re,a (i ) + ∑  Re,al (i, j ) + Re, f (i, j )
r (i, j ) 
(17)
r (i, nl ) j =1  r (i, j − 1)

where Ra and Re,a are the thermal and evaporative resistances of the outer air layer, Ral and Re,al are the
thermal and evaporative resistances of the air layer between the clothing layers, respectively. Detailed
information about these resistances may be found in McCullough et al. (1989), and convective-radiative heat
transfer coefficients for 16 individual body segments is reported by de Dear et al. (1997).

2.1 Control Signal Equations

The parameters used in the control signals are the core, skin and mean body temperatures. The blood flow
between the core and skin per unit of skin area can be expressed mathematically as (ASHRAE 1989,
Doherty and Arens 1988):

m bl = [(6.3 + 200WSIGcr / (1 + 0.5CSIGsk ))] / 3600 (18)

The heat exchange between the core and skin can be written as:

Qcr , sk = (K + c p ,bl m bl )(t cr − t sk ) (19)

563
where K is the average thermal conductance, cp,bl is the specific heat of blood. The rate of sweat production
per unit of skin area is estimated by

m rsw = 4.7 x10 −5 WSIGb exp(WSIGsk / 10.7 ) (20)

The average temperature of human body can be predicted by the weighted average of the skin and core
temperatures:

t b = αt sk + (1 − α )t cr (21)

2.2 Prediction of thermal comfort

Because comfort is a “condition of mind”, empirical equations must be used to relate comfort perceptions to
specific physiological responses. The compartment model uses empirical expressions to predict thermal
sensations (TSENS) and thermal discomfort (DISC). TSENS and DISC values can be calculated by following
equations (ASHRAE 1989):

0.4685(t b − t b ,c ) tb < tb,c (22)

TSENS = 4.7η e (t b − t b ,c ) /(t b ,h − t b ,c ) tb,c ≤ tb ≤ tb,h (23)
4.7η e + 0.685(t b − t b ,h ) tb,h < tb (24)
(25)
0.4685(t b − t b ,c ) tb < tb,c
DISC = 4.7(Qe ,rsw − Qe,rsw,req ) (26)
tb,c ≤ tb
(Qe,max − Qe,rsw,req − Qe,dif )

3. VALIDATION OF THE SIMULATION

The equations, given in the mathematical model section, are solved in the computer medium and the
simulation is performed to determine the heat and mass transfer between the body and its environment, and
also thermal comfort. In order to validate the present model, the simulation results have been compared with
the experimental data of Werner and Reents (1980), Tanabe et al. (1994), and with the numerical results of
Huizenga et al. (2001).

In Table 1, sensible heat losses from a whole human body and body segments are given, and the present
results are compared with the experimental measurements of Tanabe et al. (1994). In the calculations,
2 2
following values are used; Rt = 0.205 m K/W, Re,t = 0.0326 m kPa/W, RH=50% and ta = trd = 24.7°C. It can
be seen that the present results are very close to the measured values.

Table 1. Comparison of heat losses from each part of the body with experimental results
2
Rate of sensible heat loss (W/m )
Segment number *
Experimental Present study Difference (%)
1 70.00 78.90 12.7
2 70.00 78.90 12.7
3 50.00 49.48 1.0
4 51.35 50.77 1.1
5 45.57 44.40 2.6
6 48.60 47.28 2.7
7 35.00 32.44 7.3
8 44.43 43.57 1.9
9 57.00 58.68 2.9
10 60.71 65.97 8.7
11 35.57 32.44 8.8
12 40.00 37.67 5.8
13 37.30 35.17 5.7
14 41.43 39.32 5.1
15 38.86 36.49 6.1

564
16 43.10 41.71 3.2
Whole body 45.30 45.97 1.5
*
Tanabe et al. (1994)

Werner and Reents (1980) and Huizenga et al. (2001) investigated the effects of ambient temperatures on
the body segment temperatures at 40% relative humidity. Variation of skin temperatures with ambient
temperature is comparatively given in Fig. 1. In this calculation, it is assumed that body is in rested position
and wears only a short. Results obtained from the present model are also in good agreement with the results
of Werner and Reents (1980) and Huizenga et al. (2001).

Figure 1. Comparison of measured (Werner and Reents 1980) and simulated (Huizenga et al. 2001 and present study) skin temperatures during
ambient temperatures ranging from 10 to 40°C.

4. RESULTS AND DISCUSSION

After the comparisons and validation of the simulation, it can be passed the results which cover the humidity
effects. The relation among ambient temperature, clothing, humidity, and the influence of these parameters
on the body temperatures and also the thermal comfort are investigated in this section.

Variations of the body temperatures with the ambient temperature at two different clothing configurations (Rcl
2
=0.6 and 1.0 clo) are given in Fig. 2. In these calculations, it is assumed as Mact = 60 W/m (for seated-quiet
person), and RH = 50%, 0.05<V<0.10 m/s, ta = trd (for typical indoor conditions). In figures presenting the
temperatures of the body and segments, the curves are formed by referring the minimum value. So, it is
given the increasing rate of the temperature with respect to determined reference value rather than the real
values. Since the initial temperatures of body parts are different according to experimental conditions and
manikin situation in many studies, a qualitative comparison can be more accurate and rationalist approach
instead of quantitative. From Fig.2, the amounts of increase and decrease of body temperatures can be seen
easily. In this figure, the body temperatures at ta=20ºC and Rcl=0.6 clo are taken as a reference point. For
other cases, it is given the increasing amounts of body temperatures in respect of the reference value.
Because of the fact that the heat losses from the body to environment decrease with increasing ambient
temperature, as expected, the body temperatures (tsk, tcr and tb) increase. Moreover, variation amount in the
skin temperature is higher than that in the core and the mean body temperatures. At the ambient
temperature in the range of 20 to 30°C, the maximum change of skin temperature is 3.8°C while the core
temperature is almost stable. This figure also shows that generally the body temperatures have smaller
values for low clothing insulation, and the core temperature of the body doesn’t change markedly with the
clothing insulation. The core temperatures for Rcl =0.6 and 1.0 clo are nearly coincided. In addition, the
difference between skin temperatures for Rcl =0.6 and Rcl=1.0 clo decreases with increasing ambient
temperature although it is high at the low ambient temperatures. So, the effect of clothing on the skin
temperature is greater at low ambient temperatures.

Under the same conditions (environmental and personal), variation of the heat losses from the whole body is
shown in Fig. 3. When the ambient temperature increases, temperature difference between the ambient and
skin decreases; for this reason, sensible (convection + radiation) heat losses fall down but latent
(evaporation + diffusion) heat losses boost. To maintain heat balance between the body and its environment

565
depending on decreasing sensible heat losses from the body, evaporation heat losses increase with
increasing sweat generation. Besides, total heat losses from the skin (sensible + latent) slightly increase as
well. At low clothing insulation, the total heat loss and the sum of the convective and radiative heat losses
have higher values due to low thermal resistance. Meanwhile, the evaporative heat loss for Rcl = 0.6 clo has
lower values in comparison to that for Rcl = 1.0 clo. Because the body can easily transmit the heat at the low
clothing resistance by using sensible heat loss mechanisms, the sweat generation is lower and so the
evaporative heat loss.

Figure 2. Increment of body temperatures vs. ambient Figure 3. Heat loss vs. ambient temperature
temperature

The effects of three different relative humidity levels as low (10%), intermediate (50%) and high (90%) on the
skin wettedness are given in Fig. 4. In these calculations, the environmental and personal conditions except
for the relative humidity are assumed as the same conditions given above. As known, if the water vapor
partial pressure in the ambient air boosts, the relative humidity increases. In this circumstance, skin
wettedness increases as moisture on the skin does not penetrate easily into ambient air. Skin wettedness is
a parameter connected with evaporative potential of the environment together with generated sweat. If the
evaporative potential of the environment decreases with increasing relative humidity, skin wettedness
increases. Relative humidity value is an important factor for the thermal comfort, especially in high ambient
temperatures. While skin wettedness is 0.06 for each three relative humidity values (10, 50 and 90%) at the
ambient temperature of 20°C, the difference among skin wettedness’ increases with increasing of ambient
temperature.

Figure 4. Skin wettedness vs. ambient temperature

The effect of the relative humidity on the skin and core temperatures is smaller than that of the ambient
temperature (Figs. 5 and 6). At the ambient temperature of 30°C and the clothing insulation of 1 clo, when
the relative humidity varies from 10% to 90%, variation in the skin temperature is 0.4°C while that in the core
temperature is only 0.05°C (Fig. 5). Furthermore, the difference between the skin temperatures for clothing
resistances 0.6 and 1 clo is high at the low ambient temperatures. The temperature difference decreases

566
with increasing ambient temperature. So, clothing resistance has more effect on the skin temperature at low
ambient temperatures. As it can be seen from Fig. 6, while relative humidity has a small effect on the body
core temperature at the low ambient temperatures, it has a greater effect on this temperature at the high
ambient temperatures. As the ambient temperature increases, the difference among the core temperatures
increases for various relative humidity values.

Figure 5. Increment of skin temperature vs. ambient Figure 6. Increment of core temperature vs. ambient
temperature temperature

The effects of relative humidity on the thermal comfort perception are shown in Figs. 7 and 8. In these
figures, variation of TSENS and DISC thermal comfort indices with the ambient temperature are given for
different relative humidity and clothing insulation values. Since body temperatures and skin wettedness
increase with increasing ambient temperature, comfort sense gets worse. While relative humidity has a small
effect on the TSENS and DISC values at the low ambient temperatures, it has a greater effect on these
values at the high ambient temperatures. Relative humidity has a strong influence on the DISC which takes
into account regulation by sweating. At the ambient temperature of 30°C and the clothing insulation of 1 clo,
while DISC is 1.6 for the 10% relative humidity, it is 5.3 for the 90% relative humidity. DISC, which is related
to skin wettedness, has a high values at the high temperature and humidity.

Figure 7. TSENS vs. ambient temperature Figure 8. DISC vs. ambient temperature

In the following figures, some of the body parts are investigated in detail for different humidity values at Rcl =
1.0 clo. Variations of the skin temperature and wettedness of, for example, the pelvis and fibula are shown in
Figs. 9 and 10. If the ambient temperature gets high values, the body segments temperatures increase
similar to the mean body skin temperature. Since pelvis has more clothes than fibula, skin temperature and
wettedness of the pelvis have higher values. Skin temperature of the pelvis rapidly increases at 30°C
ambient temperature and 50% RH. Because the skin wettedness of the pelvis reaches value of 1, which is
the maximum value, at a little lower than ta=30°C, it is not increase anymore at the higher ambient
temperatures. As the skin wettedness reaches the maximum value, the evaporation loss remains constant,
and as a result of this, skin temperature increases. A similar situation is shown in Fig. 11 for back segment.

567
At the same environmental conditions (~30°C, 90%), skin temperature of the back suddenly increases.
Because, as it can be seen from Fig. 12, the skin wettedness of the back reaches value of 1, which is the
maximum value, at a temperature lower than 30°C ambient temperature, and since the heat balance is not
maintained at the higher ambient temperatures, the skin temperature of the back rapidly increases. When
the ambient temperature increases, increasing ratio of the head skin temperature is relatively faster than that
of the back skin temperature. So, the head is more affected from the environmental conditions, because it is
naked.

Figure 9. Increment of Fibula and Pelvis skin temperatures vs. Figure 10. Fibula and Pelvis skin wettedness vs. ambient
ambient temperature temperature

Figure 11. Increment of Head and Back skin temperatures Figure 12. Head and Back skin wettedness vs. ambient
vs. ambient temperature temperature

Variation of the required ambient temperature for the thermal comfort conditions with humidity ratio in the air
for two different clothing (Rcl = 0.6 and 1 clo) is given in Fig. 13. Values of TSENS and DISC are equal to
zero on these curves presenting for Rcl = 0.6 clo and Rcl = 1 clo. As it can be seen, as the humidity ratio in
the air increases, the required ambient temperature should be decreased for applying thermal comfort
conditions. At Rcl = 0.6 clo, while the required ambient temperature is 23.3°C for the relative humidity of 20%,
it is 23°C for the relative humidity of 60%. It is also seen in this figure that for the thermal comfort conditions,
the required ambient temperature decreases with increasing clothing resistance. These temperatures are
required temperatures keeping heat balance between the body and environment. But, apart from the heat
balance, there are various factors affecting the thermal comfort in different ways. For instance, at the high
and low humidity cases, human can feel uncomfortable due to respiration arduous, skin, eyes and throat

568
trouble. Therefore, even if energy balance is maintained, in general the cases, which are relative humidity
higher than 70% and lower than 30%, do not considered as a thermal comfort zone.

Figure 13. Thermal comfort conditions in psychometric chart

5. CONCLUSION
The main results derived from this study are as follows:
1. Relative humidity of air affects the heat balance between the body and environment in numerous ways.
2. Relative humidity value is an important factor for the thermal comfort, especially in high ambient
temperatures. The effect of this factor increases with clothing resistance.
3. Water vapor partial pressure increases with increase in relative humidity at constant temperature. As
the difference between the skin and environment water vapor partial pressures decrease, total latent
heat losses from the skin decreases. For that reason, the core and skin temperature of the body
increase slightly with increasing the relative humidity. But, variation of the core temperature is quite
lower than that of the skin temperature.
4. With increasing the relative humidity, the difference between the skin and environment water vapor
partial pressures decrease, and so greater portion of the body is covered by sweat. For that reason, if
the relative humidity increases, skin wettedness boosts.
5. When the clothing has lower thermal resistance values, the total heat loss and the sum of the
convective and radiative heat losses rise. Meanwhile, the evaporative heat loss related to the sweat
generation of the body increases with increasing clothing resistance.
6. With both the body temperatures and skin wettedness increase, thermal comfort perception is
negatively influenced. As a result of this, TSENS and DISC values increase with the ambient
temperature, relative humidity and clothing resistance.
7. Increase in the DISC, which is parallel with the skin wettedness, is higher than that in the TSENS.
8. Since pelvis and back have more clothes from the other body segments, the skin temperature and
wettedness of these segments is higher.
9. After the skin wettedness reaches the maximum value (w=1), if the ambient temperature keeps on
increasing, skin temperature increases rapidly. This effect is shown in back and pelvis at 90% relative
humidity for approximately 30°C ambient temperature.
10. Since the head has no cloth, it is more affected from the environmental conditions than the other
segments. Increasing ratio of the skin temperature in this segment is higher than that of the back
segment.
Due to the fact that an understanding of human thermoregulatory processes and effects of humidity on the
energy and water balance of the body facilitates the design and development of improved heating and
cooling systems, this study can be useful for HVAC engineers and researchers.

569
 2
NOMENCLATURE Ral thermal resistance of air layer, (m .K)/W
2 2
Ab body surface area, m Rcl thermal resistance of clothing, (m .K)/W
2
cp,b specific heat of body, kJ/(kg.K) Rf thermal resistance of fabric, (m .K)/W
DISC discomfort Rt total thermal resistance of clothing,
2
fcl clothing area factor (m .K)/W
hcv convective heat transfer coefficient, Re,a evaporative resistance of outer air,
2 2
W/(m .K) (m .kPa)/W
hrd radiative heat transfer coefficient, Re,al evaporative resistance of air layer,
2 2
W/(m .K) (m .kPa)/W
i segment number Re,cl evaporative resistance of clothing,
2
icl vapor permission efficiency of clothing (m .kPa)/W
j air or fabric layers number Re,f evaporative resistance of fabric,
2
K effective conductance between core and (m .kPa)/W
2
skin, W/(m .K) Re,t total evaporative resistance of clothing,
2
LR Lewis Ratio, °C/kPa (m .kPa)/W
m body mass, kg Ra gas constant of air, kJ/(kg.K)
2
M rate of metabolic heat production, W/m
Mact rate of metabolic heat production due to a Rv gas constant of vapor, kJ/(kg.K)
2
ctivity, W/m RH relative humidity
2
m bl 2
mass flow rate of blood, kg/(s.m ) Scr rate of heat storage in the core, W/m
2
Ssk rate of heat storage in the skin, W/m
m res mass flow rate of air inhaled, kg/s ta air temperature,°C
m rsw mass flow rate of regulatory sweat tb body temperature,°C
generation, kg/(s.m )
2 tb,c lower limit temperature of evaporative
nl number of layers covering segment regulation zone,°C
pa water vapor pressure in the ambient air, tb,h upper limit temperature of evaporative
kPa regulation zone,°C
pa,s saturated water vapor pressure in the tcr core temperature,°C
ambient air, kPa tex exhaled air temperature,°C
patm atmosphere pressure, kPa to operative temperature,°C
psk,s saturated water vapor pressure at the skin trd mean radiant temperature,°C
temperature, kPa
2
tsk skin temperature,°C
Qcd heat flow rate due to conduction, W/m TSENS thermal sensation
Qcr,sk heat flow rate between core and skin, w total skin wettedness
2
W/m wdif skin wettedness due to diffusion
2
Qcv heat flow rate due to convection, W/m wrsw skin wettedness due to regulatory
2
Qrd heat flow rate due to radiation, W/m sweating
Qres total heat flow rate due to respiration, Wa humidity ratio of ambient air, kg H2O / kg
2
W/m dry air
Qe,dif latent heat flow rate due to diffusion, Wex humidity ratio of exhaled air, kg H2O / kg
2
W/m dry air
Qe,max maximum latent heat flow rate due to W rate of external work accomplished,
2
evaporation, W/m W/m
2
Qe,res latent heat flow rate due to respiration, WSIGb warm signal from body
2
W/m WSIGcr warm signal from core
Qe,rsw latent heat flow rate due to regulatory WSIGsk warm signal from skin
2
sweat generation, W/m α fraction of total body mass concentrated
Qe,rsw,req required latent heat flow rate due to in skin compartment
2
regulatory sweat generation, W/m
2 θ time, s
Qe,sk total latent heat flow rate from skin, W/m
ηe evaporative efficiency
Qs,res sensible heat flow rate due to respiration,
2
W/m
r outer radius of fabric layer, m
2
Ra thermal resistance of outer air, (m .K)/W

REFERENCES
ASHRAE 1989. ASHRAE handbook – Fundamentals, chapter 8. Atlanta: American Society of Heating,
Refrigeration and Air-Conditioning Engineers. 29p.

Burch S.D., Pearson J.T., Ramadhyani S. 1991. Analysis of passenger thermal comfort in an
automobile under severe winter conditioning. ASHRAE Transactions 97:239-246.

Berglund L.G. 1998. Comfort and humidity. ASHRAE Journal 40(8): 35-41.

570
Berglund L.G., Cain W. 1989. Perceived air quality and the thermal environment. Proc. IAQ’89, San
Diego, Ca, 93-99.

de Dear R.J., Arens E., Hui Z., Ogura M. 1997. Convective and radiative heat transfer coefficients for
individual human body segments. Int J Biometeorol 40: 141-156.

de Dear R.J., Knudsen H.N., Fanger P.O. 1989. Impact of air humidity on thermal comfort during step-
changes. ASHRAE Transactions 95(2): 336-350.

Doherty T.J., Arens E. 1988. Evaluation of the physiological bases of thermal comfort models.
ASHRAE Transactions 94 Part 1: 1371-1385.

Fang L., Clausen G., Fanger P.O. 1996. The impact of temperature and humidity on perception and
emissions of indoor air pollutants. Indoor Air, Tokyo, Inst of Public Health 4: 349-354.

Fanger P.O. 1972. Thermal comfort. New York: McGraw-Hill.

Fiala D., Lomas K.J., Stohrer M. 1999. A computer model of human thermoregulation for a wide range
of environmental conditions: the passive system. J App Physiol 87(5): 1957-1972.

Fountain M.E., Arens E.A., Xu T., Bauman F.S., Oguru M. 1999. An investigation of thermal comfort at
high humidities. ASHRAE Transactions 94: 94-102.

Gagge A.P., Stolwijk J.A.J., Nishi Y. 1971. An effective temperature scale based on a simple model of
human physiological response. ASHRAE Transactions 77 Part1: 247-262.

Guan Y., Hosni M.H., Jones B.W., Gielda T.P. 2003 Literature review of the advances in thermal
comfort modeling. ASHRAE Transactions 109: 908-916.

Höppe P., Martinac I. 1998 Indoor climate and air quality. Int J Biometeorol 42: 1-7.

Huizenga C., Hui Z., Arens E. 2001. A model of human physiology and comfort for assessing complex
thermal environments. Building and Environment 36: 691-699.

Kaynakli O., Kilic M. 2005. Investigation of indoor thermal comfort under transient conditions. Building
and Environment 40(2): 165-174.

Kaynakli O., Pulat E., Kilic M. 2005. Thermal comfort during heating and cooling periods in an
automobile. Heat and Mass Transfer 41: 449-458.

Kaynakli O., Unver U., Kilic M. 2003. Evaluating thermal environments for sitting and standing posture.
Int Comm Heat Mass Transfer 30(8): 1179-1188.

Liviana J.E., Rohles F.H., Bullock O.D. 1988. Humidity, comfort and contact lenses. ASHRAE
Transactions 94(1): 3-11.

McCullough E.A., Jones B.W., Tamura T. 1989. A data base for determining the evaporative
resistance of clothing. ASHRAE Transactions 95(2): 316-328.

Olesen B.W., Hasebe Y., de Dear R.J. 1988. Clothing insulation asymmetry and thermal comfort.
ASHRAE Transactions 94(1): 32-51.

Tanabe S., Arens E.A., Bauman F.S., Zang H., Madsen T.L. 1994. Evaluating thermal environments
by using a thermal manikin with controlled skin surface temperature. ASHRAE Transactions 100(1):
39-48.

Teodosiu C., Hohota R., Rusaouena G., Woloszyn M. 2003. Numerical prediction of indoor air
humidity and its effect on indoor environment. Building and Environment 38: 655-664.

571
Toftum J., Fanger P.O. 1999. Air humidity requirements for human comfort. ASHRAE Transactions
105(2): 81-86.

Toftum J., Jorgensen A.S., Fanger P.O. 1998a. Upper limits for indoor air humudity to avoid
uncomfortably humid skin. Energy and Buildings 28: 1-13.

Toftum J., Jorgensen A.S., Fanger P.O. 1998b. Upper limits of air humudity for preventing warm
respirotory discomfort. Energy and Buildings 28: 15-23.

Werner J., Reents T. 1980. A contribution to the topography of temperature regulation in man. Europ J
App Physiology 45: 87–94.

572
 INFLUENCES OF FERROMAGNETIC DEFLECTORS BETWEEN LAYERS OF
SUPERCONDUCTING POWER TRANSMISSION CABLES ON TRANSPORT CURRENT
LOSSES

1 2,*
A. Cicek and F. Inanir
1
Department of Physics, Faculty of Arts and Sciences, Mehmet Akif Ersoy University, Campus 15100,
Burdur/Turkey.
Department of Physics, Faculty of Arts and Sciences, Recep Tayyip Erdoğan University, 53100, Rize/Turkey.
2
*
Corresponding author. E-mail: inanir@ktu.edu.tr

ABSTRACT
Alternating-current losses in two-layer power transmission cables constructed by type-II
superconducting strips in the presence of ferromagnetic deflectors between the layers are numerically
investigated. Each layer comprises 15 wires with rectangular cross section of 4 mm width and 2 m
height, while the inner and outer layer radii are 20 mm and 21 mm, respectively. Deflectors are
composed of either non-magnetic or strongly-ferromagnetic material, where the width and height of
each is 5 mm and 80 m , re sLosses
ctively. pe are obtained through Finite-Element Method
simulations with respect to amplitude of the applied current with 1 Hz frequency. Use of ferromagnetic
deflectors increase total alternating-current loss in the two layers considerably for small amplitudes,
while the loss approaches that in non-magnetic case at amplitudes around the critical current.
Individual layer losses are such that outer-layer loss is significantly larger, 2.5-fold at 1/16 of the critical
current, for ferromagnetic deflectors, whereas they are almost identical in non-magnetic case at all
amplitudes. Inner and outer wires are exposed to similar self magnetic fields of wires in non-magnetic
case, while ferromagnetic deflectors accumulate magnetic field lines on themselves and increase the
losses in outer wires. The current profile is homogeneous except at the edges of the wires in the non-
magnetic case, whereas homogeneity is disrupted for the outer wire in the case of ferromagnetic
deflectors, such that current flow in the positive direction is confined to the central region.

Keywords: Two-layer superconducting cable, ferromagnetic deflector, alternating-current loss, Finite

Element Method.

1. INTRODUCTION
Among promising applications of high-temperature superconductors (HTS) is power transmission.
HTS power transmission cables generally operate at temperatures between 64 K and 77 K. The most
important advantages of these cables on conventional ones are the ability to transport high currents
and their efficiency.

In evaluation of the efficiency of HTS cables, determination of alternating-current (AC) transport losses
plays a central role. Primary factors affecting AC losses include (i) the number and dimensions of the
wires composing a cable, (ii) geometrical alignment of wires on cables and (iii) the twist-pitch length.

The critical current densities (Jc) of HTS cables depend heavily on the magnetic field distribution at
elevated temperatures. In such power transmission cables, the total critical current density is higher
than the sum of the corresponding densities of the strips forming the cable [1]. The reason for this lies
in the fact that the variation in magnetic field distribution takes place in a more favorable manner,
especially along the edges of an individual strip.

Ferromagnetic deflectors are successfully utilized in HTS coils and practical applications [2,3]. It is
reported that a correction in the critical current density (Jc) is required in strips coated by a
ferromagnetic jacket [4]. The ferromagnetic material restrains coupling between the filaments within
the superconductor [5]. This is expected to give rise to a decline in the transport AC losses in HTS
cables. This is demonstrated by occupying the space between the filaments which inhibits the
interaction between the filaments and facilitates decrease in AC losses [6]. Screening in HTS strips
coated by ferromagnetic materials is also studied [7,8]. Furthermore, influence of magnetization on
AC transport losses in superconductors over a ferromagnetic substrate due to applied magnetic field is
studied [9,10]. It is reported that an important improvement in the performance of BiSCCO-2223 strips
coated by a partially-ferromagnetic material can be achieved [11]. Vojenčiak et al [12] demonstrated
both theoretically and experimentally that introduction of ferromagnetic deflectors in the space
between the strips of a single-layer HTS cable gives rise to a reduction in AC losses.

573
Application of ferromagnetic deflectors between the layers of a two-layer HTS cable could also give
rise to a reduction in AC transport losses in such cables. This work focuses on a numerical
investigation on the influences of ferromagnetic deflectors introduced in the space between the HTS
strips in a two-layer cable along the radial direction. The variation of transport current losses with
respect to the applied current for deflectors with varying permeability is studied.

2. CABLE GEOMETRY AND COMPUTATIONAL METHODS

The two-layer HTS cable geometry is depicted in Figure 1. Each layer comprises N=30 rectangular
superconducting tapes with width and height as wsc=4.0 mm and hsc=2.0 m , whe re the
outer layers are on circular arcs with Ri= 20.0 mm and Ro=21.0 mm, respectively. The wire parameters
are compatible with the standards set by American Superconductors Inc. The core of the cable is
®
made of copper (Cu), whereas the space between the layers is filled with Kapton dielectric. The
ferromagnetic deflectors between the layers also possess rectangular cross section with wT=5.0 mm
and hT=80.0 m
, Figure 1(b).

The two-layer HTS cable in Figure 1(a) possesses discrete azimuthal symmetry with respect to a unit
of rotation by C=2 /30 (12º), wh
RT=20.5 mm, are rotated by T=C/2=2 /62 (5.8º)
Due to the with re s pe ct to the wir
above-
®
mentioned symmetry, the problem in 2 dimensions (2D) is solved through the Comsol MultiPhysics
package with AC/DC module, which is a commercial implementation of the Finite Element Method
(FEM), by only considering the periodic element in Figure 1(a). Thus, the two wires and halves of the
deflectors, which are depicted by the filled rectangles with solid borders, are considered in
computations, Figure 1(a).

Figure 1. (Color online) Geometry of the two-layer HTS cable with ferromagnetic deflectors (a) and a close-up view of the wires and
deflectors (b). Elements are not drawn to scale for clarity.

The HTS wires are composed of second-generation yttrium barium copper oxide (YBCO)
superconductors, whereas the deflectors are either non-magnetic or strongly ferromagnetic with
relative permeability of r=1.0 and 5000, respectively. The magnetic flux dependence of the
ferromagnetic deflectors is assumed as B=0rH and ferromagnetic losses are ignored. The influence
of the substrate is neglected for simplicity, so that it is not depicted in Figure 1(a). Moreover, twisting of
the wires along the z direction is also not taken into account. The choice of geometrical and physical
parameters is to facilitate a comparison between other possible configurations of the cable with
respect to the use of ferromagnetic material.

In FEM simulations, an alternating current with a frequency of f is applied along the z direction and the
component of the vector potential along this direction (Az) is treated as the independent variable. The
rotational of the vector potential yields magnetic flux. The critical current for each wire, which is
calculated through an integration of the critical current density (Jc) over the cross-sectional area (ASC)
of the wire on the xy-plane,

574
 I c = ∫ J c dASC (1)
ASC

is taken as Ic= 80 A. Dependence of Ic on magnetic flux density and position, which may be important
in case of wires with circular cross-sections, is ignored. The current is sinusoidal with
I(t)=Imax.sin(2 
ft), where Imax is its amplitude with respect to Ic and f=1 Hz.

The boundary condition for addressing the cable geometry in Figure 1(a) considers the vector potential
at an arbitrary position r due to an HTS wire in Figure 1(b) as

µ I  r   r  
Ar =
− 0  ln   + ln   + 2  (2)
2π   Ri + hSC   Ro + hSC  

Besides, the azimuthal boundary condition exploits continuity of the normal component of the
magnetic field. The superconducting current density in computations can be written as

 Ap ( x, y ) − ( Az ( x, y ) + ∇V ) 
jsc , z ( x, y ) = J c tanh   (3)
 An 
 

where Ap(x,y) is the distribution of Az(x,y) in the previous step in the time march of simulations,
whereas An is the scaling parameter which adjusts the sharpness of the translations between the
region in superconductor with positive and negative current densities and ∇V is the scalar potential.
By numerically solving Ampere’s Law in combination with Equations (2) and (3), the distributions of
jsc,z and E = −∂A / ∂t − ∇V over a wire’s cross-section are calculated and AC loss, Q (J/m), in each
wire is obtained, in turn, by:

=
Q ∫ E ⋅ jdA
Asc
sc
(4)

3. RESULTS AND DISCUSSION

Variation of the total AC transport loss (Qe) of the HTS cable on each periodic element in Figure 1(a)
with respect to the applied current amplitude is depicted in the log-log scale in Figure 2. In
computations Imax is set to 8 different values ranging from 5.0 A to 75.0 A (Ic/16 to 15Ic/16). Qe is
calculated as the sum of the losses in inner (Qi) and outer (Qo) layers.

Figure 2. (Color online) Variation of total AC transport loss on each periodic element with respect to the applied current amplitude for non-
magnetic and strongly-ferromagnetic deflectors.

Figure 2 demonstrates that use of strongly-ferromagnetic deflectors significantly increases Qe, while
the discrepancy between the two materials vanishes for large current amplitudes close to Ic. The

575
losses increase almost linearly in Figure 2 for Imax<Ic/2, whereas the increase is slightly steeper at
-6 -4 -3
higher amplitudes. The calculated losses for r=1.0 are 2.98×10 J/m, 1.74×10 J/m and 1.91×10
J/m for Imax=5.0 A (Ic/16), 35.0 A (7Ic/16) and 75.0 A (15Ic/16), respectively. Corresponding losses for
r=5000 are 3.82×10-6 J/m, 2.47×10-4 J/m and 1.93×10-3 J/m, respectively. These values are 28.2%,
41.4% and 0.8% higher than the corresponding values in the r=1.0 case.

To elucidate the loss mechanism due to the presence of deflectors, individual losses on each layer
should be considered. Figure 3 depicts the variation of Qi and Qo with respect to Imax. In case of r=1.0
Qi and Qo are almost equal to each other (and thus to Qe/2), as expected. In contrast, a significant
discrepancy is observed for r=5000, where Qi is considerably smaller at all Imax values. In fact, Qi is
also smaller than the corresponding value for the non-magnetic deflector case. The value of Qi and Qo
-6 -6
are 1.1×10 J/m and 2.7×10 J/m, respectively, such that Qo is 145.4% higher than Qi, at Imax=5.0 A
-
(Ic/16) for the ferromagnetic deflectors. The corresponding value for non-magnetic deflectors is 1.5×10
6
J/m, which is 35.8% larger than the Qi value at the same amplitude for the ferromagnetic case. Qi
and Qo for r=5000 becomes
-5 -4
5.6×10 J/m and 1.9×10 J/m, respectively at Imax=35.0 A (7Ic/16), that is Qo is almost four-fold higher
at amplitudes around Ic/2. At the highest amplitude (15Ic/16), the corresponding values are calculated
-4 -3
as 4.7×10 J/m and 1.5×10 J/m, respectively. Although presence of ferromagnetic deflectors
increases the total AC transport loss, the loss in the inner layer of the cable can be tuned by modifying
the deflector parameters. This can particularly be important in technological applications in order to
reduce process and cooling-down costs [13].

Figure 3. (Color online) Comparison of individual AC transport losses on the inner and outer layers of the HTS cable for non-magnetic and
strongly-ferromagnetic deflectors.

How the relative permeability of deflectors influence AC losses is closely related to how they
manipulate the magnetic field distribution around the HTS wires. Figure 4 presents distribution of the
equi-potential lines of the magnetic vector potential (Az) for non-magnetic and strongly-ferromagnetic
deflectors at an instant, at which the instantenous current on a wire is I(t)=18.5 A, of the sine-wave
cycle of the applied current with Imax=75.0 A.

Figure 4. (Color online) Distribution of the equi-potential lines of the magnetic vector potential (Az) around the wires for 
r=1.0 (a) and 5000
(b) at an instant of the applied sinusoidal current such that the instantaneous current takes a moderate value (18.5 A).

Non-magnetic deflectors do not affect the magnetic field lines due to current distributions in the wires
as expected, Figure 4(a). Besides, the wires on the inner and outer layers are exposed to similar self-

576
fields. Thus, Qi and Qo for 𝜇𝑟 =1.0 in Figure 3 are almost equal. On the other hand, strongly-
ferromagnetic deflectors both trap and concentrate field lines on themselves, Figure 4(b). Moreover,
the inner layer is exposed to its self-field to a lesser extent in this case. This is the reason behind the
fact that Qi for the ferromagnetic deflectors is smaller than both Qo in the same configuration and the
corresponding Qi for non-magnetic deflectors, as seen in Figure 3. Moreover, the normal component of
the magnetic field is more dominant at the edges of the outer wire for r=5000. The above facts
indicate that current distributions in the wires are affected from the presence of the ferromagnetic
deflectors, as depicted in Figure 5.

Figure 5. (Color online) Distribution of current amplitude over the inner (lower) and outer (upper) wires of the HTS wire for non-magnetic (a)
and strongly-ferromagnetic (b) deflectors at an instant of the applied sinusoidal current such that the instantaneous current takes a
moderate value (18.5 A).

Figure 5(a) depicts that current flow is homogeneous in the interiors of both wires, while higher current
amplitudes close to Ic are confined to the edges parallel to y direction in case of r=1.0. Furthermore,
the current distribution is almost identical in inner and outer wires, Figure 5(a). On the other hand, the
situation is quite different in case of r=5000, as depicted in Figure 5(b). The current flow is more
homogeneous in the inner wire, lower part of Figure 5(b), than in the case for non-magnetic deflectors.
For this wire, the high current flow at the edge is confined to a smaller area. However, the region
expands more into the wire along x direction in the outer wire, upper part of Figure 5(b). This, in turn,
indicates that magnetic field penetrates deeper into the outer wire in case of ferromagnetic deflectors.
The current distribution patterns in Figure 5 are consistent with distribution of magnetic field lines in
Figure 4.

4. CONCLUSION

Alternating-current transport losses in two-layer high-temperature superconducting cables

accompanied by ferromagnetic deflectors are numerically studied through Finite-Element Method
simulations. Presence of ferromagnetic deflectors with relative permeability as high as 5000, located in
the space between the layers and rotated by half the unit angle of rotation due to discrete azimuthal
symmetry of the cable, increases the total loss relative to the case of non-magnetic deflectors. The
increase in the total loss is due mainly to the increased self-magnetic field around the outer-layer
wires, whereas the ferromagnetic deflectors concentrating magnetic field lines on themselves lead to a
reduction in the individual losses of inner-layer wires. This fact can be utilized in tuning the losses in
the inner layers of multi-layer cables to reduce both operation and cooling-down costs. In consistent
with the distribution of magnetic field lines due to self-fields of the wires in the existence of
ferromagnetic deflectors, the current flow is largely homogeneous in the inner-layer wires, while the
region of high current flow expands more into the outer-layer wires.

Acknowledgments

This study is supported by The Scientific and Technological Research Council of Turkey (TÜBİTAK)
under the grant number 110T876.

REFERENCES
[1] L. Rostila, L. Söderlund, R. Mikkonen, J. Lehtonen, “Modelling method for critical current of YBCO
tapes in cable use”, Physica C: Superconductivity and its Applications 467 (1-2), 91-95 (2007).

577
[2] E. Pardo, J. Šouc and M. Vojenčiak, “AC loss measurement and simulation of a coated conductor
pancake coil with ferromagnetic parts”, Superconductor Science and Technology 22 (7) , 075007
(2009).

[3] H. Wen, W. Bailey, K. Goddard, M. Al-Mosawi, C. Beduz and Y .Yang, “Performance Test of a 100
kW HTS Generator Operating at 67 K–77 K”, IEEE Transactions on Applied Superconductivity 19 (3),
1652-1655 (2009).

[4] P. Kovác, I. Hušek, T. Melišek, M. Ahoranta, J. Šouc, J. Lehtonen and F. Gömöry, “Magnetic
interaction of an iron sheath with a superconductor”, Superconductor Science and Technology 16
(10), 1195–1201 (2003).

[5] B. A. Glowacki and M. Majoros, “A method for decreasing transport ac losses in multifilamentary
and multistrip superconductors”, Superconductor Science and Technology 13 (7), 971–973 (2000).

[6] M. Majoros, M. D. Sumption and E. W. Collings, “Transport AC Loss Reduction in Striated YBCO
Coated Conductors by Magnetic Screening”, IEEE Transactions on Applied Superconductivity 19
(3),3352–3355 (2009).

[7] S. V. Yampolskii, Y. A. Genenko, H. Rauh, “Penetration of an External Magnetic Field into a

Multistrip Superconductor/Soft-Magnet Heterostructure”, Physica C: Superconductivity and its
Applications 460–462 (2), 1262–1263 (2007).

[8] S. Farinon, P. Fabbricatore, F. Gömöry M. Greco and E. Seiler, “Modeling of current density
distributions in critical state by commercial FE codes”, IEEE Transactions on Applied
Superconductivity 15 (2), 2867–2870 (2005).

[9] E. Seiler and F. Gömöry, “Modelling of the flux penetration into a superconducting strip with
magnetic sheath”, Journal of Physics: Conference Series 43, 9–13 (2006).

[10] M. Suenaga, M. Iwakuma, T. Sueyoshi, T. Izumi, M. Mimura, Y, Takahashi and Y. Aoki, “Effects of
a ferromagnetic substrate on hysteresis losses of a YBa2Cu3O7 coated conductor in perpendicular ac
applied magnetic fields”, Journal of Physics: Conference Series 97, 012025 (2008).

[11] S. Safran, M. Vojenčiak, A. Gencer and F. Gömöry, “Critical Current and AC Loss of DI-BSCCO
Tape Modified by the Deposition of Ferromagnetic Layer on Edges”, IEEE Transactions on Applied
Superconductivity 20 (5), 2294-2300 (2010).

[12] M. Vojenčiak, J. Šouc and F. Gömöry, “Critical current and AC loss analysis of a superconducting
power transmission cable with ferromagnetic diverters”, Superconductor Science and Technology 24
(7), 075001 (2011).

[13] J. Gerhold and T. Tanaka, “Cryogenic electrical insulation of superconducting power transmission
lines: transfer of experience learned from metal superconductors to high critical temperature
superconductors”, Cryogenics 38 (11), 1173-1188 (1998).

578
 THERMAL COMFORT ANALYSIS OF NOVEL LOW EXERGY RADIANT HEATING COOLING
SYSTEM AND ENERGY POTANTIAL COMPARING TO CONVENTIONAL SYSTEMS

Aliihsan Koca*, Zafer Gemici*, Koray Bedir*, Erhan Böke**, Barış Burak Kanbur*, Yalçın Topaçoğlu*
*Mir Arastirma ve Gelistirme Inc.,**Istanbul Technical University
*Y.T.U Technopark A1 Building, 34220 Istanbul, Turkey, **I.T.U Mechanical Engineering Faculty, 34437 Istanbul, Turkey
ihsankoca@mirholding.com.tr, zafergemici@mirholding.com.tr, koraybedir@mirholding.com.tr, boke@itu.edu.tr,
burakkanbur@mirholding.com.tr, yalcintopacoglu@mirholding.com.tr

ABSTRACT
In this study characteristics of a model room with dimensions 6 m x 4 m x 3 m has been investigated using
Computational Fluid Dynamics method with regard to thermal comfort, in case of that the room is heated and
cooled by the means of radiant panels. Both general thermal comfort parameters (PMV (percentage mean vote),
PPD (predicted percentage of dissatisfied)) and local thermal comfort parameters (radiant temperature asymmetry,
draught, vertical air temperature difference, warm and cool floors) which are described by the standards ISO 7730
and ASHRAE 55 have been taken into consideration. Radiant panels have been placed to exterior walls for
heating system and they have been put both exterior walls and ceiling for cooling system. According to the TS
2164 standards, different regions is identified with regard to outdoor climate conditions for heating and cooling
season in Turkey and analysis of heating/cooling is done for each region. Heat fluxes from radiant panels and
corresponding values of room mean temperature required in order the conditions of thermal comfort described by
the standards ISO 7730 to be met have been determined for these regions and heating/ cooling system
configurations. Energy saving potential of radiant system have been evaluated for these regions and
heating/cooling mode.

1. INTRODUCTION
Energy saving and emission reduction are both affected by the energy efficiency of the built environment and the
quality of the energy carrier in relation to the required quality of the energy. Low exergy heating and cooling
systems allow to use of low valued energy , which is delivered by sustainable energy sources (e.g. by using heat
pumps, solar collectors, either separate or linked to waste heat, energy storage etc.) and improve their
performances.

It is vitally important to have a clear image of the exergy balance of the human body in order to understand what
the low exergy systems for heating and cooling in buildings are. It is interesting that the thermal comfortable
condition is provided with lowest exergy consumption rate with the human body. This suggests that rational
heating and cooling systems in building would well with low exergy consumption under a condition in which we
humans consume as low amount of exergy as possible. That is we may be able to establish both thermal comfort
and low exergy consuming systems at the same time.

Human body doesn’t sense just the temperature of the air, feels the operative temperature that is combination of
air temperature, air humidity, air velocity and radiant temperature. According to this relationship lowest exergy
consumption rate emerges at the point where the room air temperatures equals 18 °C and mean radiant
temperature 25 °C [1]. This suggest that use of warm radiant energy is more effective than use of convective warm
energy for a heating purpose to realise both thermal comfort and as low exergy consumption within human body
as possible.

The work concerning low exergy systems (LowEx) by Virtane and Ala-Juusela [2,3] encourages the use of low
temperature heating systems for buildings. Low exergy heating systems allow the use of low valued energy
sources. Such heating system operates at low temperature levels that are close to room temperature. These
systems can provide the occupants with comfortable, clean and healthy environment. In addition, low temperature
systems lead to efficient use of energy and provide flexibility in the choice of the energy source. To the knowledge
of the authors, there are only a few papers available about low temperature heating systems.
Low temperature heating with heat pump or other low temperature devices has several advantages. When using
for example a heat pump in low temperature systems, energy may be saved because the thermal efficiency of the
pump (COP) increases and energy losses in the distribution net decrease. Energy efficiency aside, studies show
that low temperature heating may improve indoor air quality as well as the thermal comfort conditions [4–7].

Scientist and engineers are concerned about new heating and cooling systems in buildings due to conventional air
conditioning systems consuming large amount of energy. As a result of their studies, radiant heating and cooling
systems which can produce more comfortable conditions in occupied zone and energy saving are developed.

579
Working principle of radiant system is based upon occurring via radiation at least 50% of heat transfer from radiant
panel can be placed on floor, wall or ceiling. By the radiant panel heat transfer occurs in two ways; primarily via
radiation between panel surface and occupants which constitutes 60-80 percent of heat transfer and secondarily,
heat transfer to indoor air by means of natural convection (Okamoto et al., 2010) [8]. The intent of radiant systems
is to lower thermostat set point temperature in winter and to increase it in summer, resulting in substantial energy
savings for heating and cooling as compared with conventional systems (Laouadi, 2004) [9]. A forced convection
air-conditioning system create uncomfortable environment caused by draught and air temperature differences
between the human head and foot, whereas radiant air-conditioning systems can provide lower vertical air
temperature differences and air velocity. At a given volume flow rate, water is about 4000 times more efficient for
heat transport than air. Therefore, water heating systems are very common in Europe, and hydronic cooling
systems progressively replace air conditioning (Roulet et al., 1999) [10].

2. THERMAL COMFORT
As defined by ASHRAE Standard 55, thermal comfort is that condition of the mind that expresses satisfaction with
the thermal environment. Thermal comfort is investigated considering both general and local thermal sensation of
the human body

2.1. General thermal comfort

General thermal comfort is mainly related to PMV-PPD index can be expressed as mathematically and occupant’s
thermal sensation temperature by the whole body called as operative temperature. PMV consists of six comfort
variables (metabolic rate, clothing insulation, ambient air temperature, mean radiant temperature relative humidity
and air velocity) and is expressed Eq. (1) (ISO 7730, 2005):

( M − W ) − 3.05 ⋅10−3 ⋅ {5733 − 6.99 ⋅ ( M − W ) − pa } 

 
 −0.42 ⋅ {( M − W ) − 58.15} − 1.7 ⋅10−5 ⋅ M ⋅ ( 5867 − pa ) 
(
PMV = 0.303 ⋅ e −0,036⋅M )
+ 0.028  
{ }
(1)
 −0.0014 M ( 34 − Ta ) − 3.96 ⋅10−8 f cl (Tcl + 273)4 − (Tr + 273)4 
 
 − f ⋅ h ⋅ (T − T ) 
 cl cl a 
PPD is expressed via Eq. (2) using PMV index value (ISO 7730, 2005):

PPD = 100 − 95 ⋅ exp(−0.03353 ⋅ PMV 4 − 0.2179 ⋅ PMV 2 ) (2)

Operative temperature is not equal the ambient air temperature and is affected by surfaces and objects
temperatures of indoor environment. When mean air speed is less than 0.2 m/s, operative temperature is
calculated by Eq. (3):

Tr + Ta
T0 = (3)
2
General thermal comfort criteria which described by the standards ISO 7730 are shown in Table 1.

Table 1: General thermal comfort criteria

Parameter Limit value
PMV -0.5 < PMV < 0.5
PPD (%) PPD < 10
Winter
20-24
Operative temperature (1 clo/1.2 met)
(°C) Summer
23-26
(0.5 clo/1.2 met)

2.2. Local thermal comfort

In the earlier years of thermal comfort studies, comfort was usually described as affected by the occupant’s
thermal sensation by the whole body. But aside from the overall thermal state of the body, an occupant may also
find the thermal environment unacceptable if local influences on the body from radiant temperature asymmetry,
draught, vertical air temperature differences, cold or warm floors.

580
Radiant temperature asymmetry is the difference between the maximum and the minimum radiant temperature on
the surfaces of a cube element located at a point in the space being conditioned
(Dudkiewicz and Jezowiecki, 2009).
Draught is the unwanted local cooling of the body caused by air movement.
Vertical air temperature differences is a high vertical air temperature difference between the ankle and the head
(0.1 and 1.1 m above the floor) usually cause discomfort. This air temperature difference should be less than 3°C.
Floor surface temperature is especially important for thermal comfort assessment of spaces with occupants
wearing light indoor shoes or in cases where occupants sit/ lie on the floor or walk indoors with bare feet as in
common in Asia.

3. ROOM MODELING
A CFD model with dimensions 6 m x 4 m x 3 m is designed. Two single glazed a window which is included in the
CFD model has 1.4 m height and 1.2 m weight. Four different room models are obtained to be placed on wall
surfaces of panels with dimensions 0.6 m x 1.2 m as shown Fig. 1, panels are installed all around of the window in
the first model (Fig. 1a), windowless exterior wall in the second model (Fig. 2a), either walls in the third model
(Fig. 3a), ceiling in the fourth model (Fig. 4a). The TS 2164 standards identify four different regions in Turkey with
regard to outdoor climate conditions in winter used in heating systems design. Each region has been represented
in numerical analysis by setting these climatological parameters as boundary conditions in heating mode (Table 2),
where in the case of cooling summer conditions as well as incident radiation through window are taken into
account.

Table 2: Outside air temperature in different region for cooling and heating season
Heating season Cooling season
outside air outside air
Region City
temperature temperature
(°C) (°C)
1 Antalya 3 39
2 Istanbul -3 33
3 Ankara -12 35
4 Erzurum -21 31

(a) (b)

581
 (c) (d)

Fig.1. (a) Model 1 (b) Model 2 (c) Model 3 (d) Model 4

4. NUMERICAL SOLUTION METHOD

In fluid mechanics and heat transfer, basis of conservation equations stated as mathematically treatment of fluid
depend on conservation of mass, momentum and energy. In this study, flow is steady state, three dimensional and
physical properties of the fluid such as density, viscosity, thermal conductivity are constant. Consequently,
conservation equations are obtained as below (White, 2003):
Continuity equation,

∂u ∂v ∂w
+ + =
0 (4)
∂x ∂y ∂z

Momentum equations of x, y, z direction can be stated respectively:

∂u ∂u ∂u 1 ∂p ∂ 2u ∂ 2u ∂ 2u
u +v +w =
− +ν ( 2 + 2 + 2 ) (5)
∂x ∂y ∂z ρ ∂x ∂x ∂y ∂z
∂v ∂v ∂v 1 ∂p ∂ 2v ∂ 2v ∂ 2v
u +v +w = − +ν ( 2 + 2 + 2 ) − g (6)
∂x ∂y ∂z ρ ∂y ∂x ∂y ∂z
∂w ∂w ∂w 1 ∂p ∂ 2w ∂ 2w ∂ 2w
u +v +w =
− +ν ( 2 + 2 + 2 ) (7)
∂x ∂y ∂z ρ ∂z ∂x ∂y ∂z

Energy equation can be stated as below:

∂T ∂T ∂T  ∂ 2T ∂ 2T ∂ 2T 
u +v +w α 2 + 2 + 2 
= (8)
∂x ∂y ∂z  ∂x ∂y ∂z 

Between radiant panels and indoor air heat transfer occurs via natural convection which has been modeled by
employing Boussinesq method which yields good results provided with small difference in extremum values of
temperature within the space in question. Turbulence is simulated by the standard k-ε model which is reported to
be a good approximation especially for near-wall flows. Radiation heat transfer has been represented by
employing the Discrete Ordinates method which takes account scattering, semi-transparent media, reflecting
surface and wavelength-dependent transmission (Fluent User’s Guide, 1998).

5. RESULTS AND DISCUSSION

5.1. Heating condition of room

Room has two exterior and interior walls, a ceiling and floor. This room heating analysis is done when room situate
in self-contained flat. In this case; ceiling, floor and interior walls of the room are neighbor to cockloft, ground and
unheated space, respectively. Heating analysis was made for room model 1,2,3

582
5.1.1. Analysis results heating condition of room
Analysis results will be expressed with regard to model 2. These results were obtained when panel heating
performance was 35 W/m2, considering second region’s outside air condition. Fig. 2a shows velocity streamline in
room. Natural circulation of air flow in contact with warm panel surface is obviously shown in Fig. 2a. On the panel
surface had air speed 0.13 m/s which is the highest air velocity in the room. Fig. 2b represents temperature
distribution on walls, window and panels. Warmed panel, cold window and whole wall surface mean temperatures
were found 296.4, 288, 292 K, respectively in the room.

(a) (b)

Fig.2: (a) streamline distribution in model 2 (b) model 2 surfaces temperature distribution

Investigation of general thermal conditions comfort for room model 2

General thermal comfort is based on PMV, PPD indices and operative temperature. Therefore, these parameters
must be calculated. The PMV value is calculated from Eq. 1. For the purpose of solving this equation, mean
radiant temperature which must be calculated is expressed in Eq. (9) (ASHRAE 55, 2003).

Tp + Taust
Tr = (9)
2
Panel surface temperature (Tp) was found 296.4 K as a result of numerical analysis. Area-weighted unheated
surface temperature (Taust) was calculated 292 K with Eq. (10) (ASHRAE 55, 2003).

(Ts1 ⋅ As1 ⋅ ε s1 + Ts 2 ⋅ As 2 ⋅ ε s 2 + ... + Tsn ⋅ Asn ⋅ ε sn )

Taust = (10)
( As1 ⋅ ε s1 + As 2 ⋅ ε s 2 + ... + Asn ⋅ ε sn )
Depending on these values, mean radiant temperature (Tr) was estimated 294.1 K using Eq. (9) As a result of
numerical analysis, mean air temperature and air velocity was found 292.5 K and 0.013 m/s, respectively. Clothing
in winter, metabolic rate of occupants and relative humidity was admitted 1 clo, 1.2 met, and 50%, respectively
(ISO 7730, 2005). Values of the these six comfort variables are shown in Table 3

Table 3: General thermal comfort input values for model 2

Parameter Input values

Clothing insulation (clo) 1
Mean air temperature (°C) 19.5
Mean radiant temperature (°C) 21.1
Metabolic rate (met) 1.2
Air velocity (m/s) 0.013
Relative humidity (%) 50

583
Operative temperature was computed 293.3 K with Eq. (3). PMV and PPD values was estimated by computer
program which is written based on ASHRAE 55 and ISO 7730 standards. Values of the these parameters are
shown in Table 4

Table 4: General thermal comfort output values for model 2

Parameter Output values

Operative temperature (°C) 20.3
PMV -0.28
PPD (%) 6.63

The relationship between PMV and PPD is shown in Fig. 3. Intersection of PMV and PPD values is within the
region defined by thermal comfort

100

80
PPD (%)

60

40

20

0
-3 -2,5 -2 -1,5 -1 -0,5 0 0,5 1 1,5 2 2,5 3
PMV
Fig.3: The relationship between PMV and PPD for model 2

Investigation of local thermal comfort conditions for room model 2

The occupied zone is defined as a space 0.6 m distant from all walls and up to 1.8 m above floor level
(Myhren and Holmberg, 2008). Most critical plane in the occupied zone is farthest from panels. Reference lines
were all in the critical plane and the reference lines 1, 2 and 3 were 0.6, 3 and 5.4 m distant from exterior wall
which has a window with, respectively (Fig. 4).

Fig.4: Reference line 1,2,3

584
Temperature differences between 0.1 and 1.1 m above the floor on reference line 1, 2 and 3 were found 0.4, 0.3,
0.5°C, respectively. None of the values exceed recommended limit that is given by ISO 7730 (Table 5).

Table 5: Vertical air temperature differences 0.1-1.1 m on reference line

Reference line T0.1 (K) T1.1 (K) T1.1-T0.1 (K) ISO 7730 (K)
1 292 292.4 0.4 <3
2 292.1 292.4 0.3 <3
3 291.8 292.3 0.5 <3

Maximum air velocities on reference line 1, 2 and 3 were obtained 0.013, 0.01, 0.017 m/s, respectively. These air
velocities are less than 0.18 m/s described by ISO 7730 (Table 6).

Table 6: Maximum air velocities on reference line

Maksimum hız ISO 7730
Reference line
(m/s) (m/s)
1 0.013 < 0.18
2 0.01 < 0.18
3 0.017 < 0.18

Fig. 5 represents temperature distribution in the middle plane of the room. The difference between highest and
lowest temperature is 0.34 K, which is under the standart limitation of 23 K (radiant temperature asymmetry
caused by warm vertical wall) set by ISO 7730.

Fig.5: Temperature distribution in the middle plane of the room

Room floor surface temperature was found 19.2°C which is the range of floor surface temperatures (19-29°C)
described by ISO 7730

Investigation of thermal comfort as regards clothing insulation and metabolic rate

The relationship between PPD and clothing insulation is shown in Fig. 6. Thermal comfort requirement is provided
in the range of 1-1.6 clo. Clothing insulation and PPD values of the optimal thermal comfort are 1.2 clo and 5%,
respectively.

585
 35
30
25

PPD (%)
20
15
10
5
0
0,5 1 1,5 2 2,5
Icl (clo)
Fig.6: The relationship between PPD and clothing insulation

Fig. 7 presents the relationship between PPD and metabolic rate. Thermal comfort is satisfied from 1.2 to 1.6 clo
so occupants feel comfortable themselves this metabolic rate range in room. Clothing insulation and PPD values
of the optimum thermal comfort are 1.3 met and 5.13%, respectively.

45
40
35
30
PPD (%)

25
20
15
10
5
0
0,5 1 1,5 2 2,5
Metabolic rate (met)
Fig.7: The relationship between PPD and metabolic rate

5.2. Cooling condition of room

Room cooling analysis is done when room situate in mezzanine so heat loss takes place from exterior walls and
window only. Cooling analysis was made for all room model

5.2.1. Analysis results cooling condition of room

Analysis results will be explained considering model 4. These results were obtained when panel cooling
performance was 18 W/m2, considering second region’s outside air condition Fig. 8a represents velocity streamline
in model 4 and maximum air speed was 0.111 m/s. Fig. 8b demonstrates temperature distribution on whole room
surfaces. Warm window, cooled panel and other wall surface temperatures was found 302.2, 299, 296.8 K,
respectively.

586
 (a) (b)
Fig.8: (a) streamline distribution in model 4 (b) model 4 surfaces temperature distribution

Investigation of general thermal comfort conditions for room model 4

Parameters of calculating PMV were given as follows: clothing, 0.5 clo in summer; metabolic rate of occupant, 1.2
met; relative humidity, 50% (ISO 7730, 2005). As a result of numerical analysis, mean air temperature and air
velocity was found 298.5 K and 0.018 m/s, respectively and mean radiant temperature was found 297.9 K using
with Eq. (9). Consequently, values of the these six comfort variables are shown in Table 7

Table 7: General thermal comfort input values for model 4

Parameter Input values

Clothing insulation (clo) 0.5
Mean air temperature (°C) 25.5
Mean radiant temperature (°C) 24.9
Metabolic rate (met) 1.2
Air velocity (m/s) 0.018
Relative humidity (%) 50

PMV, PPD and operative temperature values are shown in Table 8. These values show that thermal comfort is
provided

Table 8: General thermal comfort output values for model 4

Output
Parameter
values
Operative temperature
25.2
(°C)
PMV 0.21
PPD (%) 5.91

5.3. Minimum heat flux from radiant panels in the case of heating and cooling
Minimum heat fluxes from radiant panels and corresponding values of PMV, PPD, room mean temperature are
shown for different region and heating/cooling in Table 9,10. Mean air temperature of model 1 became lower in
cooling and higher in heating than other models for different regions

Table 9: Minimum heat fluxes from radiant panels for different regions in heating
Model 1 Model 2 Model 3
Region 1 2 3 4 1 2 3 4 1 2 3 4
Heat flux (W/m2) 60 70 90 110 30 35 45 55 15 20 25 30
PMV -0.47 -0.29 -0.32 -0.39 -0.35 -0.28 -0.29 -0.32 -0.48 -0.28 -0.37 -0.44
PPD (%) 9.61 6.75 7.13 8.17 7.55 6.63 6.75 7.1 9.81 6.63 7.85 9.04
Operative
20 20.4 20.4 20.2 20 20.3 20.3 20.2 20 20.2 20 20
temperature (°C)
Mean air
17.2 18.4 17.8 17 19.2 19.5 19.2 18.9 18.9 19.7 19.2 18.8
temperature (°C)

587
 Table 10: Minimum heat fluxes from radiant panels for different regions in cooling
Model 1 Model 2 Model 3 Model 4
Region 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
2
Heat flux (W/m ) 54 47 49 44 25.5 23.5 24 21 15.5 15 15 13.5 20.5 17.5 18.5 15.5
PMV 0.25 0.35 0.35 0.37 0.32 0.33 0.44 0.43 0.31 0.29 0.4 0.4 0.36 0.38 0.39 0.37
PPD (%) 6.3 7.55 7.55 7.85 7.13 7.26 9.04 8.86 7 6.75 8.33 8.33 7.7 8.01 8.17 7.85
Operative
25.2 25.6 25.6 25.7 25.6 25.6 25.9 25.9 25.5 25.5 25.8 25.8 25.7 25.8 25.8 25.5
temperature (°C)
Mean air
temperature 27 27 27 26.9 26.1 26.2 26.6 26.4 26 25.8 26.2 26.1 26 26.1 26.1 26
(°C)

5.4. Comparison of radiant and conventional systems under similar thermal conditions
Having compared radiant and conventional heating systems under similar thermal conditions, indoor air
temperature was found to be 1.1°C less in case of radiant heating than of conventional heating because mean
radiant temperature in radiant heating was 1.1°C higher than in convective heating.

Table 11: Comparison of radiant and conventional systems

Conventional
Parameter Radiant heating
heating
Clothing insulation (clo) 1 1
Mean air temperature (°C) 20.9 22
Mean radiant temperature (°C) 22.3 21.2
Input
Metabolic rate (met) 1.2 1.2
Air velocity (m/s) 0.028 0.028
Relative humidity (%) 50 50
Operative temperature (°C) 21.6 21.6
Output PMV 0 0
PPD (%) 5 5

5.5. Annual energy saving in the case of radiant heating and cooling
Table 11 demonstrates annual energy savings using radiant heating system for different regions. There is
remarkable energy saving using radiant system for heating. Energy saving was larger at regions where heating
load was higher and could reach maximum level in the coldest region. The more energy saving could be achieved
using model 3.

Table 12: Annual energy saving use of radiant heating systems

Energy saving (Turkish currency)

Region
Model 1 Model 2 Model 3
1 79.0 79.0 100.7
2 143.2 143.2 179.4
3 178.5 178.5 233.8
4 268.2 268.2 333.7

Annual energy savings are shown for radiant cooling system in Table 12. Model 3 has maximum energy saving,
whereas model 4 has minimum energy saving. Energy saving is larger at regions where cooling load is higher.

Table 13: Annual energy saving use of radiant cooling systems

Energy saving (Turkish currency)

Region
Model 1 Model 2 Model 3 Model 4
1 25.7 30.0 36.4 15.1
2 15.0 15.0 16.7 10.1
3 11.9 12.5 14.5 7.6
4 1.3 1.5 1.6 1.1

588
6. CONCLUSION
Thermal comfort and energy efficiency performance of the novel radiant heating & cooling systems investigated
with CFD method according to different building and heating/cooling system configurations in this study, The
following conclusions have been reached:

• In the use of radiant heating and cooling systems in a model room fulfills general and local thermal
comfort requirements with lower air temperature for heating, higher air temperature for cooling than
conventional system.
• All investigated radiant heating and cooling systems gave acceptable an indoor environment, which has
smaller vertical temperature differences and almost no air movement field
• Heat fluxes from radiant panels required in order the conditions of thermal comfort obtained numerically
for diffferent regions
• There is remarkable energy saving using radiant system for heating, even much more in colder areas in
Turkey. In cooling season energy saving associated with radiant system is lower than it is in heating
season, since there is a smaller difference between values of indoor and outdoor temperature, and
furthermore, cooling period is shorter than its heating counterpart.
• Both in heating and cooling cases, energy saving incrase when panels are placed heat loss occured
either exterior walls

NOMENCLATURE
As surface area
M metabolic rate, W/m2
W effective mechanical power, W/m2
pa water vapour partial, (Pa)
Ta air temperature, °C
fcl clothing surface area factor
Taust area-weighted unheated surface temperature, °C
Tcl clothing surface temperature, °C
To operative temperature
Tp panel surface temperature, °C
Tr mean radiant temperature, °C
Ts surface temperature
T0.1 air temperature at 0.1 m above floor level (°C)
T1.1 air temperature at 1.1 m above floor level (°C)
h convective heat transfer coefficient, W/m2°C
u velocity component in x direction, (m/s)
v velocity component in y direction, (m/s)
w velocity component in z direction, (m/s)
x,y,z cartesian coordinates
p pressure, (Pa)
g acceleration of gravity
Greek Letters
ρ density, kg/m3
α thermal diffusivity coefficient
ε emissivity

589
REFERENCES

[1] IsawaK., Komizo T. And Shukuya M. 2002. Low exergy systems will provide us with the lowest human-body
exergy consumption and thermal comfort. Lowex News, publication of IEA Annex 37, No. 5; July 2002: pp. 8-11
[2] M. Virtanen, M. Ala-Juusela, Increased energy efficiency and improved comfort, Paper No. 1494, 9th REHVA
World Congress- Clima 2007, 10–14 June, Helsinki, Finland.
[3] M. Ala-Juusela, Heating and Cooling with Focus on Increased Energy Efficiency and Improved Comfort,
Guidebook to IEA ECBCS, Annex 37, Low Exergy Systems for Heating and Cooling of Buildings, VTT Technical
Research Centre of Finland, 2003, (http://www.vtt.fi/inf/pdf/tiedotteet/2004/T2256.pdf).
[4] B.W. Olesen, E. Mortensen, J. Thorshauge, Thermal comfort in a room heated by different methods, Technical
Paper no. 2256, Los Angeles Meeting, ASHRAE Transactions 86, 1980.
[5] P.E. Nilsson (Ed.), Achieving the Desired Indoor Climate, IMI Indoor Climate and Studentlitteratur, 2003, ISBN
91-44-03235-8.
[6] E. Hutter, Comparison of different heat emitters in respect of thermal comfort and energy consumption, in:
Proceedings of the International Centre for Heat and Mass Transfer, Heat and Mass Transfer in Building Materials
and Structures, 1991, pp. 753–769.
[7] H.H.E.W. Eijdems, A.C. Boerstra, Low Temperature Heating Systems: Impact on IAQ, Thermal Comfort and
Energy Consumption, Annex 37, Newsletter 1, 2000.
[8]Laouadi, A. (2004). Development of a radiant heating and cooling model for building energy simulation software.
Building and Environment 39: 421-431.
[9] Myhren, J. A. and Holmberg, S. (2008). Flow patterns and thermal comfort in a room with panel, floor and wall
heating. Energy and Buildings 40: 524-536.
[10] Okamoto, S., Kitora, H., Yamaguchi, H. and Oka, T. (2010). A simplified calculation method for estimating
heat flux from ceiling radiant panels. Energy and Buildings 42: 29-33.
[11]Roulet, C. A., Rossy, J. P. and Roulet, Y. (1999). Using large radiant panels for indoor climate conditioning.
Energy and Buildings 30: 121-126.
White, F. M. (2003). Fluid Mechanics. New York:Hemisphere-McGraw-Hill.
ASHRAE Standart 55 (2003). Thermal environment conditions for human occupancy, ASHRAE.
Dudkiewicz, E. and Jezowiecki, J. (2009). Measured radiant thermal fields in industrial spaces served by high
intensity infrared heater. Energy and Buildings 41: 27-35
EN ISO 7730 (2005). Ergonomics of thermal environment-Analytical determination and interpretation of thermal
comfort using calculation of the PMV and PPD indices and local thermal comfort criteria. International Organization
for Standardization, Geneva.
Fluent User’s Guide (1998). Fluent incorporated Centerra Resource park, Lebanon.

590
 EXPERIMENTAL INSULATION PERFORMANCE EVALUATION OF AEROGEL FOR
HOUSEHOLD REFRIGERATORS
GEDİZ İLİŞ G.
Indesit Company R&D Center
Manisa Organize Sanayi Bölgesi 45030 Manisa, Turkey
gamze.gedizilis@indesit.com

ABSTRACT
Energy saving is one of the most important topic for the refrigeration technology. For this reason, insulation of a
household refrigerator should be improved by innovative insulation materials. Aerogel insulation is one of the most
promising high performance thermal insulation materials for different applications nowadays. Within this work, a
different application area of aerogel is studied. Aerogel insulation sheets were applied to a household refrigerator as
an additional insulation material. Aerogel insulation was applied on the evaporators and different surfaces of the two
types of static and no frost types household refrigerators. According to the energy consumption test results, the
application of 10 mm thick aerogel sheet to the outer backside surface of the refrigerator decreases the energy
consumption of the refrigerator down to 5%.

INTRODUCTION
Over the past two decades, energy and energy efficiency are the main important topic around the world and many
researchers and manufacturers devote significant efforts to improve the energy efficiency of their appliances. Due to
the global warming and ozone depletion, many energy efficiency procedures have been written as energy labeling of
the appliances and many standards were written and have been applied to different sectors. ISO 15502 standard is
one of them and used for energy labeling of household refrigerating appliances.

The consumers are interested in the energy efficient appliances and this idea forces the manufacturers to make
developments and investments on energy efficient instruments. Interest in more energy efficient appliances for the
household refrigerators encourage the manufacturers to focus on many applications like modern appliances with
many sensors (Xiao et al. 2010), new compressor designs (Lee et al. 2008), new working fluids (Jung et al. 2008),
additives to the refrigerant (Mahbubul et al. 2013) or the lubricant oil (Bi et al. 2008) etc. Also, the innovative
insulation materials or vacuum insulation panels to decrease the energy consumption of the domestic refrigerators
are another ways to achieve more energy saving domestic refrigerator.

Polyurethane is the base insulation material of a household refrigerator. It is used not only for the insulation but also
for the rigidity of the refrigerator. Due to its long reaction time, expensive application procedure, and construction
constrains, manufacturers look for another insulation alternatives to achieve more energy efficient refrigerators.
Vacuum insulation panels are one of the alternatives on this way (Clodic and Zoughaib 2000). Another alternative can
be the Aerogel insulation sheets.

Aerogels were discovered approximately a hundred year ago (Kistler 1931). Its density is very low (50-200 kgm-3) and
they have significant thermal, acoustic and optical properties (Tajiri and Igarashi 1998, Anderson et al. 2009, Reim et
al. 2004). Aerogels are also used in various areas of science such as vacuum insulation panels, acoustics, and have
significant application area for possible building applications (Beatens et al. 2011, Buratti and Moretti 2012, 2013).
Highly energy efficient windows, translucent aerogel insulation for the buildings are one of the examples for building
applications of the aerogel. Aerogel insulation has distinct advantages due to its ability with low thermal conductivity
(down to 0.010 Wm-1K-1).

Although Aerogel has such a various kinds of application area, almost no application was performed in the household
refrigerator sector. In this paper, Aerogel insulation sheet was applied to the domestic refrigerator in order to decrease
the energy consumption. For this purpose, Aerogel insulation sheet with 10 mm thickness was applied to the
probable critical heat loss areas of the household refrigerator. Energy consumption tests were performed for two
kinds of refrigerator; one is static type cooler refrigerator, other one is the four door no frost type refrigerator with
freezer and cooler. The energy consumptions of the refrigerators with and without Aerogel were compared and
necessary discussions were performed.

THE THERMOPHYSICAL PROPERTIES OF AEROGEL SHEET

In this work, Aspen Aerogels sheet was applied. The Cryogel Z type aerogel sheets with 10 mm thickness were used.
As can be seen from Figure 1, the sheet is flexible with aluminum sheet cover on one side. It is hydrophobic and has
a density of 128 kg/m3. The thermal conductivity of the used aerogel varies with temperature, i.e. λ=138 mWm-1K-1 at
0oC and λ=129 mWm-1K-1 at -50oC. It is non-flammable and non-reactive and not hazardous to human health.

591
Figure 1. The Cryogel Z type aerogel (Aspen Aerogels)

THE TESTED REFRIGERATORS

The tests were performed with two types of refrigerator; one is static only with cooler (R1), other one is no frost type
refrigerator both with cooler and freezer (R2). Both refrigerators work with 230 V-50 Hz and the refrigerant is R600a.
The refrigerant charge of the R1 and R2 is 26 and 50 gr, respectively. The R1 has 341 lt. cooler capacities and has
no freezer. The R2 has 292 lt. cooler with chiller compartment and 110 lt. freezer capacities.

THE AEROGEL APPLICATION PROCEDURE TO THE REFRIGERATORS

Totally five different energy consumption tests were performed with both types of the refrigerators. Figure 2 illustrates
the test numbers and the procedure of the experimental work for these refrigerators. For the static type refrigerator,
two tests were performed;

a) reference refrigerator without aerogel (test#1),

b) aerogel located on the evaporator and the porous side contact with the polyurethane (test#2),

For the no frost type refrigerator three different types of tests with load were performed;

c) reference refrigerator without aerogel (test#3),

d) aerogel porous side contact with the polyurethane (test#4),
e) aerogel porous side not contact with the polyurethane located at the back part of the refrigerator (test#5).

As given in Figure 1, aerogel sheet has two sides, one is covered with aluminum sheet and the other side has porous
structure. Aerogel sheet was located to allow the polyurethane contact with the porous side of the aerogel.

Figure 2. The procedure of the experiments and the test numbers.

Due to lack of freezer for the static type refrigerator, there is no need to load the refrigerator for the EC test (ISO
15502). The evaporator was covered with the aerogel sheet and the energy consumption test, test#2, was performed.
As illustrated in Figure 3, the aerogel porous side is located to allow contacting the polyurethane. The aerogel
application onto the evaporator of the static type refrigerator is illustrated in Figure 3.

592
 (a) (b)
Figure 3. The aerogel sheet application onto the evaporator of the static type refrigerator, a) uncovered evaporator, b) covered evaporator.

The application of the Cryogel Z aerogel sheet to the basement and the backside of the refrigerator for test#4 is
illustrated in Figure 4. The porous side of the sheet is located to be able to contact with the polyurethane.

Figure 4. The aerogel sheet application to the backside and the basement of the no frost type refrigerator (test#4).

The test#5 was performed by covering the aerogel sheet between the back part of the refrigerator and the condenser.
The Figure 5 shows the application of the Cryogel sheet on the refrigerator. The application is performed by covering
the aerogel sheet on to the reference refrigerator.

Figure 5. The aerogel sheet application to the backside of the no frost type refrigerator for test#5.

EXPERIMENTAL PROCEDURE-TEST CHAMBER

The refrigerators were tested according to the standard ISO 15502. The steady stated Energy Consumption (EC) test
procedures were applied. M packages were used for the refrigerator with freezer. For static type refrigerator (R1) the
EC test were performed without M packages due to existence of freezer. For this purpose, the M packages were
placed into the R2. The M package storage plan was given in Figure 6. The thermocouples were instrumented on the
M packages and the M packages placed both in cooler and freezer as highlighted in Figure 6. The tests were held

593
within a climate chamber. The climate chamber temperature can be varied between 10 – 48oC with ±0.2 uncertainty
band. A compressor, a condenser, an expansion device and an evaporator control the test chamber cooling system.
Fans and dampers obtain the air circulation in the chamber and an electrical heater provides the warm air. Many
thermocouples were located into the chamber to provide the homogenous air temperature inside of the chamber and
the chamber temperature was fixed at 25oC. The schematic view of the test chamber is given in Figure 7. The air
chamber velocity should not exceed 0.25 m/s according to the international standards (ISO15502 2005).

Figure 6. No frost four door refrigerator (R2) load plan of the M packages.

Figure 7. The schematic view of the test chamber.

T types thermocouples were located on the refrigerator; 10 for freezer, 4 for chiller, 3 for cooler, 1 for crisper.
Evaporator inlet and outlet, suction line, compressor outlet, condenser inlet and outlet and compressor top point
temperatures are also measured by T type thermocouples. The data that monitored and collected was performed
with the Marel Climatic Chamber program.

RESULTS AND DISCUSSION

In this section the temperature variation analysis of the internal volume and the temperature variation of the
evaporator inlet-outlet of the refrigerator were given. The energy consumption tests and the power analysis of the
refrigerators were obtained and illustrated via tables and figures. As mentioned before, two types of refrigerator were
taken into consideration and five tests were performed at 25oC.

594
TEMPERATURE VARIATION ANALYSIS OF THE INNER TEMPERATURE AND AT THE M PACKAGES

STATIC TYPE REFRIGERATOR

In this section the temperature variation analysis of the internal volume and the inlet and outlet temperature of the
evaporator, crisper, suction line, dryer, compressor outlet, compressor shell, and space temperature variations of R1
for test#1 and test#2 were measured and the collected data were illustrated in Figure 8 and 9.
For the R1 refrigerator, EC tests were performed (test#1-2). During these tests, the variation of temperature for
different locations was also plotted. As can be seen from Figure 9, the inner temperatures of the refrigerator T1, T2,
T3, and crisper temperatures were at 5oC during the EC test. The suction line temperature varies approximately
between 15 and 31oC.

Figure 8. The temperature variations inside and at the important locations of the static type reference refrigerator (test#1).

The second test of the R1 was performed for the same refrigerant charge amount with the reference static type
refrigerator as 26 gr. The test result is illustrated in Figure 9a. After a period that was needed for the stabilization of
the refrigerator to a steady state condition, the test results were plotted for 300 min. and five operation cycles were
observed during this period. As can be seen from Figure 9a, the suction line temperature starts to vary between 19-
33oC. The suction line temperature is higher when compared with the reference refrigerator. Also the inlet and outlet
temperatures of the evaporator are lower when compared with the reference refrigerator, test#1. After this amount of
refrigerator charge, the charge was increased to 29 gr. As can be seen from the Figure 9b, the inlet and outlet
temperatures are lower and the suction line temperature is higher than the reference refrigerator temperatures. This
means that the refrigerant could not be able to vaporized. The charge reduction should be done. Finally, the charge
was reduced to 22 gr as the temperature variation was given in Figure 9c. When compared with the 26 gr and 29 gr
tests, the 22 gr gas charge has convergence with the reference refrigerator.

595
 (a)

(b)

(c)
Figure 9. The temperature variations inside and the important locations of the R1 refrigerator (test#2) with, a)26 gr, b)29 gr, c)22 gr refrigerant charge.

596
NO FROST TYPE REFRIGERATOR
In this section the temperature variation analysis for the no frost type reference refrigerator (R2) were given and
illustrated in Figure 10, 11 for the test#3 and #4, and 12 for the test#5. As seen from Figure 10, the M package
temperatures inside the freezer and the chiller, the temperature variation of the crisper, and the shelve temperature
inside the cooler during the EC test is given. The temperature variations of the M packages inside the freezer and the
chiller are around -18oC and 0oC, respectively. The temperature is measured in the cooler at three different locations
as T1, T2 and T3 as illustrated in the figure. The temperatures of these points vary during the start/stop period of the
compressor. During the whole cycle, the averages of these points are around 5oC. The same temperature variation
behavior with the reference refrigerator is seen for the test#4 in Figure 11. The test#4 was performed and the
temperature variations are also given when the aerogel is applied inside of the polyurethane. The M packages
temperatures are around -18oC and 0oC for the freezer and chiller, respectively. The average inner temperature of the
cooler is around 5oC. The test results for the test#5 were also given in Figure 12 and as expected, approximately the
same temperature variations with test#3 were seen.

Figure 10. The temperature variations inside and the important locations of the no frost type reference refrigerator (test#3).

Figure 11. The temperature variations inside and the important locations of the no frost type R2 refrigerator (test #4).

597
Figure 12. The temperature variations inside and the important locations of the no frost type R2 refrigerator (test #5).

ENERGY CONSUMPTION AND POWER ANALYSIS OF THE REFRIGERATOR

In this section the EC test results for both static and no frost refrigerators and the power analysis of the reference
refrigerators and for test#2, 4, and 5 were plotted and illustrated via figures and tables.

STATIC TYPE REFRIGERATOR

The measured compressor power consumption of static type reference refrigerator and test#2 for the first 100 min is
illustrated in Figure 13. As can be seen from the figure, the compressor on and off time of the test#2 is longer than
the reference test. The compressor cycle time, on-off-on period, and the compressor running time was calculated and
the related results were given in Table 1.

Figure 13. The power of the static type refrigerator, reference refrigerator (test#1) and for test#2 for 100 min.

As mentioned before, the reference refrigerator for static type has 26 gr gas charge and the EC results showed that
for the test#2, 22 gr refrigerant charge is more suitable. The comparison of power, compressor on/off time and the
EC result comparisons are given according to these refrigerant charges. As seen from Table 1, the average inner
cabin temperatures for reference and test#2 refrigerators are 5oC for both and the chamber temperatures for
reference and test#2 refrigerators are 25.2oC and 24.3oC, respectively. Although the compressor off time of test#2
refrigerator is longer than the reference refrigerator, the on time of the compressor is longer, too. For this reason, the
EC result of the aerogel applied R1 refrigerator is 5% higher than the reference refrigerator. Although the refrigerant
charge is decreased, the EC is increased. Otherwise, with 26 gr charge caused liquidification at the suction line and
icing problem is observed on the pipe of the compressor. That’s why the reduction of charge should be necessary
and it affects and reduces the EC of the refrigerator.

598
Table 1. The average cabin and chamber temperature, compressor on/off time and the EC results of reference and aerogel applied R1 refrigerator.

NO FROST TYPE REFRIGERATOR

Figure 14 shows the measured compressor power consumption of no frost refrigerator in reference test (test#3),
test#4, and test#5, for the first 200 min. By this result, the compressor cycle time, on-off period, and the compressor
running time is calculated and presented in Table 2. As clearly seen from Figure 14, the off time of the R2 for test#4
and test#5 is longer than the reference refrigerator. The EC result cannot be predicted from this figure due to the
longer on times of the test#4 and test#5. As given from Table 2, the cabin and the chamber temperatures of all tests
are around 5 and 25oC, respectively. The hottest M packages inside the freezer is around -18oC for all tests. The
test#4 and test#5 show increase in both compressor off and on times when compared with the reference test. If
reference test is compared with test#4, it is clearly seen that, the running ratio of the compressor causes 3% high EC
result. The EC test of test#5 shows that the application of aerogel is not suitable in polyurethane. It gives good results
at the outer surface of the refrigerator and 5% decrease in EC is observed.

Figure 14. The power of the no frost type refrigerator, reference refrigerator (test#3), R2 for test#4, R2 for test#5 for 200 min.

Table 2. The average cabin, M package, and chamber temperature, compressor on/off time and the EC results of reference and aerogel applied
refrigerators.

CONCLUSION
The application of the aerogel sheet to a household refrigerator in order to decrease the energy consumption was
carried out in this study. Two types of refrigerator as static and no frost were held and with these refrigerators, totally
five main energy consumption tests were performed. The cooler, chiller, M packages’ temperature variations, the
compressor on/off time, and the power during the EC tests were measured and collected by using ISO 15502
standard procedures. The test results performed during this study can be concluded that

the application of aerogel:

• can be suitable for the household refrigerators in order to decrease the EC where the maximum heat loss is
dominant for the refrigerator.

599
 • in contact with the polyurethane has no effect on the EC of the refrigerator. 10 mm aerogel is used instead of
10 mm polyurethane and almost no thermal conductivity was changed.
• to the static type refrigerator is not suitable thus the thermodynamic cycle cannot be controlled with an
electronic code or command.
• is suitable when applied on the outer surface of the refrigerator.
• can be applied where polyurethane cannot flow and fill the gap due to the constructional constraints like
small spaces (e.x. water dispenser backside space) that causes heat loss.
• can be applied where the heat bridges occur (e.x. vacuum panels contact edges).

The future works can be held for the heat bridges of the refrigerators where vacuum insulated panels applied to the
refrigerators.

Acknowledgement
The author is grateful to the support given to this work by Mustafa Tavukçuoğlu and Taner Çetinkaya from Ayvaz
Company, for their invaluable effort in obtaining the aerogel samples.

NOMENCLATURE
EC Energy consumption
NF No Frost
R1 Static type refrigerator
R2 NF type refrigerator
T Temperature (oC)
Greek Letters
λ thermal conductivity (W/mK)

REFERENCES
Anderson A.M., C.W. Wattley and M.K. Carroll. 2009. Silica aerogels prepared via rapid supercritical extraction:
effect of process variables on aerogel properties. J Non-Cryst Solids 2:101–108.

Beatens R., B.P. Jelle, and A. Gustavsen. 2011. Aerogel insulation for building applications: A state of the art
review. Energy and Buildings 43:761-769

Bi S., L. Shi, and L. Zhang. 2008. Application of nanoparticles in domestic refrigerators. Applied Thermal
Engineering 28:1834–1843

Buratti C.,and E. Moretti. 2012. Experimental performance evaluation of aerogel glazing systems. Applied Energy
97: 430-437.

Buratti C., and E. Moretti. 2013. Glazing with silica aerogel for energy saving in buildings. Applied Energy 98: 396-
403.

Clodic D., and A. Zoughaib. 2000. Technical and economical evaluation of vacuum insulated panels for a
European freezer. International Refrigeration and Air Conditioning Conference.

ISO15502, Household Refrigerating Appliances-Characteristics and Test Methods, International Organization for
Standardization, Geneva, Switzerland, 2005.

Jung D., C. Kim, K. Song, and B. Park. 2000. Testing of propane/isobutane mixture in domestic refrigerators.
International Journal of Refrigeration. 23: 517-527

Kistler S.S. 1931. Coherent expanded aerogels and jellies. Nature 127-741.

Lee H., S. Ki, S. Jung, and W. Rhee. 2008. The Innovative Green Technology for Refrigerators Development of
Innovative Linear Compressor. International Compressor Engineering Conference.

Mahbubul I.M., S.A. Fadhilah, R. Saidur, K.Y. Leong, and M.A. Amalina. 2013. Thermophysical properties and
heat transfer performance of Al2O3/R-134a nanorefrigerants. International Journal of Heat and Mass Transfer 57:
100–108.

Reim M., G. Reichenauer, W. Körner, J. Manara, M. Arduini-Schuster, and S. Korder et al. 2004. Silica-aerogel
granulate – structural, optical and thermal properties. J Non-Cryst Solids 350:358–363.

Tajiri K, and K. Igarashi. 1998. The effect of the preparation conditions on the optical properties of transparent
silica aerogels. Sol Energy Mater Sol C 4: 189–195.

600
Xiao J, W. Wang, Q.C. Guo, and Y.H. Zhao. 2010. An experimental study of the correlation for predicting the frost
height in applying the photoelectric technology. International Journal of Refrigeration. 33: 1006-10

601
 ECOLOGICAL PERFORMANCE ANALYSIS OF IRREVERSIBLE BRAYTON CYCLE

Yusuf YEĞİNER, Serkan KENÇ, Güven KÖMÜRGÖZ, İbrahim ÖZKOL

*

Istanbul Technical University, Maslak/Istanbul, Turkey

*
e-mail : ozkol@itu.edu.tr
Keywords :Brayton Cycle, Ecological Performance Criterion

ABSTRACT
This paper presents a performance analysis and optimization for irreversible Brayton cycle based on the
Ecological Coefficient of Performance (ECOP) criterion. The ECOP objective function is defined as the ratio
of power output to the loss rate of availability. The maximum of ECOP and performance parameters are
derived analytically under the assumption of optimal design. The graphical representations of the developed
ECOP function in terms of governing parameters are discussed in detail. The effect of dominant parameters
on optimal ecological performances has been investigated. The comparisons of ECOP, dimensionless
ecological function and power output with respect to the dimensionless entropy generation rate and the
enthalpy ratio are presented.

INTRODUCTION
Finite Time Thermodynamics (FTT) Method, providing performance bounds (and extremes/boundaries) for
irreversible thermodynamic processes, can be applied to various engineering fields such as refrigerators,
heat pumps, engines with heat regeneration, several coupled engines, solar engines, and photovoltaic cells.
The Finite Time Thermodynamics can be used to find the optimal time path of any cyclic processes with
friction and heat leakage (all possible irreversibilities) by providing maximization of the work output per cycle.
In irreversible thermodynamics processes, FTT modeling gives more realistic results than those provided by
reversible thermodynamic. By applying FTT processes it is possible to get more physical understanding of
how irreversibilities affect the performance of thermodynamic processes [1-4].

Brayton cycles have been used extensively in gas turbine power plants and aircraft propulsion systems.
Detailed works on Brayton heat engines have made tremendous progress using finite-time thermodynamics
as well as smaller size and higher thermal efficiency heat engine design [5-7]. Detailed information about
optimization- works based on different objective functions for Brayton and the other heat engine models can
be found in literature by Bejan [2], Chen et al. [3] and recently by Durmayaz et al. [8]. During the last decade,
some authors have studied the ecological performance of endoreversible and irreversible heat engines by
considering finite-time and finite-size constraints [9-13]. The ecological performance criteria was first
introduced by Angulo-Brown [9] for finite-time Carnot heat-engines. In addition, it has been found that the
thermal efficiency for maximum ecological function is almost equal to the average of Carnot efficiency
introduced by Curzon and Ahlborn[14]. It has also been derived the general characteristics of the internal
reversible thermal cycles by Angulo-Brown [9]. Later Yan has developed the ecological function found by
Brown [15].

According to the result of these studies, it's observed that the objective function can have negative values
contrary to the reasonable rate of the loss term which is always possible. It can be defined that the difference
between the actual and the theoretical power output is the negative reasonable rate of loss. Ust has
described a new ecological function called ecological coefficient of performance (ECOP) [16]. By this
definition he shows that the objective function must have positive values, as in the thermal coefficient of heat
engines or performance coefficient of refrigerators and heat pumps, in order to determine the effect of the
reasonable rate of loss on the power output. This objective function is called as the coefficient of ecological
performance. It is expressed as power output divided by the reasonable loss rate. Ust et al. have also
performed performance analysis applications for irreversible dual cycle, based on ECOP [16].

In this paper, the thermo-ecological optimization technique introduced by Ust [16] and Ust et al. [17] has
been applied to an irreversible Brayton cycle. A performance analysis of Brayton cycle using a finite-time
ecological optimization technique based on enthalpy approach has not been published yet, to the author's
knowledge. Application of the maximum ecological objective function to a Brayton cycle gives a sense how
the natural resources can effectively be used and how a sense the work output of existence system output.

THE THEORETICAL MODEL

The purpose of this section is to examine a steady-flow irreversible open Brayton cycle with an external heat
source. Let the heat engine cycle (1–2–3–4–1) consists of two constant-pressure processes and two non-

602
isentropic processes (the compression 1–2 and the expansion 3–4). The temperature–entropy diagram and
pressure-volume is shown in Fig. 1.

Fig. 1.Brayton cycle model and T-S diagrams

The cycle operates between two heat reservoirs of temperatures TH and TL . In the cycle, Q̇ H is the rate of
heat transfer from the heat source at the temperature TH to the working fluid in processes 2–3, Q̇ L is the rate
of heat transfer from the working fluid to the heat sink at the temperature TL in process 4–1. Q̇ H and Q̇ L can
be defined as below;

Q̇ H = ṁ (h3 − h2 ) (1)
Q̇ L = ṁ (h3 − h2 ) (2)

Where ṁ is mass flow rate and h2 , h3 are enthalpy values at state 2, 3 respectively. The effectiveness of hot
and cold-side heat exchangers εH and εL for counter-flow heat exchangers are defined as

εH = 1 − e−NH (3)
εL = 1 − e−NL (4)

and the numbers of heat transfer units of hot- and cold-side heat exchangers are

NH = (UH AH )/Ċ w (5)

NL = (UL AL )/Ċ w (6)

Where UH AH is the hot-side heat-exchanger conductance and UL AL is the cold-side heat-exchanger

conductance, Ċ W is the capacity rate of the working fluid. The power produced by the Brayton heat engine
according to the first law of thermodynamics can be given as

Ẇ = Q̇ HT − Q̇ LT = ṁ [(h3 − h2 ) − (h4 − h1 )] (7)

The thermal efficiency becomes,

Ẇ (h −h1 )
η= = 1 − (h4 (8)
Q̇HT 3 −h2 )

The entropy generation rate can be obtained as follows

Q̇ Q ̇
ṁ(h4 −h1 ) ṁ(h3 −h2 )
Ṡ g = LT − HT = − (9)
TL TH TL TH

Enthalpy values at state 3,4 can be expressed in following form

h3 = εH TH Cp + h2 (1 − εH ) (10)

h1 −εL TL Cp
h4 = (11)
1−εL

603
Where Cp is pressure coefficient. The isentropic efficiencies of turbine and compressor are

h2s −h1
ηc = (12)
h2 −h1

h3 −h4
ηT = (13)
h3 −h4s

By using Eqs. (12) and (13)

h2S = (1 − ηc )h1 + ηc h2 (14)

h4 1
h4S = + �1 − � h3 (15)
ηT ηT

h2s h3
ϕ= = (16)
h1 h4s

Where ϕ is the enthalpy ratio of the Brayton cycle.By using Eqs.(14)–(16), enthalpy values can be obtained
as

h1 [ϕ−(1−ηC )]
h2 = (17)
ηC
h1 [ϕ−(1−ηc )]
h3 = εH TH Cp + � � (1 − εH ) (18)
ηk

By substituting Eqs. (10), (11), (14), (15), (17) and (18) into Eq.(16)
a1 ϕ+a2
h1 = (19)
a3 ϕ2 +a4 ϕ+a5

Where the simplification parameters are defined as

a1 = ηc Cp [εL TL − (ηT − 1)(1 − εL )εH TH ] (20)

a 2 = ηC ηT (1 − εL )εH TH Cp (21)

a 3 = (ηT − 1)(1 − εL )(1 − εH ) (22)

a 4 = ηC − (ηT − 1)(1 − εL )(1 − εH ) (23)

a 5 = ηT (ηT − 1)(1 − εL )(1 − εH ) (24)

The ECOP (ecological coefficient of performance) objective function defined as the ratio of power output to
the loss rate of availability can be written as

Ẇ
ECOP = (25)
T0 Ṡg

Where T0 is the environment temperature. Using Eqs. (7) and (9) in Eq. (25), the ECOP function is derived
as

Ẇ ṁ[(h3 −h2 )−(h4 −h1 )]

ECOP = = ṁ(h4 −h1 ) ṁ(h3 −h2 ) (26)
T0 Ṡg T0 � − �
TL TH

Finally, ECOP function can be reduced as simple algebraic relation,

b1 ϕ2 +b2 ϕ+b3
ECOP = (27)
b4 ϕ2 +b5 ϕ+b6

Where;

b1 = a 3 a6 + a1 a 7 (28)

604
b2 = a 4 a 6 + a1 a 8 + a2 a 7 (29)
b3 = a 5 a 6 + a2 a 8 (30)
b4 = a 3 a 9 + a1 a10 (31)
b5 = a 4 a 9 + a1 a11 + a2 a10 (32)
b6 = a 5 a 9 + a2 a11 (33)

Then the defined parameters a 6 a7 a8 a 9 a10 and a11 for simplicity:

a 6 = ηC [TH TL εH TH Cp (1 − εL ) + TH TL εL TL Cp ] (34)
a 7 = TH TL (−εH )(1 − εL ) (35)
a 8 = −(1 − ηC )TH TL (−εH )(1 − εL ) − ηC TH TL εL (36)
a 9 = −ηC �TH T0 εL TL Cp + T0 TL εH TH Cp (1 − εL )� (37)
a10 = −T0 TL (−εH )(1 − εL ) (38)
a11 = ηC T0 TH εL + (1 − ηC )T0 TL (−εL )(1 − εL ) (39)

RESULTS AND DISCUSSION

In this section, analytical results are presented and compared to illustrate the application of thermo-
ecological optimization technique for Brayton cycle. The variation of ECOP functions with respect to the
enthalpy ratio (ϕ) and the dimensionless power output (W �̇ = Ẇ /(Ċ T ))for different values of isentropic
w L
efficiencies (ηC ,ηT ), heat sources temperature ratio (τ = TH /TL ) and the total number of heat transfer units
(NT = NH + NL ) are demonstrated in Figs.2-3. Variation of the ECOP function, the ecologic performance
function, work output are presented graphically and discussed in detail. All the graphs were constructed for
the values of variables (TL ,TH ,εL ,εH ) and given in Table 1.

Table.1. The value of the variables

Low temperature heat source (TL ) 300 K
High temperature heat source (TH ) 1200 K
Low temperature heat exchanger effectiveness (εL ) 0.9
High temperature heat exchanger effectiveness (εH ) 0.9

It can easily be noticed from the Fig. 2(a)-(c) that the ECOP function has a maximum for a certain ϕ value for
the chosen set of operation parameters. The maximum ECOP value is getting higher for the higher heat
sources temperature ratio (τ), the total number of heat transfer units (NT = NH + NL ) and isentropic
efficiencies (ηC ,ηT ).

The maximum ECOP with respect to ϕ can be found analytically by setting d(ECOP)/dϕ= 0. The optimum
value ϕ at the maximum ECOP is found as below;

b4 b3 −b1 b6 (b2 b6 −b5 b3 )(b1 b5 −b4 b2 )

ϕopt = × �1 − �1 − (b4 b3 −b1 b6 )2
� (40)
b1 b5 −b4 b2

The maximum of ECOP can be obtained by substituting Eq. (40) into Eq. (27), i.e.

b1 ϕopt 2 +b2 ϕopt +b3

ECOPmax = (41)
b4 ϕopt 2 +b5 ϕopt +b6

As heat sources temperature ratio (τ), the total number of heat transfer units (NT = NH + NL ) and isentropic
efficiencies (ηC ,ηT ) increase the performances in terms of ECOP and Ẇ increase too. The maximum power
output (Ẇ max ) is higher than the power output at the maximum ECOP conditions (Ẇ ∗ ), as expected, but the
ECOPmax is higher than the ECOP at Ẇ max conditions (ECOPmp ). By taking into account the power and entropy
generation rate (loss rate of availability) of the heat engine together the optimal design intervals in terms of
ECOP and Ẇ yields:

ECOPmax ≥ ECOP ≥ ECOPmp or Ẇ ∗ ≥ Ẇ ≥ Ẇ max (42)

605
 (a)

(b)

(c)
Fig. 2. Variation of the ECOP function with respect to the enthalpy ratio for various values (a) τ (b) NT (c)
ηC = ηT (TL = 300 K, TH =1200 K, ηc =0.9, ηt =0.95, εL = εH =0.9)

606
 (a)

(b)

(c)

Fig.3. Variation of the ECOP function with respect to the power output for various values (a) τ (b) NT (c)
ηC = ηT (TL = 300 K, TH =1200 K, ηc =0.9, ηt =0.95, εL = εH =0.9)

607
To make a correct decision that which optimal design parameters of a heat engine will be important one
should take into consideration the followings; If the power is the substantial interest then the design
parameters must be selected near to the Wmax conditions or if the aim is to keep the entropy generation
minimum design parameters must be chosen to make ECOP maximum. The maximum of thermal efficiency
is an important parameter due to engineering viewpoint. For the given temperatures, ECOP and thermal
efficiency are dependent. In addition to this, maximum of them overlaps, despite the fact that they have
distinctive physical bases. Obtaining the identical performance conditions at the maximum ECOP and
thermal efficiency are a foreseen and reasonable conclusion. For the reason of the fact that the maximum
thermal efficiency conditions for a definite power results in minimum fuel consuming that leading to minimum
environmental destruction. Consequently, the ECOP takes into account environmental influence properly. As
ECOP provides knowledge about the entropy generation (the loss rate of availability), thermal efficiency
provide knowledge about the essential fuel consumption to provide certain power.

Similar to Eq. 40-41, it is possible to obtain the optimum value of power at the maximum ECOP,

Ẇopt b1 ϕopt 2 +b2 ϕopt +b3

= (43)
Ċw b4 ϕopt 2 +b5 ϕopt +b6

The comparisons of ECOP, dimensionless ecological function (E �̇ = Ė /(Ċ T )) proposed by Angulo-Brown [9]
W L
�̇ ) with respect to the dimensionless entropy generation rate (S�̇ ), and the enthalpy ratio
and power output (W g
(ϕ), are demonstrated in Fig. 4(a) and (b). The optimum enthalpy ratio (ϕ∗ ) at the (ECOPmax ) conditions is
always greater than enthalpy ratio (ϕmef ) at the maximum of ecological function (Ė max ) and (ϕmp ) at the
maximum power output (Ẇ max ) conditions in Fig. 4(a). Limits can be given as follows;

ϕ∗ ≥ ϕmef ≥ ϕmp (44)

∗
It can be observed from Fig.4(b) that the entropy generation rate (Ṡ g ) at the ECOPmax conditions is lower than
the entropy generation rate (Ṡ g,mp ) at Ẇ max conditions and (Ṡ g,mef ) at the Ė max conditions. In order to minimize
the entropy generation rate, limits can be described as below;
∗
Ṡ g < Ṡ g,mef < Ṡ g,mp (45)

608
 (a)

(b)

Fig.4. Variations of the ECOP, dimensionless power output and the ecologic performance functions with
respect to the (a) enthalpy ratio and (b) dimensionless entropy generation rate (ηC =0.9, ηT =0.95)

CONCLUSION
A performance analysis was carried out for an irreversible Brayton heat engine based on the thermo-
ecological objective function, that is, the Ecological Coefficient of Performance (ECOP) is defined as the ratio
of power to the loss rate of availability. In the analysis, the maximum of ECOP and performance parameters
are derived analytically under the assumption of optimal design and presented graphically. For the
environmental view, the design parameters based on the maximum of ECOP conditions can be used as a
measure of the compromise between the power output and the entropy generation. The paramount
importance of this study is to provide guidance for optimal design in terms of entropy generation rate, power
and thermal efficiency of Brayton cycles.

NOMENCLATURE
2
A area, m
−1
Ċ w product of mass flow rate and specific heat capacity, = ṁ Cp , kW·K
Ė ecological performance function
ECOP ecological coefficient of performance
h enthalpy, Cal
−1
ṁ mass flow rate of the working fluid, kg·s
NT total number of heat transfer units
Q̇ rate of heat transfer, kW
−1
S entropy, kJ·K
T temperature, K
−2 −1
U overall heat transfer coefficient . kW·m ·K
Ẇ power output, kW
Greek Letters
ε heat exchanger effectiveness
ϕ enthalpy ratio
η thermal efficiency
τ heat sources temperature ratio, = TH/TL
Subscripts
c compressor
g generation

609
H high temperature heat source
L low temperature heat source
max maximum
mef maximum ecological function conditions
mp maximum power conditions
opt optimum
T turbine
0 environment conditions

Superscripts
* maximum ECOP conditions
− dimensionless

REFERENCES
[1] C.Y. Cheng, C.K. Chen, The ecological optimization of an irreversible Carnot heat engine, J. Phys. D:
Appl. Phys. 30 (1997) 1602–1609.
[2] Bejan, A. (1996) Entropy Generation Minimization: The New Thermodynamics of Finite-Size Devices
and FiniteTime Processes, Journal of Applied Physics, 79 (3):1191–1218.
[3] Chen L, Wu C, Sun F., “Finite-time thermodynamic optimization or entropy-generation minimization
of energy systems”, J Non-Equilibrium Thermodynamic, 1999; 24(4):327–59.
[4] Chen L, Sun F., Advances in finite-time thermodynamics: analysis and optimization, New York:Nova
Science Publishers; 2004.
[5] Chen, L., Zheng, J., Sun, F., Wu, C. (2002b) Performance Comparison of an Irreversible Closed
Brayton CycleUnder Maximum Power Density and Maximum Power Conditions, Exergy, an
International Journal, 2: 345-351.
[6] Kaushik, S. C., and Tyagi, S. K., 2002, "Finite time thermodynamic analysis of a nonisentropic
regenerative Brayton heat engine," Int. J. Sol. Energy, 22, pp. 141–151.
[7] Chen L., Sun F., Wu C.: Theoretical analysis of the performance of a regenerated closed Brayton
cycle with internal irreversibilities. Energy Conversion and Management 18(1997), 9, 871-877.
[8] A. Durmayaz, O.S. Sogut, B. Sahin, H. Yavuz, Optimization of thermal systems based on finite-time
thermodynamics and thermoeconomics, Progr. Energy Combust. Sci. 30 (2004) 175–217.
[9] F. Angulo-Brown, An ecological optimization criterion for finite-time heat engines, J.Appl. Phys. 69
(1991) 7465.
[10] Z. Yan, Comment on ‘An ecological optimization criterion for finite time heat engines’, J. Appl. Phys.
73 (1993) 3583.
[11] F. Angulo-Brown, R. Paez-Hernandez, Endoreversible thermal cycle with a nonlinear heat transfer
law, J. Appl. Phys. 74 (1993) 2216–2219.
[12] C.Y. Cheng, C.K. Chen, Ecological optimization of an endoreversibleBrayton cycle, Energy Convers.
Management 39 (1998) 33.
[13] C.Y. Cheng, C.K. Chen, Ecological optimization of an irreversible Brayton heat engine, J. Phys. D:
Appl. Phys. 32 (1999) 350–357.
[14] F.L. Curzon, B. Ahlborn, Efficiency of a Carnot engine at maximum power output, Amer. J. Phys. 43
(1975) 22–24.
[15] Yan, Z. J., 1993: Comment on An ecological optimization criterion for finite-time heat engines, J.
Appl. Phys., 73, 3583
[16] Y. Ust, Ecological performance analysis and optimization of power generation systems, PhD thesis,
Progress Report, Yildiz Technical University, Turkey, 2004 (in Turkish).
[17] Y. Ust, B. Sahin, O.S. Sogut, Performance analysis and optimization of an irreversible dual cycle
based on ecological coefficient of performance criterion, Appl. Energy (2005), in press.

610
 SHORT AND LONG-TERM SOLAR RADIATION ESTIMATION METHOD

Güçlü Y.S. , Dabanlı I. ,Şişman E.

a a b

a
Istanbul Technical University, Faculty of Civil Engineering, Department of Hydraulics, Maslak 34469, Istanbul, Turkey,
gucluya@itu.edu.tr
b
Yildiz Technical University, Faculty of Civil Engineering, Department of Hydraulics, Davutpasa 34220, Turkey,
esisman@erdemir.com.tr

ABSTRACT
In this study, Angstrom equation solar radiation estimation values corresponding to the sunshine durations are
achieved by a linear equation, which is an improvement over the classical model. The results turn out to be better
than the classical Angstrom equation. In the suggested equation, 1 to 10 days’ solar irradiation values are
assumed to affect the present-day values as short-term estimates. The proposed model for summer months
(June-July-August) has provided estimations that are in close agreement with observations. Both long- and short-
term estimations have been obtained rather successfully. In the proposed methodology there are two extra factors,
as dependence coefficients. Finally, the classical Angstrom equation remains as a special case of the model.

Key Words: Solar Radiation, Sunshine Duration, Angstrom Equation, Short-Term Estimate, Dependence

INTRODUCTION
th
Especially in the 20 Century and during its second half energy crisis has been experienced due to extreme
consumption of fossil fuels. As a result, people began to search for alternative energy sources. Solar energy
is one of the clean energy sources. To benefit from solar energy as the most important energy source in the
future, present calculation and estimation of the radiation amounts have been a problem waiting for solution.
In order to determine the potential of the solar energy for an area, solar radiation values are prerequisites.
Thus, the decisions of whether or not to invest in a region should be given by the potential of solar energy
plan possibilities in the area.
Angstrom (1924) made the first study on this subject and proposed a linear equation that expresses the
relationship between the sunshine duration and amount of solar radiation. Later, Page (1964) changed the
definition of an expression by using radiation on the horizontal area out of the atmosphere instead of clear
day radiation. Similarly, many researchers have demonstrated studies linearly (Swartman and Ogunlade,
1967; Sabbagh et al., 1977; Rietveld, 1978; Dogniaux and Llemoine, 1983; Gopinathan, 1988; Lewis, 1989;
Soler, 1991; Wahab, 1993; Hinrichsen, 1994; and Sen, 2001). In addition to these studies, Şahin and Şen
(1998) brought dynamism to the Angstrom equation. Some of the scientists have suggested approaches by
proposing the second order equations. Ögelman et al (1984) stated that second order equation better
expresses the relationship between solar radiation and sunshine duration. Akınoğlu and Ecevit (1990) and
Yıldız and Öz (1994) studied the same problem by using second order equation too. Samuel (1991)
proposed a non-linear equation in third order. Şen (2007) contributed to the non-linear model estimates of
the solar radiation. Additionally, Şen (1998) has estimated solar radiation using fuzzy logic modeling
techniques. The aforementioned studies, respectively, are linear and non-linear, with no dependencies

611
between successive values. However, the solution suggested in this article takes into consideration the
linearity with dependencies.
The purpose of this article is to show the present day’s solar radiation and sunshine duration depending on
the previous days’ values and to provide new equation that yields better results than Angstrom approach. In
the equation, two important dependency coefficients are implemented. This method reduces to the classical
Angstrom equation when the dependencies are equal to zero.

METHODOLOGY
The following linear Angstrom equation that estimates solar radiation is the starting point of the present
study.

 H   S 
 =a+b  (1)
 H0   S0 
H, Solar Radiation, H0, Monthly Mean Solar Radiation, S, Sunshine Duration, S0, Monthly Mean Sunshine
Duration, a and b are coefficients.
To refine this equation, the time-effect is considered in other words, today's sunshine duration and solar
radiation are thought of dependent on previous or earlier days’ sunshine duration and solar radiation,
amounts. In equation 2, additionally, t, day, c and d coefficients.

 H   S   S    H 
  =a+b   − d    + c  (2)
 H0 t  S 0 t  S 0  t −i   H 0  t −i

Where 'i' value ranges from 1 to 10 for short-term estimates. However, for some stations, it can be further
extended to include the range (t-1, t-2, t-5, t-10, t-15, t-30, t-60, etc.). Long-term estimates are suitable for
the summer months from t-1 to t-731.
In the first stage of creating the equation, c and d dependency coefficients are obtained, where c is the solar
radiation dependency coefficient and d is sunshine duration dependency coefficient.
To obtain the solar radiation dependency coefficient c,

 H 
 
 H0  t −i
is plotted against

 H 
  ,
 H0 t
where the slope of the drawn regression line yields c coefficient. Similarly, to obtain the sunshine duration
dependency coefficient d,

 S 
 
 S0  t −i

612
is plotted versus

 S 
  ,
 S0 t

where the slope of the drawn regression line gives coefficient d.

Finally, a and b coefficients are determined similar to the Angstrom case. For this purpose, the following
equations are substituted into Equation (2), then into Equation (5), which yields,

 S   S 
X =   − d   (3)
 S0 t  S 0  t −i

 H   H 
Y =   − c  (4)
 H 0 t  H 0  t −i

Y = a + bX (5)

To obtain coefficients a and b, X values are pointed on the horizontal axis and Y values on the vertical axis.
The slope of the drawn regression line provides coefficient b and the intersection point on the vertical axis
gives the coefficient, a. New version and the classical Angstrom equation are now available for estimating
the solar radiation.

APPLICATION
Daily data used in this study are taken from DMI and measured in Antakya, Turkey from 2000 to 2010. For
Antakya a, b, c, and d coefficients are obtained by using approximately half of the data and then the rest of
the data are used for validation. The same procedure is applied by using Angstrom equation. In addition, for
each t-i, Angstrom equation is re-determined. First, coefficients are determined according to t-1.
In Figure 1, the horizontal values are solar radiations for t-1 day; the vertical values are solar radiations for t
day and then regression line is drawn. Coefficient c is determined as the slope of the regression line, as
0,541.

613
 2000-2004
0,60

0,50

0,40
t H/H0

0,30 y = 0,541x + 0,18

0,20

0,10

0,00
0,00 0,10 0,20 0,30 0,40 0,50 0,60
t-1 H/H0

Figure 1 – Lag-one scatter diagram of solar radiation for c coefficient

In Figure 2, the horizontal values are sunshine durations for t-1 day; the vertical values are sunshine
durations for t day and then regression line is drawn. Coefficient d is determined by the slope of the
regression line as 0,591.

2000-2004
1,00
0,90
0,80
0,70
0,60
t S/S0

0,50 y = 0,591x + 0,247

0,40
0,30
0,20
0,10
0,00
0,00 0,20 0,40 0,60 0,80 1,00
t-1 S/S0

Figure 2 – Lag-one scatter diagram of sunshine duration for d coefficient

614
In Figure 3, the horizontal values are X’s, the vertical values are Y’s and hence the regression line is drawn.
Coefficient b is determined by the slope of the regression line as 0,321 and the intercept point on the vertical
axis yields the coefficient a as 0.1.

2000-2004
0,60

0,50

0,40

0,30
Y=(H/H0)t-c*(H/H0)t-1

0,20
y = 0,321x + 0,100
0,10

0,00
-0,75 -0,50 -0,25 0,00 0,25 0,50 0,75 1,00
-0,10

-0,20

-0,30
X=(S/S0)t - d*(S/S0)t-1

Figure 3 – Scatter diagram for a and b coefficients of the proposed method

In Figure 4, the horizontal values are sunshine durations for t day, whereas the vertical values are solar
radiations for t day and then the regression line can be drawn. The coefficient b is then determined by the
slope of the regression line as 0,329 and the intercept point on the vertical axis gives the coefficient as
0,193.

615
 2000-2004
0,60

0,50

0,40
t H/H0

0,30 y = 0,329x + 0,193

0,20

0,10

0,00
0,00 0,20 0,40 0,60 0,80 1,00
t S/S0

Figure 4 – Scatter diagram for classical Angstrom coefficients

After that verification process begins. In this stage, solar radiation is estimated at time t. In Figure 5, the
horizontal values are estimations with Angstrom (red) and the new equation (blue), whereas the vertical
values are measurements. It is obvious that the new estimation straight line is closer to measurements line
than Angstrom straight line.

Estimations and Measurements 2005-2008

0,80

0,70

0,60
Measurements

0,50 new meth. est.

y = 1,112x - 0,035
0,40
angst. est.
0,30 y = 1,143x - 0,034

0,20

0,10

0,00
0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80
Estimations

Figure 5 – Scatter diagram for Angstrom and the proposed methods comparison

616
The daily error amounts are determined as follows;

2
Daily Error = (Measurement–Estimation)

The total errors of the two methods are compared. After estimation of the solar radiation for all days from
2005 to 2008, the dependency model as developed in the article has a total error equal to 2,37 but the
Angstrom model has a total error of 3,14 for t-1. Therefore, Angstrom model has been more defective than
the model of dependency as suggested in this study. The comparison of square root errors between the two
methods yields as follows 13%,
(√3,14 − √2,37) / √3,14 = 0,13 (13%).
In the same way, the above steps are applied for times instances t-2, t-3, t-4, t-5, t-6, t-7, t-8, t-9 and t-10 but
still Angstrom model appeared more defective than dependency model. Error ratios are respectively as
follows; 9%, 8%, 7%, 7%, 6,5%, 6%, 7%, 7% and 7%. According to measuring station’s compatibility, t-i
values can be t-15, t-30 or t-60.
Proposed new model in the article has estimates very close to the measurements in June, July and August.
However, the short-term estimation model during the summer solar energy production has very high
estimates of the solar radiation with error almost zero with the highest estimation error in summer months.
This is of great importance for energy planning.
In case of t-1, after estimating of solar radiation for all summer days from 2005 to 2008, the dependency
model has a total error of 0,2226 but the Angstrom model has a total error equal to 0,5651. Therefore,
Angstrom model has been more defective with 37% than the model of dependency. It seems that the new
model has good estimation for summer days, so in the cases of t-366 and t-731 the estimates are obtained.
In case of t-366, after estimation of the solar radiation for all summer days from 2007 to 2010, the
dependency model has a total error; 0,2407 but the Angstrom model has a total error of 0,3097. Therefore,
the Angstrom model has been more defective with 12% than the model of dependency. Proposed model
also has the ability for the long-term estimations during the summer months. However, when the time
distance increases from t onwards, then the dependence reduces.

CONCLUSION
In this study, the Angstrom equation is improved significantly. The new equation represents additional c and
d dependency coefficients the values of which are multiplied by previous or earlier days’ values of radiation
as well as the time. The days before the sun are thought to be dependent due to the solar radiation and
sunshine duration dependence on previous days. The new equation has estimates better than estimates of
the Angstrom equation. Previous day’s solar radiation and sunshine duration values enter into the
calculations in the suggested method. Dependency model brings well short-term estimates, in particularly in
the event of loss of short-term solar radiation values. As noted, before maximum 10 days, short-term
estimate can be made because after 10 days, dependency decreases but error increases. However, as
mentioned in the application part, new model has less error especially in the summer months. So that during
the summer months dependency model is compared to the Angstrom model in case of short-term
estimations and as a result, the Angstrom model has more error, which is almost 40%. Additionally, during

617
the summer months dependency model is also compared to the Angstrom model in case of long-term
estimations and hence it is found that the Angstrom model has more error at about 12%. If i values in t-i
increase than time error increases too, because of that dependency decrease. In summary, for one year
difference, the new model estimations have less error during the summer months.
In the case of no dependencies, the proposed model reduces to the classical Angstrom equation.

REFERECES
Akınoğlu B.G., and Ecevit, A. (1990) Construction of a quadratic model using modified Angström coefficients
to estimate global solar radiation, Solar Energy, 45:85–92

Angström, A. (1924) Solar terrestrial radiation, Q J R Meteorol. Soc., 50:121–126

Rietveld, M. R. (1978) A new method for estimating the regression coefficients in the Formula relating solar
radiation to sunshine, Agric. Meteorol., 19:243–252

Şahin, A. D., and Şen, Z. (1998) Statistical analysis of the Angström formula coefficients and application for
Turkey, Solar Energy, 62:29–38

Gopinathan, K. K. (1988) A general formula for computing the coefficients of the correlation connecting
global solar radiation to sunshine duration, Solar Energy, 41:499–502

Wahab, A. M. (1993) New approach to estimate Angström coefficients, Solar Energy, 51:241–245

Hinrichsen, K. (1994) The Angström formula with coefficients having a physical meaning, Solar Energy,
52:491–495

Şen, Z. (2001) Angström equation parameter estimation by unrestricted method, Solar Energy, 71:95–107

Ögelman, H., Ecevit A., and Taşemiroğlu, E. (1984) Method for estimating solar radiation from bright
sunshine data, Solar Energy, 33:619–625

Şen, Z. (2007) Simple nonlinear solar irradiation estimation model, Renewable Energy, 32:342–350

Şen, Z. (1998) Fuzzy algorithm for estimation of solar irradiation from sunshine duration, Solar Energy,
63:39–49

Page, J. K. (1964) The estimation of monthly ea values of daily total short wave radiation on vertical and
inclined surfaces from sunshine records for latitudes 40°N–40°S, Proceedings of the UN conference on new
sources of energy, paper no. 598, 4, pp 378–390

Swartman, R. K., and Ogunlade, O. (1967) Solar Radiation Estimates from Common Parameters, Solar
Energy, 11, 170-172

Sabbagh, J. A., Sayigh, A. A. M., and El-Salam, E. M. A. (1977) Estimation of the total solar radiation from
meteorological data, Solar Energy, 19: 307-311.

Samuel T. D. M. A. (1991) Estimation of Global Radiation for Sri Lanka, Solar Energy, 45,5,333

Dogniaux, R., and Lemoine, M. (1983) Classification of radiation sites in terms of different indices of
atmospheric transparency. In Palz W. (éditeur), Solar Energy R&D in the European Community, Series F,
Vol. 2, Solar Energy Data. D. Reidel Publ. Co., Dordecht, 94-107.

Lewis, G. (1989) The Utility of the Angström –Type Equation fort he Estimation of Global Radiation, Solar
Energy, 43,5,297-299,

Soler, A. (1990a) Monthly specific Rietveld’s Correlations. Solar and Wind Technology, 2/3, 305-306

618
Yıldız, M., and Öz, S. (1994) Evaluation of the solar energy potential of Turkey. In: Proceedings of the 6th
National Energy Congress;. p. 250–60 [in Turkish].

619
 ENERGY AND EXERGY ANALYSES OF COPPER-CHLORINE (Cu-Cl) BASED INTEGRATED
SYSTEMS FOR HYDROGEN PRODUCTION

A. Ozbilen, I. Dincer and M. A. Rosen

Faculty of Engineering and Applied Science
University of Ontario Institute of Technology
2000 Simcoe Street North, Oshawa, Ontario, L1H 7K4, Canada
Ahmet.Ozbilen@uoit.ca, Ibrahim.Dincer@uoit.ca, Marc.Rosen@uoit.ca

ABSTRACT
Hydrogen production via thermochemical water splitting using the Cu-Cl cycle is a promising alternative to
conventional hydrogen production methods. Multi-generation systems are often attractive due to their higher
energy and exergy efficiencies than individual cycles (e.g. steam or gas turbine cycles). An increase in
efficiency often allows, for the same output, less resources (or exergy) to be used. This, in return, reduces the
extraction from the environment of energy resources (e.g. fossil fuels), and decreases the associated
environmental impacts. In this study, comprehensive thermodynamic analyses of two Cu-Cl based integrated
systems for hydrogen production are reported. The systems considered here are developed to produce
hydrogen and provide cooling. Hot water, drying air and oxygen are also by-products of the system. The first
system also has the capability of generating electricity using a two-stage steam turbine cycle. The main sub-
units of the system are the Cu-Cl thermochemical water splitting cycle and a LiBr-H2O absorption cooling
system. Solar energy drives the first system using a heliostat solar tower, whereas waste/process heat from a
Generation IV nuclear super-critical water cooled reactor (SCWR) is the energy source of the second system.
The energy efficiency of the first system is found to be approximately 70%, while the exergy efficiency is
approximately 58%. The second system, on the other hand, has an energy efficiency of 63% and an exergy
efficiency of 41%.

Keywords: Hydrogen production, thermochemical water splitting, Cu-Cl cycle, exergy, multi-generation.

1. INTRODUCTION
Global energy demand tends to increase due to increasing living standards and a growing world population.
Concerns about the energy supply security are increasing due to reductions in supplies of fossil fuel resources
which increases energy carrier prices. These resources contribute to air pollution and global climate change.
Petroleum is a central concern, with a share of more than one third of global primary energy consumption and
more than 95% of the energy consumption in the transport sector. The negative environmental impact of coal
mining and use combined with the large contribution of coal usage to global carbon dioxide emissions, as well
as potential future risks of the declining reserves of natural gas, are other factors of concern (Urbaniec et al.,
2010).
Hence, alternatives to fossil fuels have been sought. Increases in energy demand will likely lead to a
growth in nuclear and renewable energy utilization, partly to meet the objective of sustainability. The shift from
fossil fuels to nuclear and renewable resources is expected due to the increase in energy demand as well as
concerns over environmental issues such as global warming. In the future, energy systems are expected to be
hybrid systems when it is economically feasible. Hybrid systems combine various energy resources and energy
conversion methods to help increase efficiency while reducing wastes and associated environmental impact.
Hydrogen is a promising candidate as an energy carrier that helps expand markets for renewable and nuclear
energy resources and contributes to sustainability and environmental stewardship, and that can act as a link
between these technologies when they are utilized in hybrid systems (Dincer, 2012; Rosen, 2010). Integrated
energy conversion systems can help fulfill the increasing need for energy sustainability.
Hydrogen exists in abundance in nature in the form of water. However, pure hydrogen needs to be
produced and there are several methods of achieving this including steam reforming of natural gas, coal
gasification, water electrolysis and thermochemical water decomposition. Dufour et al. (2009) indicate that 96%
of world’s hydrogen is produced using fossil fuels, and steam reforming of natural gas is the most commonly
used method.
Hydrogen production using thermochemical water splitting cycles has the potential to be cleaner and
more cost-effective than other production methods. Although hydrogen production systems using
thermochemical cycles have not yet been commercialized, studies have shown that such systems can be
expected to compete with conventional H2 production methods including steam methane reforming (Lewis et
al., 2009; Pilavachi et al., 2009). Due to its lower temperature requirements (around 530°C), the Cu-Cl
thermochemical water decomposition cycle has some advantages over other cycles (Naterer et al., 2013).
Multi-generation systems are often attractive due to their higher energy and exergy efficiencies than
individual cycles (e.g. steam or gas turbine cycles). An increase in efficiency often allows, for the same output,
less resources (or exergy) to be used. This, in return, reduces the extraction from the environment of energy
resources (e.g. fossil fuels), and decreases the associated environmental impacts. The objective of this study is

620
to conduct a comprehensive thermodynamic analysis of a novel Cu-Cl based integrated system for multi-
generation, to improve understanding of the system and its potential applications.

2. SYSTEM DESCRIPTION
The systems discussed here produce hydrogen as the main output. Product hydrogen is considered to be
exported as a commodity for industry and/or as a fuel. All systems also have the capability of hydrogen storage
for energy management. Thus, the produced hydrogen is always a product; and when energy management is
needed, some of the hydrogen will be stored and converted to electricity using fuel cells. Oxygen as an output
of the Cu-Cl cycle is also treated as byproduct which is sufficiently pure for use or sale (Naterer et al., 2008). All
the systems also have the capability of providing cooling, hot water and drying air. A LiBr-H2O absorption
refrigeration system is considered in the study for both systems which uses the excess energy of solar/nuclear
heat transfer fluid to obtain cooling effect. The absorption cooling system (ACS) is used instead of a
conventional cooling system to utilize surplus heat. A steam turbine cycle (STC) is included in System I to
generate power. Further details on the ACS and STC are presented elsewhere (Gebreslassie et al., 2010 and
Nezammahalleh et al., 2010). Hot water at 42°C is supplied to a community by the integrated system. Also,
drying air is obtained by heating ambient air to drying temperature (50°C) using excess heat of the exhaust
gases. Drying air can be used in many applications, mainly drying fruits and vegetables.

2.1 System I – Cu-Cl Based Multi-generation System Using Solar Energy

Solar thermal energy, concentrated using a heliostat solar tower, is the energy source of System I (Fig. 1).
Molten salt (which has composition of 60% NaNO3 and 40% KNO3, on a mass basis) is considered as the heat
transfer fluid (HTF) to supply heat to the Cu-Cl cycle. Molten salt has an advantage in that the solar heat can be
stored for tens of hours for use at night, or when sunlight is not available (Wang et al., 2011). First, heat is
supplied to the copper oxychloride decomposition step (step 4) since this step has the highest temperature
heat requirement (530°C) in the cycle. Second, the heat is transferred to the hydrolysis step (step 3) of the Cu-
Cl cycle and then hydrogen production (step 1) and drying (step 2). The temperature of the molten salt is
increased to 650°C, so as to match the heat requirements of the Cu-Cl cycle (Wang et al., 2011). The
temperature of the molten salt in a low temperature storage tank is higher than 250°C which is about 30°C
higher than the melting point of the molten salt. A hydrogen storage tank and fuel cell unit is also integrated to
the Cu-Cl cycle for energy management. Energy management with a hydrogen storage option is promising,
since hydrogen can be converted to electricity efficiently via fuel cells during peak hours. System I also
comprises a steam turbine cycle (STC), which has a low pressure and a high pressure steam turbine, and a
LiBr-H2O ACS. The steam turbine cycle and absorption system also use the solar energy, while the ORC
utilizes waste heat from the steam turbine cycle. Fig. 1 shows the system diagram, and Table 1 identifies the
flows in the system. Water streams in Table 1 are differentiated in terms of its use:
• 13-30: H2O is working fluid of the STC.
• 37-40: H2O is working fluid of the ACS.
• 41-44: H2O is for heat rejection from ACS.
• 47-49: H2O is used in hot water production.
• 51-52: H2O is feed to the Cu-Cl cycle.

2.1 System II – Cu-Cl Based Multi-generation System Using Nuclear Energy

The second Cu-Cl based integrated system considered here utilizes nuclear power. A Generation IV
supercritical water cooled reactor (SCWR) is a suitable match for the Cu-Cl cycle. The HTF can be both water
and molten salt in this case. Unlike solar based systems, nuclear power does not have intermittent
characteristics. The nuclear plant can continuously supply heat to the Cu-Cl cycle using water as the HTF. The
coolant (water) inlet and exit temperatures are defined as 350 and 625°C, respectively. Also, the coolant
pressure and mass flow rate are 25 MPa and 1320 kg/s (Pioro and Duffey, 2007). Thus, water is selected as
HTF for System II. The LiBr-H2O ACS is also introduced to provide cooling, so there are multiple outputs in
System II. A hydrogen storage and fuel cell unit are also used in System II. Fig. 2 shows the system diagram,
and Table 2 identifies the flows in the system. Similarly, water streams in Table 2 are differentiated in terms of
its use:
• 1-6: H2O is heat transfer fluid.
• 13-16: H2O is working fluid of the ACS.
• 17-20: H2O is for heat rejection from ACS.
• 23-25: H2O is used in hot water production.
• 27-28: H2O is feed to the Cu-Cl cycle.

621
Fig. 1. Schematic diagram of System I.
Table 1: Flows in System I.
Steam
turbine
Solar cycle Absorption Cooling
Sub-unit tower (STC) system (ACS) Drying air Hot water Cu-Cl cycle
54-
Stream number 1-12 13-30 31-36 37-40 41-44 45-46 47-49 51,52 57 58
Stream Molten LiBr-
H2O H2O H2O Air H2O H2O H2 O2
material salt H2O

Fig. 2. Schematic diagram of System II.

622
Table 2: Flows in System II.
Nuclear Absorption cooling system Drying Hot
Sub-unit Cu-Cl cycle
HTF (ACS) air water
Stream
1-6 7-12 13-16 17-20 21-22 23-25 27,28 30,33 34
number
Stream LiBr-
H2O H2O H2O Air H2O H2O H2 O2
material H2O

3. ANALYSIS
Analyses of the components in the systems (Figs. 1 and 2) are conducted using mass, energy, entropy and
exergy balances. The assumptions considered throughout the analysis are as follows:
• Constant ambient temperature and pressure.
• Steady state operation.
• Adiabatic compressors, turbines and heat exchangers.
• Isentropic efficiency for turbines and compressors of 0.9.
• Heat exchanger effectiveness of 0.85 for all heat exchangers.
• Hydrogen storage and fuel cell unit is not included to thermodynamic calculations. Since, only steady
state operations are considered.
• Efficiency values of heliostat solar tower and SCWR are not taken into account. Hence, inputs for both
systems are considered to be the energy content of the heat transfer fluid.

The energy balance used in the analysis considering each component is given as follows:

𝑄̇𝑐𝑣 − 𝑊̇𝑐𝑣 = ∑𝑖𝑛 𝑚̇𝑖 ℎ𝑖 − ∑𝑜𝑢𝑡 𝑚̇𝑖 ℎ𝑖 (1)

where 𝑚̇𝑖 and ℎ𝑖 represent the mass flow rate and specific enthalpy of the flow stream into and out from each
component in the system.
Exergy is defined as the maximum amount of work which can be produced by a system or a flow of
matter or energy as it comes to equilibrium with a reference environment. Exergy is a measure of the
potential of the system or flow to cause change as a consequence of not being in stable equilibrium with the
reference environment (Rosen and Dincer, 2001, 2013). The exergy balance used in the analysis follows:

0 = 𝐸𝑥̇𝑄,𝑐𝑣 − 𝑊̇𝑐𝑣 + ∑𝑖𝑛 𝑚̇𝑖 𝑒𝑥𝑖 − ∑𝑜𝑢𝑡 𝑚̇𝑖 𝑒𝑥𝑖 − 𝐸𝑥̇𝑑 𝑖 (2)

The exergy quantities in an exergy balance are described below by Dincer and Rosen (2013):

Exergy of thermal energy (ExQ): The exergy associated with a thermal energy transfer Q can be expressed
as:

𝑇
𝐸𝑥𝑄 = ∑𝑛𝑖=1 �𝑄𝑖 ∗ �1 − 0�� (3)
𝑇𝑖

Specific exergy of a matter flow (exi): The exergy of flow can be expressed in terms of physical, chemical,
kinetic and potential components. Neglecting the kinetic and potential energy changes in the components
considered in this study, exi of flow streams can be defined as

𝑒𝑥𝑖 = (ℎ𝑖 − ℎ𝑜 ) − 𝑇𝑜 (𝑠𝑖 − 𝑠𝑜 ) + 𝑒𝑥𝑐ℎ,𝑖 (4)

where 𝑒𝑥𝑐ℎ,𝑖 is the specific chemical exergy of flow stream i.

Exergy destruction (Exd): Exergy destruction in a component is proportional to the entropy generation due to
irreversibilities and can be written as

𝐸𝑥𝑑 𝑖 = 𝑇𝑜 . 𝑆𝑔𝑒𝑛,𝑖 (5)

where Sgen,i denotes the entropy generation in each component.

623
The energy efficiency of System I can be expressed as follows:

𝑚̇𝐻2 × 𝐿𝐻𝑉𝐻2 +𝑊̇𝑛𝑒𝑡 +𝑄̇𝑐𝑜𝑜𝑙𝑖𝑛𝑔 +𝑚̇𝑎𝑖𝑟 ×�ℎ𝑎𝑖𝑟−ℎ𝑎𝑖𝑟,0 �+𝑚̇ℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 ×�ℎℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 −ℎ𝑤𝑎𝑡𝑒𝑟,0 �

𝜂𝑒𝑛 = (6)
𝑚̇ℎ𝑡𝑓 ×(ℎ1 −ℎ12 )

where 𝑊̇𝑛𝑒𝑡 = 𝑊̇𝑛𝑒𝑡,𝑆𝑇𝐶 − 𝑊̇𝑖𝑛,𝐶𝑢𝐶𝑙

The exergy efficiency of System I can be expressed as follows:

𝐸𝑥̇𝐻2 +𝑊̇𝑛𝑒𝑡 +𝐸𝑥̇𝑄𝑐𝑜𝑜𝑙𝑖𝑛𝑔 +𝑚̇𝑎𝑖𝑟×�𝑒𝑥𝑎𝑖𝑟 −𝑒𝑥𝑎𝑖𝑟,0 �+𝑚̇ℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 ×�𝑒𝑥ℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 −𝑒𝑥𝑤𝑎𝑡𝑒𝑟,0 �

𝜂𝑒𝑥 = (7)
𝑚̇ℎ𝑡𝑓 ×(𝑒𝑥1 −𝑒𝑥12 )

Similarly, the energy and exergy efficiencies respectively of System II follow:

𝑚̇𝐻2 × 𝐿𝐻𝑉𝐻2 +𝑄̇𝑐𝑜𝑜𝑙𝑖𝑛𝑔 +𝑚̇𝑎𝑖𝑟×�ℎ𝑎𝑖𝑟 −ℎ𝑎𝑖𝑟,0 �+𝑚̇ℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 ×�ℎℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 −ℎ𝑤𝑎𝑡𝑒𝑟,0 �

𝜂𝑒𝑛 = (8)
𝑚̇ℎ𝑡𝑓 ×(ℎ1 −ℎ12 )+𝑊̇𝑖𝑛,𝐶𝑢𝐶𝑙

𝐸𝑥̇𝐻2 +𝐸𝑥̇𝑄𝑐𝑜𝑜𝑙𝑖𝑛𝑔 +𝑚̇𝑎𝑖𝑟×�𝑒𝑥𝑎𝑖𝑟 −𝑒𝑥𝑎𝑖𝑟,0 �+𝑚̇ℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 ×�𝑒𝑥ℎ𝑜𝑡𝑤𝑎𝑡𝑒𝑟 −𝑒𝑥𝑤𝑎𝑡𝑒𝑟,0 �

𝜂𝑒𝑥 = (9)
𝑚̇ℎ𝑡𝑓 ×(𝑒𝑥1 −𝑒𝑥12 )+𝑊̇𝑖𝑛,𝐶𝑢𝐶𝑙

4. RESULTS AND DISCUSSION

The results of the comprehensive thermodynamic analysis of the Cu-Cl based integrated multi-generation
system are presented in this section. EES (Engineering Equation Solver) software is used to study the
performance of the system and its components. The parametric studies are carried out by varying some of
the system parameters.
Fig. 3 shows the energy and exergy efficiencies of System I and its sub-units. The energy efficiency
of the overall system is 70% whereas the exergy efficiency is 57%. However, if the heat released by the
condenser of the steam turbine cycle is not utilized as hot water, the energy efficiency value would be 51%.
The COP of the absorption cooling system is 0.77 and the exergetic COP is 0.30. The exergy efficiency of
the Steam turbine cycle (STC) is greater than its energy efficiency, since the exergy of output power is the
power itself although exergy of input heat is lower than the energy of the heat input.
0,9
Energy efficiency
0,8
Exergy Efficiency
0,7

0,6

0,5

0,4

0,3

0,2

0,1

0
Rankine Cu-Cl ACS (COP) Overall
Fig. 3. Energy and exergy efficiency of System I and its sub-units.

The energy and exergy efficiencies of the System II and its sub-units are shown in Fig. 4. The overall
energy efficiency of System II is 51% and the exergy efficiency is 41%. The absorption cooling system and
the Cu-Cl cycle used in System II have similar efficiency values as the ones in System I.

624
 0,9
Energy efficiency
0,8
Exergy efficiency
0,7

0,6

0,5

0,4

0,3

0,2

0,1

0
Cu-Cl ACS (COP) Overall
Fig. 4. Energy and exergy efficiency of System II and its sub-units.

Figures 5 and 6 show the percentage contributions of System I and System II outputs to flows of (a)
total energy and (b) total exergy. The energy percentages of all outputs for both systems are close to each
other. The exergies of power and hydrogen dominate the System I exergy share, whereas the exergy of the
output is 94% of total exergy output for System II. Energy of cooling is defined as the amount of heat
extracted from the cooling heat transfer fluid (water for both systems).

(a) (b)
Fig. 5. Percentage contributions of System I outputs to (a) total energy and (b) total exergy flow.

(a) (b)
Fig. 6. Percentage contributions of System II outputs to (a) total energy and (b) total exergy flow.

625
 Figures 7 (a) and (b) shows the impact of heat exchanger effectiveness on energy and exergy
efficiencies. In both figures variation of efficiencies with “no hot water” cases are also presented. Both energy
and exergy efficiencies increase with increasing effectiveness due to reduction in lost energy to the
environment. Figures 7 (a) and (b) also show that utilizing the rejected heat out of the system as a system
product as hot water significantly increases the energy efficiencies of the systems. Exergy efficiency results,
however, are very close for the both “utilized hot water” and “no hot water” cases in System I and II.

CONCLUSIONS
Thermodynamic analyses based on energy and exergy are successfully performed to investigate the
performance of the Cu-Cl based, integrated multi-generation systems. The overall energy efficiency of System
I is 70% whereas the exergy efficiency is 57%. The overall energy efficiency of System II is 51% and its
exergy efficiency is 41%. The produced hydrogen is the main contributor to the exergy output flows of both
systems. It is concluded that utilizing rejected heat as hot water to the utilities significantly increases the
energy efficiency, but does not have a major impact on exergy efficiency, mainly since the temperature of the
hot water is close to reference-environment temperature.

Acknowledgement
The authors acknowledge gratefully the financial support provided by the Ontario Research Excellence Fund.

0.9

h en,system1
h ex,system1
0.8
h en,system,nohotwater
h ex,system,nohotwater

0.7
Efficiency

0.6

0.5

0.4
0.75 0.8 0.85 0.9 0.95
ehex
(a)

626
 0.9
h en,system1

0.8 h ex,system1

h en,system,nohotwater

0.7 h ex,system,nohotwater
Efficiency

0.6

0.5

0.4

0.3
0.75 0.8 0.85 0.9 0.95
ehex
(b)
Fig. 7. Variation of energy and exergy efficiencies with heat exchanger effectiveness for (a) System I and (b)
System II.

NOMENCLATURE
ex Specific exergy, MJ/kg
Exd Exergy destruction, MJ
exch Specific chemical exergy, MJ/kg
ExQ Exergy of thermal energy, MJ
h Specific enthalpy, kJ/kg
LHV Lower heating value, MJ
𝑚̇ Mass flow rate, kg/s
𝜂𝑒𝑛 Energy efficiency
𝜂𝑒𝑥 Exergy efficiency
Q Heat, MJ
Sgen Entropy generation, MJ/K
Ti System temperature, K
T0 Reference environment temperature, K
W Work, MJ
𝑊̇ net Net power output, MW

Acronyms
ACS Absorption cooling system
EES Engineering equation solver
HP High pressure

627
HTF Heat transfer fluid
LP Low pressure
SCWR Supercritical water cooled reactor
STC Steam turbine cycle

REFERENCES
Dincer I. and Rosen M. A. 2013. Exergy: Energy, Environment and Sustainable Development, 2d ed. UK: Elsevier.

Dincer I., 2012. Green methods for hydrogen production. International Journal of Hydrogen Energy 37: 1954-1971.

Dufour J., Serrano D. P., Galvez J.L., Moreno J. and Garcia C. 2009. Life cycle assessment of processes for
hydrogen production: Environmental feasibility and reduction of greenhouse gases emissions. International
Journal of Hydrogen Energy 34: 1370-1376.

Gebreslassie B. H., Medrano M. and Boer D. 2010. Exergy analysis of multi-effect water–LiBr absorption systems:
From half to triple effect. Renewable Energy 35: 1773-1782.

Kotas T. J. 1995. The Exergy Method of Thermal Plant Analysis. Krieger: Malabar, Florida.

Lewis M. A., Masin J. G. and O’Hare P. A., 2009. Evaluation of alternative thermochemical cycles. Part I: The
methodology. International Journal of Hydrogen Energy 34: 4115–4124.

Moran, M.J. 1989. Availability Analysis: A Guide to Efficient Energy Use. American Society of Mechanical
Engineers: New York.

Naterer G. F., Fowler M., Cotton J. and Gabriel K. 2008. Synergistic roles of off-peak electrolysis and
thermochemical production of hydrogen from nuclear energy in Canada. International Journal of Hydrogen Energy
33: 6849-6857.

Naterer G. F., Suppiah S., Stolberg L., Lewis M., Wang Z., Dincer I., Rosen M. A., Gabriel K., Secnik E., Easton E.
B., Pioro I., Lvov S., Jiang J., Mostaghimi J., Ikeda B. M., Rizvi G., Lu L., Oduyaka A., Spekkens P., Fowler M. and
Avsec J. 2013. Progress of international hydrogen production network for the thermochemical Cu-Cl cycle.
International Journal of Hydrogen Energy 38: 740-759.

Nezammahalleh H., Farhadi F. and Tanhaemami M. 2010. Conceptual design and techno-economic assessment
of integrated solar combined cycle system with DSG technology. Solar Energy 84: 1696-1705.

Pilavachi P. A., Chatzinapagi A. I. and Spyropoulou A. I. 2009. Evaluation of hydrogen production methods using
the analytic hierarchy process. International Journal of Hydrogen Energy 34: 5294-5303.

Rosen M. A. and Dincer I. 2001. Exergy as the confluence of energy, environment and sustainable development.
Exergy, an International Journal 1: 3-13.

Rosen M. A. 2010. Advances in hydrogen production by thermochemical water decomposition: A review. Energy
35: 1068-1076.

Urbaniec K., Friedl A., Huisingh D. and Claassen P. 2010. Hydrogen for a sustainable global economy. Journal of
Cleaner Production 18: S1-S3.

Pioro, I. L., and Duffey, R. B. 2007. Heat Transfer and Hydraulic Resistance at Supercritical Pressures in Power
Engineering Applications. New York: ASME Press.

Wang Z. L., Naterer G. F., Gabriel K .S., Secnik E., Gravelsins R. and Daggupati V. N. 2011. Thermal design of a
solar hydrogen plant with a copper-chlorine cycle and molten salt energy storage. International Journal of
Hydrogen Energy 36: 11258-11272.

628
 EXERGETIC PERFORMANCE OF A LOW BYPASS TURBOFAN ENGINE AT TAKEOFF
CONDITION

H. Aydın, O. Turan, A. Midilli and T. H. Karakoc

TUSAS Engine Industries (TEI), TR-26003, Eskisehir, Turkey
E-mail: tei.hakan@gmail.com
Faculty of Aeronautics and Astronautics, Anadolu University, TR-26470 Eskisehir, Turkey
E-mail: onderturan@anadolu.edu.tr
Advanced Energy Technologies, Energy Division, Mechanical Engineering Department, Faculty of Engineering, Recep
Tayyip Erdogan University, Rize, 53000, Turkey
E-mail: adnan.midilli@erdogan.edu.tr
Faculty of Aeronautics and Astronautics, Anadolu University, TR-26470 Eskisehir, Turkey
E-mail: hikmetkarakoc@gmail.com

ABSTRACT
In this article, exergetic methodology is applied for a low bypass turbofan engine at maximum power setting. The
engine is a low-bypass (0.96 to 1) turbofan engine and its variants fitted to the 737-100/200 all comprise six low-
pressure compressor (LPC) stages, seven high pressure compressor (HPC) stages, a single HP turbine (HPT), and
finally three LPT stages. At the end of the analysis, the most irreversible units in the system are found to be the
combustor and the fan/low pressure compressor, with exergy loss rates of 18.7 MW and 2.486 MW, respectively.
The exergy efficiencies of the fan/LPC, the HPC and combustor are 0.856, 0.845 and 0.744, respectively. For the
HPT and LPT, the exergy efficiencies are calculated to be 0.98 and 0.963, respectively.

INTRODUCTION
With over two billion people travelling safely around the world every year and some 23,000 aircraft in commercial
service. Worldwide passenger traffic will average 5.1 percent growth and cargo traffic will average 5.6 percent
growth 2–5% of the world energy consumption belongs to aviation industries (Boeing, 2011; Enviro, 2011; IATA,
2005; Lee et al., 2007; USHP, 2006). Total scheduled world revenue tonne kilometers (RTK) increased by 119 per
cent, with scheduled passenger revenue passenger kilometers (RPK) and cargo (RTK) traffic rising by 108 and
140 per cent, respectively (Macintosh and Wallace, 2008). Effects of energy consumption in aviation sector give
rise to potential environmental hazards. Therefore energy consumption plays a crucial importance role to achieve
sustainable development; balancing economic and social development with environmental protection. The
importance of energy efficiency is also linked to environmental problems, such as global warming and atmospheric
pollution (Ahmadi et al., 2011; Ptasinski et al., 2006).

Energy intensity can be related to some important measures operational and technological efficiency in the aircraft
and its propulsion system. An aero-engine converts the flow of chemical energy contained in the kerosene fuel
and the air drawn into the engine into propulsion power. Nearly one-fourth or one-third of fuel energy is used to
propel the aircraft. The remaining energy is expelled as waste heat in the exhaust. Specific fuel consumption
(SFC) is more relevant to consider propulsion power in terms of payload carried per unit range. Energy intensity
( EI ) is a suitable parameter with comparing efficiency and environmental impact. It consists of two components-
energy use, EU and load factor, L f as shown in Eq.1 (Joosung et al., 2004).
EU MJ MJ RPK (1)
=
EI = = /
L f RPK ASK ASK
Where MJ is mega joules of kerosene fuel energy, RPK is revenue passenger-kilometers, ASK is available seat-
kilometers, and L f is load factor. To have a model of aircraft, it is necessary to show EI as a function of the
engine, aerodynamic, and structural efficiency of the aircraft system as well as load factor. These parameters play
important role in the energy intensity of an aircraft. Energy efficiency in commercial aircrafts is improved by
averaging 1.5% percent annually with the introduction of bypass turbofan engines. However, as the bypass ratio
increased, engine diameter has also increase, leading to an increase momentum drag. Other way to propulsion
system improvement is to increase turbine inlet temperature, which is limited by materials and cooling technology,
and improving engine component efficiencies. Between the introduction B707 and B777, commercial aircrafts
have been constructed exclusively of aluminum and are currently about 90% metallic by weight. So improvements
of structural efficiency are less evident (Joosung et al., 2004).
The environmental impact of emissions can be reduced by increasing the efficiency of resource utilization (Rosen,
2002). Using energy with better efficiency reduces pollutant emissions. Energy and exergy concepts have been
utilized in environmental sustainability, economics and engineering. Exergy is a quantity which follows from the
First and Second Laws of Thermodynamics and analyses directly impact process design and improvements
because exergy methods help in understanding and improving efficiency, environmental and economic
performance as well as sustainability. The potential usefulness of exergy analysis in addressing sustainability

629
issues and solving environmental problems is substantial (Ao et al., 2008; Dincer and Rosen, 1998; Midilli and
Dincer, 2009, 2010; Norberg et al., 2009). The exergy studies related to gas turbines have first been done on
stationary gas turbines. In the literature, the various exergy and exergo-economic analysis of aero engines have
been reported (Aydin et al., 2012a, 2012b; Ballı et al., 2008; Bejan and Seems, 2001; Brilliant, 1995; Cesare et al.,
2010; Diango et al., 2011; Etele and Rosen, 2001; Figliola, 2003; Riggins, 1996; Roth and Mavris, 2000, 2001;
Schiffmann and Favrat, 2010; Turan, 2012a, 2012b).

Through a literature review, it is noticed that there is no work to be studied about exergy analysis for a low bypass
turbofan engines in the open literatures. The present assessment, therefore, aims to provide a practical framework
for the use of such exergy analysis in low bypass engines. Lack of exergy analysis for low bypass turbofan engine
makes the paper original and becomes main motivation.

In this paper, first the detailed exergy analysis of JT8D low bypass turbofan engine has been performed. In this
analysis, exergy efficiency, exergy losses/destructions have been calculated at maximum power setting, i.e.
takeoff condition. Exergy analysis of JT8D has first been studied in this paper. Moreover, there is no previous work
about exergy related to low bypass turbofan engine in the aircraft engine in the literature.

SYSTEM DESCRIPTION FOR LOW BYPASS TURBOFAN ENGINE

JT8D series engines are one of the most popular modern commercial engines ever made. More than 14,750 of
them have been built, amassing more than 673 million hours of reliable service since 1964. The eight models that
make up the JT8D family cover a thrust range from 62 to 76 kN. The newer JT8D-200 engine offers 18,500 to
21,700 pounds of thrust, and is the exclusive power for the popular MD-80 aircraft (B-1-7.pdf)

An illustrated diagram, station numbering and main component of the high bypass turbofan engine is shown in Fig.
1. It consist of fan (F), axial low pressure compressor (LPC), axial high pressure compressor (HPC), an annular
combustion chamber, high-pressure turbine (HPT) and low pressure turbine (LPT).

HPC: High pressure compressor; CC: combustion chamber; HPT: High pressure turbine; LPT: Low pressure turbine; EN: Exhaust nozzle; FN: Fan
nozzle

This engine operates according to the Brayton cycle, which includes four processes under the ideal conditions
given below:
a. isentropic compression (fan and HPC)
b. combustion at constant pressure (CC)
c. isentropic expansion (HPT and LPT)
d. heat transfer at constant pressure (EN and FN).
There are two drive shafts in this engine. The first, N2, connects the HPT and HPC and constitutes the HP
system, while the second, N1, connects the LPT to the fan and constitutes the LP system. While the high pressure
turbine runs the high pressure compressor, fuel pump, starter generator and reduction gearbox, LPT runs the fan.
In the schematic diagram of the high bypass turbofan engine given in Fig. 1, the bypass ratio is defined as
Bypass airflow m fan m cold
α
= = = (2)
Primary airflow m core m hot
 core , then the bypass air mass flow rate is (α m core ) .
Thus if the air mass through the core (HPC) is m

MATHEMATICAL FRAMEWORKS FOR THERMODYNAMIC ANALYSIS

630
The energy and exergy methods in practice: some useful tools and definitions

Thermodynamic first-law analysis is energy-based approach in thermal systems. It is based on the principle of
conservation of energy applied to the system. For a general steady state, steady-flow process, the four balance
equations (mass, energy, entropy and exergy) are applied to find the work and heat interactions, the rate of exergy
decrease, the rate of irreversibility, the energy and exergy efficiencies (Balkan et al., 2005; Dincer et al., 2004;
Wall, 2003).

The mass balance equation can be expressed in the rate form as

∑ m =∑ m
in out
, (3a)
where m  is the mass flow rate, and the subscript in stands for inlet and out for outlet. The general energy
balance can be expressed below as the total energy inputs equal to total energy outputs.

∑ E =∑ E
in out
(3b)
Energy conservation suggests that for a steady-state process the First Law may be represented by (Hammond and
Stapleton, 2001):

∑ ( h + ke + pe ) in
m in − ∑ ( h + ke + pe )out m out + ∑ Q − W =
0 (3c)

where m in and m out denote the mass flow rate across the system inlet and outlet, respectively, Q represents the
heat transfer rate across the system boundary, W is the work rate (including shaft work, electricity, and so on)
transferred out of the system, and h, ke, and pe denote the specific values of enthalpy, kinetic energy, and
potential energy, respectively.
This energy balance can be simplified, assuming negligibly small changes in kinetic and potential energy and no
heat or work transfers, to (Hammond and Stapleton 2001):

∑H i ,in = ∑ H j ,out (3d)

where Hi,in represents the various energy (or enthalpy) streams flowing into the system, and Hj,out the different
energy outputs.

First Law or ‘energy’ analysis takes no account of the energy source in terms of its thermodynamic quality. It
enables energy or heat losses to be estimated, but yields only limited information about the optimal conversion of
energy. In contrast, the Second Law of Thermodynamics indicates that, whereas work input into a system can be
fully converted to heat and internal energy, not all the heat input can be converted into useful work (Hammond,
2007). The exergy loss in a system or component is determined by multiplying the absolute temperature of the
surroundings by the entropy increase (Hepbasli, 2008; Hermann, 2006; Kilkis, 1999). Exergy methods also help in
understanding and improving efficiency, environmental and economic performance as well as sustainability
(Genoud and Lesourd, 2009).

Note that, whereas energy is a conserved quantity, exergy is not and is always destroyed when entropy is
produced. In the absence of electricity, magnetism, surface tension and nuclear reaction, the total exergy of a
system Ex  can be divided into four components, namely (i) physical exergy Ex
 PH (ii) kinetic exergy Ex
 KN (iii)
 and (iv) chemical exergy Ex
potential exergy Ex PT  CH
(Hepbasli, 2008).

 = Ex
Ex  PH + Ex
 KN + Ex
 CH + Ex
 PT (4a)
Although exergy is extensive property, it is often convenient to work with it on a unit of mass or molar basis.
The total specific exergy on a mass basis may be written as follows (Hepbasli, 2008):
ex = ex PH + ex KN + exCH + ex PT (4b)

The general exergy balance can be written as follows (Hepbasli, 2008):

∑ Ex

in − ∑ Ex

out = ∑ Exdest + ∑ Exloss
  (5a)
     
Exheat − Exwork + Exmass ,in − Exmass , out = Exdest + Exloss (5b)
 T0   ,
=
Exheat ∑ 1 − T Qk (5c)
 k 
 ,
work = W
Ex (5d)

631
 mass ,in = ∑ minψ in

Ex  , (5e)

mass , out = ∑ moutψ out


Ex  . (5f)

where Q is the heat transfer rate through the boundary at temperature Tk at location k and W is the work rate.
The flow (specific) exergy is calculated as follows:
ex =( h − h0 ) − T0 ( s − s0 ) (6)
where h is enthalpy, s is entropy, and the subscript zero indicates properties at the restricted dead state of P0
and T0.
The rate form of the entropy balance can be expressed as (Hepbasli, 2008)
Sin − Sout + Sgen =
0 (7)

where the rates of entropy transfer by heat transferred at a rate of Q k and mass flowing at a rate of m are
Sheat = Q k / Tk and Smass = ms  , respectively (Hepbasli, 2008).
Taking the positive direction of heat transfer to be to the system, the rate form of the general entropy relation
given in Eq. (7) can be rearranged to give (Hepbasli, 2008)
Q
Sgen = ∑ m out sout −∑ m in sin −∑ k (8)
Tk
Also, it is usually more convenient to find Sgen first and then to evaluate the exergy destroyed or the irreversibility
rate I directly from the following equation, which is called Gouy–Stodola relation (Szargut et al., 1988):

= = T S
I Ex (9)
dest 0 gen

Assuming air to be a perfect gas, the specific physical exergy of air is calculated by the following relation (Kotas,
1995)
 T  P
exair=
, per C p ,a  T − T0 − T0 ln  + RaT0 ln (10)
 T0  P0
Numerous ways of formulating exergy (or second-law) efficiency for various energy systems are given in detail
elsewhere (Cornelissen, 1997). It is very useful to define efficiencies based on exergy. There is no standard set of
definitions in the literature. Here, exergy efficiency is defined as the ratio of total exergy output to total exergy input,
i.e.
E x
ηex =  out (11)
Exin

Assumptions

In this study, the assumptions made are listed below

(a) The air and combustion gas flows in the engine are assumed to behave ideally.
(b) The combustion reaction is complete
(c) Compressors and turbines are assumed to be adiabatic
(d) Ambient temperature and pressure values are 288.15 K and 101.35 kPa, respectively.
(e) The exergy analyses are performed for the lower heating value (LHV) of kerosene (JET A1) which is accepted as
42,800kJ/kg.
(h) Engine accessories, pumps (fuel, oil and hydraulic) are not included in the analysis
(i) The kinetic and potential exergies are neglected
(j) Chemical exergy is neglected other than combustor.

Airflow

The total airflow mass is 142.7 kg/s that includes 74.74 kg/s fan air and 67.95 kg/s core air. Air is taken into LPC at
ambient temperature of 288.15 K and ambient pressure of 101.35 kPa. In gas turbine engines, a part of
compressed air is extracted to use for ancillary purposes, such as cooling, sealing and thrust balancing. In this
study the cooling airflow is neglected since it doesn’t have meaningful effect on exergy and sustainability
analyses.

Combustion balances and emissions

632
As fuel the kerosene (JET A) is burned. Its chemical formula is as C12H23. The value of LHV is 42,800 kJ/kg. Fuel
flow is 1.05 kg/s that results in air/fuel ratio as 64. Combustion balance equation is calculated by Eq.(12),

 0.7748 N 2  12.11CO2 
   
 + 0.2059O2   + 18.51H 2O 
C12 H 23 + 369 ⇒
+ 0.0003CO2   + 58.22O2 
(12)
   
 + 0.019 H O   + 285.9 N 
 2   2 

The mass of combustion gases are obtained as 3.43 kg/s for CO2, 2.14 kg/s for H2O, 11.98 kg/s for O2 and 51.46
kg/s for N2 after combustor chamber.

Specific heat capacities of air and combustion gas

The cold air specific heat capacity is calculated by Eq. (2) as follow

9.45378T 2 5.49031T 3 7.92981T 4 (13)

cP ,air (T ) =
1.04841 − 0.000383719T + − +
107 1010 1014
The specific heat capacity of the combustion gases after combustion chamber for JT8D is calculated from the
composition of Eq (1) of each emissions’ mass percentage as follow.

2.043T 1.551T 2 6.717T 3 (14)

cP , gas (T ) =0.9886 + + −
105 107 1011
Where the unit of temperature is K. For hot gases, R value is calculated as 0.2901 kJ/kg.

EXERGY ANALYSIS

The exergy analysis of JT8D gas turbine engine’s Fan, HPC, combustor, HPT and LPT will be performed. The
exergy parameters of JT8D for two investigated operating conditions will be calculated by Eqs. (13-19)
i. Fan;

∑ E x in , fan − ∑ E x out , fan = ∑ E xdest , fan (15a)

∑ E x in , fan − ∑ E x out , fan = W fan + E x2 − ( E x13 + E x25 )

(15b)
m fan (h13 − h2 ) + m A (h25 − h2 )
W fan =
MA (15c)

E x + E x25 − E x2
η ex , fan = 13
W fan (15d)

ii. High Pressure Compressor;

∑ E x in , HPC − ∑ E x out ,HPC = ∑ E xdest ,HPC

(16a)
∑ E x in , HPC − ∑ E x out ,HPC = W HPC + E x25 − E x3
(16b)
m (h − h25 )
W HPC = HPC 3
MA (16c)

E x3 − E x2.5
ηex , HPC =
W HPC (16d)

Combustor (CC);
iii.

633
 ∑ E x in ,CC − ∑ E x out ,CC = ∑ E xdest ,CC
(17a)

∑ E x in ,CC − ∑ E x out ,CC = E x3 + E x3, fuel − E x4

(17b)
E x4
η ex ,CC =
E x3 + E x3, fuel (17c)

iv. High Pressure Turbine (HPT):

∑ E x in , HPT − ∑ E x out ,HPT = ∑ E xdest ,HPT
(18a)
∑ E x in , HPT − ∑ E x out ,HPT = E x4 − (W HPT + E x45 )
(18b)
W HPT
η ex ,HPT = (18c)
E x4 − E x45

v. Low Pressure Turbine (LPT):

∑ E x in , LPT − ∑ E x out ,LPT = ∑ E xdest ,LPT (19a)

∑ E x in , LPT − ∑ E x out ,LPT = E x45 − (W LPT + E x5 )

(19b)
W LPT
ηex , LPT =
E x45 − E x5 (19c)

JT8D thermodynamic parameters are listed in Tables 1 and 2.

Table 1. JT8D Turbofan engine thermodynamic data for takeoff thrust running.

Mass flow Temperature Pressure Exergy flow

Station No Location
(kg s-1) (K) (kPa) (MW)
0 Air 677.2 288.15 101.35 0
2 FAN inlet 677.2 288.15 101.35 0
1.3 FAN bypass outlet 565.3 327.6 155.8 21.51
2.5 FAN core outlet 111.9 372 221.3 8.39
2.5 HPC inlet 111.9 372 221.3 8.39
3 HPC outlet 111.9 744.2 2178.7 50.55
3 Combustor inlet 111.9 744.2 2178.7 50.55
3 Fuel 2,110 288.15 2000 94.37
4 Combustor outlet 114,2 1350 2082.2 1103
4 HPT inlet 114.2 1350 2082.2 110.3
4.5 HPT outlet 114.2 985 535 59.43
4.5 LPT inlet 114.2 985 535 59.43
5 LPT outlet 114.2 727.6 144.1 24.51
Source: (Farokhi, 2009)

Table 2. Exergy values of the JT8D turbofan engine and its component for takeoff thrust running.

Exergy
Inlet exergy Outlet exergy Exergy dest.
Component Efficiency
(MW) (MW) (MW)
(%)

FAN 32.43 29.90 2.53 0.922

634
 HPC 56.66 50.55 6.11 0.873
CC 144.92 110.30 34.62 0.761
HPT 110.30 108.18 2.12 0.958
LPT 59.43 57.26 2.17 0.938

JT8D 94.37 27.92 0.296

RESULTS AND CONCLUSIONS

In this paper, first the exergy analysis of JT8D high bypass turbofan engine at takeoff thrust power has been
carried out. In this analysis main exergetic parameters are energy and exergy flows, exergy destruction and exergy
efficiency. Now, it is necessary to definite the phases of flight for an aircraft. The phase of flight definitions given in
Table 3 consist of broad operational phases. Most of them have sub-phases. Considering the flight phases as a
function of engine power, the flight phases can be split into seven parts in this study: a) landing b) climb c)
maximum cruise d) normal take-off e) maximum continuous f) automatic power reverse g) maximum take-off.
Concerning the classification of flight phases as an engine power, it is difficult to see many examples in the
literature.

Table 2. Standard definitions the phases of a flight.

Phase Symbo Definition Sub-phases

l
prior to pushback or taxi, or after arrival, engine(s) i) not operating ii) start-
Standing STD at the gate or parking area, while the aircraft is up
stationary iii) operating iv) shut-down
engine(s) i) not operating ii) start-
Pushback/Towin aircraft is moving in the gate, ramp, or parking area,
PBT up
g assisted by a tow vehicle
iii) operating iv) shut-down
i) power back ii) taxi to runway iii)
aircraft is moving on the ground under its own
Taxi TXI taxi to take-off position iv) taxi from
power prior to take-off and after landing
run way
from the application of take-off power through
Take-off TOF rotation and to an altitude of 35 feet above runway i) take-off ii) rejected take-off
elevation
from the end of the take-off to the first prescribed
Initial Climb ICL power reduction, or until reaching 1,000 feet above -
runway elevation
from completion of Initial Climb through cruise i) climb to cruise ii) cruise iii)
En Route ENR altitude and completion of controlled descent to the change of cruise level iv) descent
Initial Approach Fix v) holding
Maneuvering MNVR low altitude/aerobatic flight operations i) aerobatics ii) low flying
From the Initial Approach Fix to the beginning of the i) initial approach ii) final approach
Approach APPR
landing flare. iii) missed approach/go-around
from the beginning of the landing flare until aircraft
exits the landing runway, comes to a stop on the i) flare ii) landing roll iii) aborted
Landing LDG
runway landing after touchdown

Emergency a controlled descent during any airborne phase in

EMG -
descent response to a perceived emergency situation
Uncontrolled a descent during any airborne phase in which the
UND -
descent aircraft does not sustain controlled flight
any of that portion of the flight which occurs after
Post-impact PIM -
impact with a person, object, obstacle

Source: (EADS, 2012)

635
 120

98,20 96,40
100
85,70 84,60
80
Exergy efficiency (%)

74,50

60

40
29,80

20

0
FAN HPC CC HPT LPT JT8D

Fig.2. Exergy efficiencies (%) of JT8D turbofan engine at takeoff thrust.

Fig.3.
Exergy
destruction (MW) of JT8D turbofan engine at takeoff thrust.

Fig. 2 demonstrates the exergy efficiencies of the fan, HPC, combustor, HPT, LPT and JT8D turbofan engine at
takeoff condition.

636
 80,0
74,3
70,0

60,0
Exergy destruction rate (%)

50,0

40,0

30,0

20,0 9,8 11,9

10,0

0,0 1,4
2,6
FAN
HPC
CC
HPT
LPT

Fig.4. Exergy destruction rates (%) of JT8D turbofan engine at takeoff thrust.

Figs. 3 and 4 presents the exergy destructions and exergy destruction rates for the fan, HPC, combustor, HPT,
LPT.
The exergy efficiency, one of the most important indicators for the sustainability of the engine, is mainly based on
the exergy input and the required output. It is noticed that the exergy efficiency of the turbofan engine highly
affected by the input-output exergetic values of the each engine component at all phases of a flight as shown in
Table 2. The results in Fig.2 show that the exergy efficiency ranges from 0.745 to 0.982 in engine components. As
can be seen in Fig.2, HPT and LPT are good exergy efficiencies changes between 0.964-0.982 due to higher
isentropic efficiencies. For the fan and HPC exergy efficiencies are found to be 0.857 and 0.846, respectively. On
the other hand, minimum exergy efficiency is observed in combustor (to be 0.745) due to internal irreversibilites in
CC.
The unit with greatest exergy loss is found to be CC (to be 18.87 MW) as shown in Fig.3. The exergy destructions
for the other units are found to be HPC (to be 3.03 MW), fan (to be 2.49 MW), LPT (to be 0.66 MW) and HPT (to
be 0.36 MW).If so, greatest exergy destruction rate is calculated in the CC (to be 74.3%) as shown in Fig.4. It is
clear from Fig.4 that HPT has minimum exergy destruction rate with value of 1.4%.

The results should provide a realistic and meaningful in the exergetic takeoff performance evaluation of low
bypass turbofan engines, which may be useful in the analysis of similar propulsion systems. In a future study, we
will focus on exergo-environmental and exergo-sustainability analysis of the low bypass turbofan engine. It is
noted that, to obtain more comprehensive conclusions, exergo-economics must be considered. In particular, an
exergo-economic analysis would be useful. An exergo-environmental analysis can help improve the environmental
performance of the low bypass engine, and consequently should be considered in future assessments.

Acknowledgement

The authors would like to express their appreciation to TUSAS Engine Industries (TEI) in Eskisehir city of Turkey
for full support throughout the preparation of this study. They are also grateful for the support provided for the
present work by Anadolu University, Eskisehir and Recep Tayyip Erdoğan University, Rize, Turkey.

NOMENCLATURE
-1
cp Specific heat (kJ. (kg K) )
E Energy rate (MW)
-1
ex Specific exergy (kJ.kg )
E x Exergy rate (MW)
f fuel-air ratio; fuel exergy factor
-1
hPR Fuel heating value (kJ.kg )
H Enthalpy (kJ)
I Irreversibility rate (kW)
ke Kinetic energy

637
m Mass flow rate (kg.s )
-1
2
A Area (m )
MA Molecular weight
pe Potential energy
P Pressure (bar or kPa); product exergy
-1
R Specific gas constant (kJ. (kg. K )), diameter (m)
-1
S Entropy (kJ.K )
T Temperature (K)
W Work rate (MW)

Subscripts
a Air
ch Chemical
dest Destruction
f Fuel
gen Generated
k kth component
LPC Low Pressure Compressor
HPC High Pressure Compressor
LPT Low Pressure Turbine
HPT High Pressure Turbine
in Inlet
ke Kinetic energy
LD Loss and destruction
kn Kinetic
out Outlet
per Perfect
ph Physical
pe Potential energy
tot Total

A area, m2
cp specific heat, J/kg°C
h heat transfer coefficient, W/m2°C
Greek Letters
η Efficiency
ρ
-3
Air density (kg.m )

REFERENCES

Boeing (2011) See also http://www.boeing.com. [accessed 15/08/2011].

Enviro (2011) http://www.enviro.aero/Content/Upload/File/BeginnersGuide_Biofuels_Web. [accessed
05/12/2011].
IATA (2011) http://www.boeing.com/commercial/cmo/forecast_summary.html. [accessed 20/09/2011].
Lee, J., Ian J., Waitz, A., Brian,Y., Kim, C., Gregg ,G., Fleming, L., Curtis, M., Holsclaw, A. 2007. System for
assessing aviation’s global emissions (SAGE), Part 2: Uncertainty assessment. Transportation Research D-Tr
E 12: 381–395.
USHP (2009) http://www.house.gov/transportation/aviation/02-15-06/02-15- 06memo.html. [accessed
25/08/2009].
Macintosh, A., Wallace, L. 2009. International aviation emissions to 2025: Can emissions be stabilised without
restricting demand?. Energy Policy 37:264–273.
Ahmadi, P., Dincer, I., Rosen, M.A. 2011. Exergy, exergoeconomic and environmental analyses and evolutionary
algorithm based multi-objective optimization of combined cycle power plants, Energy 36: 5886-5898.
Ptasinski, K.J., Koymans, M.N., Verspagen, H.H.G. 2006. Performance of the Dutch energy sector based on
energy, exergy and extended exergy accounting. Energy 31:3135–3144.

Joosung, I. L., Stephen P. Lukachko, and Ian, A. W., Aircraft and Energy Use. 2004.
http://web.mit.edu/aeroastro/people/waitz/publications/AircraftEnergyUse.pdf
Rosen, M.A.2002. ‘Assessing energy technologies and environmental impacts with the principles of
thermodynamics. Applied Energy.72: 427–441.

638
Ao, Y., Gunnewiek, L., Rosen, M. A. 2008. Critical review of exergy- based indicators for the environmental
impact of emissions. International Journal of Green Energy 5:1-2: 87-104.
Dincer, I., Rosen, M.A. 1998. Worldwide perspective on energy, environment and sustainable development,
International Journal of Energy Research 22: 1305-1321.
Midilli, A., Dincer, I. 2009. Development of some exergetic parameters for PEM fuel cells for measuring
environmental impact and sustainability, International Journal of Hydrogen Energy 34: 3858 – 3872.
Midilli, A., Dincer, I. 2010. Effects of some micro-level exergetic parameters of a PEMFC on the environment and
sustainability, International Journal of Global Warming, 2 (1): 65-80.
Norberg, S., Tamm, G., Highley, J., Rounds, M., Boettner, D., Arnas, O. 200).Teaching thermodynamics via
analysis of the west point power plant. International Journal of Green Energy, 6:3: 230-244.
Aydın, H., Turan, O., Midilli, A., Karakoç, T.H. 2012). Exergetic and exergo-economic analysis of a turboprop
engine: a case study for CT7-9C International Journal of Exergy 11, 1: pp.69-88.
Aydın, H., Turan, O., Midilli, A., Karakoç, T.H. 2012b. Component-based exergetic measures of an experimental
turboprop/turboshaft engine for propeller aircrafts and helicopters. International Journal of Exergy.11,
N3:.322 – 348.
Ballı, O., Aras, H., Hepbaslı, A. 2008. Exergetic and exergoeconomic analysis of an aircraft Jet Engine (AJE).
International Journal of Exergy 5(6): 567-581.
Bejan, A., Siems, D.L. 2001. The need for exergy analysis and thermodynamic optimization in aircraft
development. International Journal of Exergy 1(1):14-24.
Brilliant, H.M. 1995a.Second law analysis of present and future turbine engines, AIAA Paper, 95-3030, July.
Cesare, T., Paolo, A.R., Luiz, F. and Oliveira, S.D.J. 2010. Exergy and thermoeconomic analysis of a turbofan
engine during a typical commercial flight., Energy 35: 952–959.
Diango, A, Perilhon, C., Descombes, G., Danho, E. 2011. Application of exergy balances for the optimization of
non-adiabatic small turbomachines operation. Energy 36: 2924-2936.
Etele, J., Rosen, M.A. 2001. Sensitivity of exergy efficiencies of aerospace engines to reference environmental
selection. International Journal of Exergy 1(2):91-99.
Figliola, R.S., Tipton, R., Li, H. 2003. Exergy approach to decision-based design of integrated aircraft thermal
systems. Journal of Aircraft 40(1):49-55.
Riggins, D.W. 1996. High-speed engine/component performance assessment using exergy and thrust-based
methods, NASA-CR-198271, January.
Roth, B.A., Mavris, D.N. 2000. A comparison of thermodynamic loss models applied to the J79 Turbojet Engine;
Joint Propulsion Conference and Exhibit, 36th, Huntsville, July, AL, pp.16–19.
Roth, B., Mavris, D. 2001. A work availability perspective of turbofan engine performance. American Institute of
Aeronautics and Astronautics; AIAA publication, No. 0391.
Schiffmann, J., Favrat, D. 2010. Design, experimental investigation and multi-objective optimization of a small-
scale radial compressor for heat pump applications. Energy 35:436–450.
Turan, O. 2012a. Effect of reference altitudes for a turbofan engine with the aid of specific-exergy based method,
International Journal of Exergy 11.2: pp. 252-270.
Turan, O. 2012b. Exergetic effects of some design parameters on the small turbojet engine for unmanned air
vehicle applications. Energy, The International Journal 46:51-61.
Balkan, F, Colak, N,, Hepbasli, A. 2005. Performance evaluation of a triple effect evaporator with forward feed
using exergy analysis. International Journal of Energy Research 29: 455–470.
Dincer, I, Hussain, M,M,, Al-Zaharnah, I. 2004. Energy and exergy use in public and private sector of Saudi Arabia’,
Energy Policy 32(141):1615–1624.
Wall, G. 2003. Exergy tools. Proceedings Institution of Mechanical Engineering :125–136.
Hammond, G.P, Stapleton, A,J. 2001. Exergy analysis of the United Kingdom energy system. Proceedings of the
Institution of Mechanical Engineers Part A: Journal of Power and Energy 215, 2:141–162.
Hammond, G,P. 2007. Industrial energy analysis, thermodynamics and sustainability. Applied Energy 84: 675–
700.
Hepbasli, A. 2008. A key review on exergetic analysis and assessment of renewable energy resources for a
sustainable future. Renewable and Sustainable Energy Reviews 12: 593-661.
Hermann, W.A. 2006. Quantifying global exergy resources Energy31(12):1685–1702.
Kilkis, I.B. 1999. Utilization of wind energy in space heating and cooling with hybrid. Energy and Buildings 30:147–
153.
Genoud, S., Lesourd, J.B. 2009. Characterization of sustainable development indicators for various power
generation technologies.International Journal of Green Energy.6, No.3:257-267.
Cornelissen, R.L. 1997. Thermodynamics and sustainable development: the use of exergy analysis and the
reduction of irreversibility. PhD thesis, University of Twente, The Netherlands.
Szargut, J,M., Morris, D,R., Steward, F.R. 1988. Exergy analysis of thermal, chemical and metallurgical processes.
Hemisphere, New York.
Kotas, T.J. 1995. The exergy method of thermal plant analysis, Reprint ed., Krieger, Malabar, Florida.
Farokhi, Saeed, Aircraft propulsion, J. Wiley, c2009.
EADS 2012. Preliminary Aircraft Evaluation for SASEB, BLM National Aviation Office, Technical review.

639
 DEVELOPING SOME NEW EXERGETIC INDICES FOR ASSESSING THE
PERFORMANCE OF BUILDINGS

A. Hepbasli
Department of Energy Systems Engineering, Faculty of Engineering,
Yaşar University, 35100 Bornova, Izmir, Turkey

ABSTRACT
Buildings account for a large part of the yearly energy utilization in modern societies. Within the
European Union, the energy utilization by the built environment amounts to more than 40% of the total
energy consumption. In this context, for evaluating the whole performance of buildings, various
national and international standards as well as energetic indices (or metrics or indicators or indexes)
have been developed. In addition, some building performance metrics that can be used by building
owners, energy managers, and operators to track the performance of commercial facilities have also
been proposed.

In recent years, the concept exergy has become very popular and been considered a very useful tool
in the design, analysis and performance assessment of energy-related systems. Among these
systems, buildings play a very essential role in terms of the relatively low number of exergetic
applications. The present contribution aims at developing some new exergy-based performance
assessment indicators for buildings. In this regard, some energy-based indices are given first.
Modifying these from the exergetic points of view, new exergy-based ones are then proposed. Finally,
an illustrative example of a shopping center is presented while listing some concluding remarks. It may
be concluded that the most effective way of assessing the performance of a building seems to apply
the low-exergy approach.

Keywords: Buildings, Exergy, Building Performance, Exergy Indice, Exergy Indicator.

INTRODUCTION
Buildings play a very essential role in the EU's energy efficiency policy because about 40% of final
energy consumption (and 36% of greenhouse gas emissions) is in houses, offices, shops and other
buildings. Moreover, buildings offer the second largest untapped cost effective potential for energy
savings after the energy sector (EC, 2012). Since the beginning of the 1990s, the Member States in
Europe have dealt with the legal regulations on energy consumption to reduce carbon dioxide
emissions, according to Kyoto Protocol. Turkey is also responsible to provide regulations to comply for
the latest European Energy Performance of Buildings Directive 2010/31/EC (EC, 2010; Erlalelitepe et
al., 2011). Over 55 years, literally hundreds of building energy programs have been developed,
enhanced, and are used throughout the building community. The whole-building energy simulation
programs are the core tools in the building energy field. These programs provide users with key
building performance indicators, such as energy use and demand, temperature, humidity, and costs
(Crawley et al., 2005). In measuring building performances, various mathematical calculation models
have been utilized. These models include a wide range of studies (i.e., from simplified statistical
models to specialized physical simulations) (Schlueter and Thesseling, 2009).

Nowadays, most of analyses in building energy systems are based on the energy conservation
principle. The classical exergy analysis method enables to pinpoint the location, to understand the
causes of inefficiencies, and to establish the true magnitude of waste and losses. Exergy analysis is,
therefore, an important and efficient tool for the design of systems due to providing the designer with
answers to two important questions of where and why the losses occur (Costeexergy, 2013).

Exergy may be defined in various ways, as listed in Ref. (Hepbasli, 2012). According to one definition,
it is the maximum theoretical useful work (shaft work or electrical work) obtainable as the system is
brought into complete thermodynamic equilibrium with the thermodynamic environment while the
system interacts with this environment only (Tsatsaronis, 2007). The environmental benefits and
economics of energy technologies may also be identified better than energy through exergy. In this
context, exergy should be utilized by engineers and scientists, as well as decision and policy makers,
involved in green energy and technologies in tandem with other objectives and constraints (Rosen et
al., 2008).

In the literature, various attempts have been made towards developing energetic and exergetic indices
(or indexes or indicators or metrics) for assessing and comparing the performance of buildings. In
this regard, Olofsson et al. (2004) studied on how to rate the energy performance of buildings. They
presented a brief review of recent approaches to energy rating while different strategies of setting
energy efficiency standards were discussed. Entrop et al. (2010) compared the methodologies and
accuracies of three Dutch energy performance indicators and applied them to to eight houses. They

640
linearly correlated the actual domestic energy use with the estimated energy consumption given by
the energy performance indicators, being 7–25% lower.
Greensfelder et al. (2010) investigated building performance metrics that can be used by building
owners, energy managers, and operators to track the performance of commercial facilities. They
prepared a report and aimed at clarifying and simplifying the options for promoting market adoption.
This report included the methods and findings of this research, a discussion on important topics
related to selecting, comparing, and tracking metrics, lists of recommended. As far as exergy-based
performance indices are concerned, their number is relatively low compared to the energy-based
ones, as will be seen in the following section. The main objectives of the present contribution are
twofold, namely (i) to present available exergetic indices for buildings, and (ii) to develop and propose
some new ones.

SOME ENERGETIC AND EXERGETIC INDICES FOR BUILDINGS

Baselines and Normalizing the Indexes
Baselines are used to compare a building with internal data sets to understand how the building is
performing. This term has been widely used in industry for defining the performance of a building
before energy efficient upgrades are established. The main aim behind this is to measure the impact
of utility-sponsored projects/programs. In creating the baseline data, past measured performance of a
building is considered to understand if building performance is improving, degrading, or constant. For
this purpose, historic utility and metric data should be stored for future comparisons. In this regard,
properly normalizing the data is very essential for making proper comparisons. In most cases, the
only normalization needed will be based on climate, but it may also be necessary to normalize data
for changing operational patterns or building physical characteristics (Greensfelder et al., 2010).

In the following, some energy and exergy-related indices used to assess the performance of various
buildings are briefly described while their main relations are presented.

Energetic Indices
a) Energy utilization (or use or unit ) index (EUI)
This is a widely used metric expressing the total energy consumption for each building. It is defined as
the total energy use normalized to the floor area. In other words, each energy type is converted to a
common unit (kJ) for comparison and calculation of total energy consumed. The EUI is expressed in
kJ/m2∙yr and can be used to compare energy consumption relative to similar building types or to track
consumption from year to year in the same building. The EUI is calculated by converting annual
consumption of all fuels to kJ and then dividing by the gross square meter of the building. It can be a
good indicator of the relative potential for energy savings. A comparatively low EUI indicates less
potential for large energy savings. The metric can be studied based on statistical averages, medians,
percentiles etc. Analyses with median metrics prove to be successful, because of less sensitivity to
individual buildings (Sharp, 1996; Thumann and Younger, 2003; Olofsson et al., 2004).

b) Energy indicators developed for the Netherlands

Three indicators have been used in the Netherlands to express the energy performance of buildings,
as listed in Table 1. These indicators are evaluated in MJ/m2∙yr.

Table 1. Various energy performance indices for the Netherlands (Adapted from Entrop et al. (2010)).
Type of energy
performance Relation used Factors Eq.
indice no
Energy Etotal , EPC C1, C2: numerical
Performance EPC = correction factor (330
Coefficient for CFEPC (C1 Atgs , EPC + C2 Atts, EPC ) 2
MJ/m , 65 MJ/m )
2
(1)
new buildings CFEPC: correction factor to
regulated in 1995 fit past EPC results
Old Energy Index Etotal , EI ,old Atts, EI ,old C3
for existing EI old = C3, C4, C5: numerical
buildings C4 A2 tts, EI ,old + C5 Etotal , EI old + Atgs , EI ,old correction factors 0.13, 56
2 (2)
voluntary (MJ/m ), 0.06
proposed in 2000
New Energy Index Etotal , EI , new C6, C7, C8: numerical
for existing EI new = correction factors 155
buildings C6 Atgs, EI , new + C7 Atts, EI , new + C8 2 2
(MJ/m ), 106 (MJ/m ) and (3)
regulated in 2008 9560 (MJ).

641
c) Categorizing energy metrics
Within the framework of the building performance tracking project, funded by the California Energy
Commission, various building performance metrics were investigated. Based on a comprehensive
literature review and phone interviews with professionals from the building controls industry,
performance metrics were divided into four categories:, namely (i) basic energy metrics, (ii) advanced
energy metrics, (iii) basic system metrics, and (iv) advanced system metrics (Greensfelder et al.,
2010).

Table 2 lists some basic energy and system metrics for the whole building. Advanced metrics tables
include more detailed metrics, providing a more in-depth picture of building performance. the full list of
metrics may be found in Ref. (Greensfelder et al., 2010).

Table 2. Some basic energy and system metrics for the whole building (Adapted from Greensfelder et al.
(2010)).
Type of Sub- Metric Units
metric system (Indicator)
2
Energy use kJ/m ·yr
NA 2
Energy use kJ/annual operating hours·m ·yr
Basic 2
Electric Energy use kWh/m ·yr
energy Gas Energy use
2 2
kJ/m ·yr or therms/m ·yr
2
Electric Utility cost $/m ·yr
2
Gas Utility cost $/m ·yr
% Hours uncomfortable % hours outside of space temperature
Basic NA range
system Comfort index Average comfort index (0-100)

Basic energy metrics: These utilize the whole building utility meter data and are relatively easy to
track, and should be tracked by all buildings.

Advanced energy metrics: These utilize the whole building utility meter data, but require more
advanced analysis than basic energy metrics, or metrics that require additional meters beyond whole
building energy use meters.

Basic system metrics: These utilize existing points from the building automation system (BAS) or
maintenance management system while they are relatively easy to track.
Advanced system metrics: These require additional points to be added to the BAS or maintenance
management system while this category also includes metrics that use both meter data and BAS data.

d) Primary energy ratio

The primary energy ratio (PER) is defined as the ratio of delivered useful energy (thermal, electric,
mechanic) to the supplied primary fossil energy (derived from natural gas, fuel oil or coal), as
expressed below (Gonçalves et al., 2013):

Qheating
PER = (4)
E primary

where Qheating is the heating load.

e) Building energy index

The building energy index (BEI) was developed by McLean et al. (2011) as an overall performance
indicator for building design strategy. It consists of the climate related and climate unrelated energy
loads, which are respectively derived from the climate energy index (CEI) and benchmark data for
nonspace conditioning energy uses. The CEI is calculated based on psychrometric chart. It calculates
the energy required to condition any weather hourly ordinate to the nearest boundary of a human
comfort zone and operates on a unitary air flow. The BEI may be obtained from (McLean et al., 2011)

CEI [(kWh / yr ) /(m 3 / h )] Vbd [m 3 / h ]

BEI = + NBL [kWh / m 2 yr ] (5)
A floor [m 2 ]
where Vbd is the building design airflow rate and NBL is the normalized benchmark load for actual
occupancy hours (normalized non-space conditioning benchmark load).

642
f) Normalized performance indice
The normalized performance indice (NPI) is defined as the ratio of the annual building energy use to
the floor area. Two separate indices are calculated; one for electricity and the other for fossil fuels
while additionally, CO2 and cost performance indexes are used. The NPI is also normalised for
weather and exposure, indicating the actual performance of the building (Canbay et al., 2004). For
example, performance assessment (yardstick) for fossil fuels and electricity in the supermarkets are
given as follows: Fossil fuels: <160, 160–290 and >290 Electricity: < 670, 670–920 and >920 for
low, medium and high consumptions in kWh/m2 (EEO, 1994).

Exergetic Indices
Exergy analysis has the potential to optimize building energy demand, since exergy consumption can,
to a certain extent, be minimized. As an initial step of exergy application for building design, a
reference environment of buildings needs to be properly defined (Sakulpipatsin, 2009).

a) Total exergy system efficiency index

This is defined as the ratio of the exergy demand of the room to the total exergy load of the room as
follows, while its detailed calculation has been given in more detail elsewhere (Schlueter and
Thesseling, 2009):

Ex room
ε system = (6)
Extotal

In other words, dividing the remaining exergy that leaves the room through the envelope by the total
exergy load of the room results in an exergy system efficiency index. It displays the ratio of exergy that
is actually used to heat the room. The exergy demand of the room is estimated by multiplying the heat
demand with the quality factor of the room. The quality factor is estimated by the Carnot efficiency
using the outside and inside temperature of the room. The total exergy load indicates the total amount
of exergy necessary to supply the building and depends on building construction, geometry and
system selection. Optimization can be achieved by balancing between these three different fields
(Schlueter and Thesseling, 2009).

b) Primary exergy ratio

This is defined as the useful exergy demand rate (exergy heat demand rate at room) to the primary
exergy input rate, as given below (Gonçalves et al., 2013):

E x
PExR =  room (7)
Ex primary

c) Exergy flexibility factor

It is defined as the ratio of the exergy demand of the heating subsystem (i.e., radiator, direct electric
heating, wall heating, radiating heating, slab heating) to the total exergy input of the building, as given
below:

E x HS
ExF = (8)

Extotal ,building

which expresses the possibility of replacing a given subsystem by any other subsystem. The flexibility
factor elaborates on the notion that not all subsystems are appropriate for meeting all types of exergy
demand while it changes between zero and one. A relatively large flexibility factor expresses a higher
ability of a system to operate with a broader range of energy sources (Sakulpipatsin et al., 2006;
Hepbasli, 2012):

d) Exergy index of sustainability

This index (ExSI) is calculated in terms of exergy index of renewability ( α ex ), which is defined as the
ratio of the annual solar exergy on the building footprint to the total annualized exergy lost due to the
building construction and operation as follows (El Shenawy and Zmeureanu, 2013):

643
 100
ExSI = (9)
1 + exp [ −λ .(α ex − 50)]

with
Annual available solar exergy
α ex = .100 (10)
Annual exergy lost

In the α ex approach, the exergy efficiency of 100% of harvesting the solar exergy is considered.
However, the theoretical potential is reduced by losses associated with the conversion from the
primary source to the secondary resource. The value of parameter λ in Eq. (9) indicates the strength
of the policy implemented for achieving sustainability in the building sector; this parameter determines
the slope and spread for the relationship between the sustainability index ExSI and the renewability
index α ex . The value of λ , between 0 and 1, is set by the developer of rating scale based on local or
national goals, market penetration of technologies and shareholders surveys. The proposed rating
scale assesses the building sustainability in five categories in terms of the ExSI as follows:
Sustainable (96%≤ ExSI ≤100%), exergy efficient (75%≤ ExSI ≤96%), average exergy efficient (25%
≤ExSI≤75%), less than average exergy efficient (4% ≤ExSI ≤25%), and unsustainable (0% ≤ExSI
≤4%) (El Shenawy and Zmeureanu, 2013).

DEVELOPING SOME NEW EXERGETIC INDICES FOR BUILDINGS

As can be seen from the developed and proposed exergy based-indices, the exergy load of the room
considered is considered and this is compared to the whole building exergy load or the heating sub-
system`s exergy rate. As far as the EXSI is concerned, there exists a categorization. In terms of
energy-based indices, yearly energy utilization (fuel or electrical) or energy cost is compared to the
building floor area.

In this context, there may be various ways of developing new exergetic indices. One way would be to
modify some energy-based indices and hence to proposed the corresponding exergy-based ones. So,
specific exergy utilization index (SExUI) may be defined for fuel and electricity as follows:

Ex yr , fuel
SExUI fuel = (11a)
Abuilding

Ex yr , elec E yr , elec
SExUI elec = = (11b)
Abuilding Abuilding

where Ex yr , fuel is exergy of the yearly fuel consumption in kJ/yr. The exergy (or chemical exergy) of
fuel should be calculated. Based approximation method for fuel chemical exergy, the following
relations are widely used (Hepbasli, 2008).

ExCH
β LHV = (12)
LHV
Ex
β HHV = CH (13)
HHV

where β is the proportionality constant (or quality factor or exergy coefficient), while LHV and HHV
denote the lower heating (net caloric) and higher heating (gross caloric) values, respectively. Because
electrical energy is not affected by ambient conditions, it is equivalent in work and can be treated as
totally convertible to work. Exergy content of electricity equals its energy content (Hepbasli, 2008).
Table 3 lists typical chemical exergy content of some fuels (Kotas,1995; Ayres, 2005; Hepbasli, 2008).

644
Table 3. Typical chemical exergy content of some fuels
(Adapted from Refs. (Kotas,1995; Ayres, 2005; Hepbasli, 2008)).
Type of fuel βLHV βLHV
(Ayres, 2005) (Kotas, 1995)
Coal 1.088
Different types of coal 1.06-1.10
Different fuel oils and 1.04-1.08
petrol
Coke 1.06 1.05
Fuel oil 1.073
Natural gas 1.04 1.04±0.5%
Diesel fuel 1.07
Fuelwood 1.15
Wood 1.15-1.30
Geothermal βHHV=0.29

Fuel and electricity utilization may be combined under the term of the normalized SExUI (NSExUI) as
follows:

Ex yr , fuel + Ex yr , elec
NSExUI = (14)
Abuilding

An Illustrative Example

In the following, an illustrative example of a shopping center in Izmir, Turkey is presented to indicate
how the proposed NSExUI will be applied to this center. One should note that the values given here
are taken from Canbay et al. (2004) and have not been updated.

Total energy consumption (kWh) 2,958,216 for LPG 4,363,104 for electricity
Normalized energy use (kWh) 3,918,688 6,735,929
Floor area (m2) 16,000 16,000
NPI (kWh/m2) 245 421

In this regard, the performance of the shopping center considered is assessed using the NPI values of
245 and 421 kWh/m2 based on the yardsticks given earlier. As a result, the building electricity and
fossil fuel (LPG) consumption values are defined to be medium level and low level, respectively.

Using Eqs. (11a) and (14), and taking β = 1.04 from Table 3, the values for the SExUI and NSExUI
are calculated to be 254.71 and 675.71 kWh/m2, respectively. The SExUI value seems to be very
close to the NPI. As expected, this is due to the factor of β = 1.04. To make a comparison between
various buildings, this may be utilized, but one also should consider a categorization in terms of
various levels, such as low, medium and high levels, by producing different yardsticks (comparative
reference values). This needs a detailed work to determine them. There may also be another
approach to determining yardsticks. For example, the values of the exergy-based indicators may be
scored with various percentage values while they may be transformed to the scale of 0 to 10.

CONCLUSIONS
Performance indices give a measure of the energy/exergy utilization of a building, which can be
compared to various energy and exergy-based yardsticks. They can present the potential for possible
improvements while they can also be used to indicate progress over the time. These indices enable
us to compare buildings in a group or estate.

We have reviewed various energy and exergy-based indices for evaluating the performance of
buildings in this study. We have also developed and proposed two exergy-based indices.

We can extract the following concluding remarks from this study:

a) The number of energy-based indices is higher than that of the exergy-based ones.
b) Some countries have developed energy-based indicators by considering their own needs. For
similar buildings operated under similar conditions, to decide which building is
more ‘low-energy’ than another building is less difficult. To generalize these for other countries, a
normalization is needed by considering various factors, such as weather, exposure.

645
c) The proposed exergetic indices here may also be combined with exergetic prices of various
energy sources and CO2-emissions.
d) Considering some exergy-based concepts, such as exergy audit (Yasni and Carrington, 1988) and
low exergy (Hepbasli, 2012), exergy management and becoming an increasingly important aspect
of exergy, one should develop more exergy-based indices.
e) As highlighted elsewhere (Hepbasli, 2012), the concept of energy active buildings may be
extended and improved through the concept of the exergy active building. In this context, it is
recommended that Leadership in Energy, Exergy & Environmental Design (LEExED) could be
established by modifying Leadership in Energy & Environmental Design (LEED).
f) The most effective way of assessing the performance of a building seems to apply the low-exergy
approach and to utilize low-exergy based indices because all the building components are
assessed from the primary energy production to the building envelope.

NOMENCLATURE
A area, m2
BEI building energy index, kWh/m2∙yr
C numerical correction factor, MJ/m2
CF correction factor,-
E characteristic yearly energy use of the house (or building), MJ/yr
EI climate energy index, (kWh/yr)/( m3/h)
EPC energy performance coefficient, MJ/m2∙yr
Ex exergy amount, kJ
ExF exergy flexibility factor,-
ExSI exergy index of sustainability, -
EUI energy utilization (or use or unit ) index, kJ/m2∙yr
HHV higher heating value, kJ/kg
HV heating value, kJ/kg
LHV lower heating value, kJ/kg
NBL normalized benchmark load for actual occupancy hour, kJ/m2∙yr
NSExUI normalized specific exergy utilization index, kJ//m2∙yr
PER primary energy ratio, -
PExR primary exergy ratio, -
Q heating load, kJ
SExUI specific exergy utilization index, kJ//m2∙yr

V volumetric air flow rate, m3/h

Greek Letters
α index of renewability, -
β proportionality constant or quality factor or exergy coefficient, -
Ɛ exergy efficiency, -

Subscripts
bd building design
CH chemical
EI energy index
elec electrical
EPC energy performance coefficient
ex exergy
HHV higher heating value
HS heating system
LHV lower heating value
tgs total ground surface
tts total thermal transmission surface
yr year

REFERENCES
Ayres R. U. 2005. Mass, exergy, efficiency in the US economy, Interim Report, Approved by L.
Hordijk, International Institute for Applied Systems Analysis, Austria, July 26, 2005.
Canbay C. S, Hepbasli A. and G. Gokcen. 2004. Evaluating performance indices of a shopping centre
and implementing HVAC control principles to minimize energy usage. Energy and Buildings
36(6):587-598.
Costexergy. 2008. Analysis and Design of Innovative Systems for Low-Exergy in the Built
Environment, Available in <http://www.costexergy.eu/background_2.html> (Access date: 20 May
2013).

646
Crawley D. B, Hand J. W, Kummert M. and B. T. Griffith. 2005. Contrasting the Capabilities of Building
Energy Performance Simulation Programs. A Joint Report , Version 1.0, 55 pages.
EC, European Commission. 2010. Directive 2010/31/EC of the European Parliament and of the
Council on the Energy Performance of Buildings.
EC, European Commission. 2012. Directorate-General for Energy, Directorate C - Renewables,
Research and Innovation, Energy Efficiency, "Consultation Paper Financial Support for Energy
Efficiency in Buildings", C.3 - Energy Efficiency, Brussels.
Energy Efficiency Office (EEO), Department of the Environment, Introduction to Energy Efficiency in
Shops and Stores, Best Practice Programme, UK, 1994.
El Shenawy, A., and R. Zmeureanu. 2013. Exergy-based index for assessing the building
sustainability. Building and Environment 60:202-210.
Entrop A. G, Brouwer H. J. H. and A. H. M. E. Reinders. 2010. Evaluation of energy performance
indicators and financial aspects of energy saving techniques in residential real estate. Energy and
Buildings 42:618–629.
Erlalelitepe I, Ekmen K. E, Turhan C, Akdemir M, Akkurt G. G. and T. Kazanasmaz. 2011. Energy
performance of residential buildings and their architectural configuration. Low-Energy Architecture
(LEA), World Renewable Energy Congress, 8-13 May, Linköping, Sweden.
Gonçalves P, Gaspar A. R. and M. G. de Silva. 2013. Comparative energy and exergy performance of
heating options in buildings under different climatic conditions. Energy and Buildings 61:288-297.
Greensfelder E, Friedman H. and E. Crowe. 2010. Building Performance Tracking in Large
Commercial Buildings: Tools and Strategies, Subtask 4.4 Research Report: Characterization of
Building Performance Metrics Tracking Methodologies, California Energy Commission, Available in
<http://www.cacx.org/PIER/documents/Subtask_4-2_Report.pdf> (Access date: 25 May 2013).
Hepbasli A. 2008. A study on estimating the energetic and exergetic prices of various residential
energy sources. Energy and Buildings 40(3):308-315.
Hepbasli A. 2012. Low exergy (LowEx) heating and cooling systems for sustainable buildings and
societies. Renewable and Sustainable Energy Reviews 16:73–104.
Kotas T. J. 1995. The Exergy Method of Thermal Plant Analysis. Krieger Publishing Company,
Malabar, Florida.
McLean D, Roderick Y, Quincey R. and D. McEwan. 2011. Climate energy index and building energy
index: New indices to assess and benchmark building energy performance. Proceedings of Building
Simulation 2011: 12th Conference of International Building Performance Simulation Association,
Sydney, 14-16 November, 792-799.
Olofsson T, Meier A. and R. Lamberts. 2004. Rating the energy performance of buildings. The
International Journal of Low Energy and Sustainable Buildings 3, Available in <http://lnu.diva-
portal.org/smash/get/diva2:208041/FULLTEXT01> (Access date: 26 May 2013).
Rosen M. A, Dincer I. and M. Kanoglu. 2008. Role of exergy in increasing efficiency and sustainability
and reducing environmental impact. Energy Policy 36:128-137.
Sakulpipatsin P, Bezuijen R, Schmidt D. and E. Boelman. 2006. Energy and exergy analysis applied
to the built environment. Technology and Innovation for Sustainable Development Conference
(TISD2006). Faculty of Engineering, Khon Kaen University, Thailand, 25-27 January.
Sakulpipatsin P, Itard L. C. M, Van Der Kooi H. J, Luscuere P. G. and E. C. Boelman. 2009. A
reference environment for the calculation of the exergy value of air in buildings. ASHRAE Transactions
CH-09-002:10-15.
Sharp T. 1996. Energy benchmarking in commercial office buildings. Proceedings of the 1996 ACE3
Summer Study 4:321-329.
Schlueter A. and F. Thesseling. 2009. Building information model based energy/exergy performance
assessment in early design stages. Automation in Construction 18:153-163.
Tsatsaronis G. 2007. Definitions and nomenclature in exergy analysis and exergoeconomics. Energy
32:249–253.
Thumann A. and W. J. Younger. 2003. Handbook of Energy Audits. The Fairmont Press, Inc. Marcel
Dekker, Inc. Sixth Edition. Lilburn, Georgia New York and Basel.
Yasni E. and C. G. Carrington. 1988. Off-design exergy audit of a thermal power station. J. Eng. Gas
Turbines Power 110(2):166-172.

647
EXERGETIC SUSTAINABILITY ANALYSIS OF A SINGLE LAYER SOLAR DRYING PROCESS
,*
A. Midilli and H.Kucuk
Energy Division, Department of Mechanical Engineering, Faculty of Engineering, Recep Tayyip Erdoğan University,
53100 Rize, Turkey
adnan.midilli@erdogan.edu.tr
haydar.kucuk@erdogan.edu.tr

ABSTRACT
The main objective of this study is to develop the exergetic sustainability indicators for a single layer solar drying
process of the shelled and unshelled pistachio samples. In this regards, the exergetic sustainability indicators such as
exergetic efficiency, waste exergy ratio, exergy recoverability ratio, exergy destruction ratio, environmental impact
factor, and exergetic sustainability index are defined and estimated by depending on the experimental data obtained
from the single layer solar drying process. Accordingly, it is noticed that the rise of waste exergy ratio decreases the
exergetic efficiency and exergetic sustainability index while increasing environmental impact factor.

Keywords: Exergetic sustainability, Single layer drying, Solar drying, Environmental impact factor, Exergetic
sustainability index

INTRODUCTION
Energy is a key element in the interactions between nature and society and considered a key input for the
environment and sustainable development. During the past decade environmental and sustainability issues have
become major burdens to overcome since they have caused regional and global consequences (Dincer, 2007).
The supply and utilization of low-priced and clean fuel is particularly significant for environmental sustainability as well
as social, economic and institutional sustainability since energy plays a vital role in industrial and technological
developments around the world (Dincer, 1999; Midilli et al., 2005a; Midilli et al., 2005b). Exergy analysis is considered
an essential tool to expose the impacts of a power generating and consuming devices on exergy-based sustainability.
On the other hand, sustainability is necessary to overcome current ecological, economic, and developmental
problems (Dincer and Rosen, 2007). Also, sustainability concerns often lead local and national authorities to
incorporate environmental considerations into energy planning. In this regard, sustainable development demands a
sustainable energy source that, in the long term, is readily and sustainably available at reasonable cost and can be
utilized for all required tasks without causing negative societal impacts (Midilli and Dincer, 2009).
In terms of exergy analysis of the single layer solar drying processes, Midilli and Kucuk (2003) examined the energy
and exergy analyses of the drying process of shelled and unshelled pistachios using a solar drying cabinet. Midilli and
Dincer (2009) developed some new exergy-based parameters, such as exergetic efficiency, exergetic stability factor,
environmental benign index and exergetic sustainability index for PEM fuel cells in order to study how some of their
operating aspects and system characteristics affect the environment and sustainability. Also, Midilli et al. (2012)
presents some new exergy-based parameters for Recirculating Aquaculture System (RAS) for Black Sea trout
(Salmotruttalabrax). In this regard, the exergetic parameters, such as exergetic efficiency, waste exergy ratio, exergy
recoverability ratio, exergy destruction ratio, environmental impact factor, and exergetic sustainability index are
proposed and investigated. This paper presents the effects of some exergetic sustainability indicators on the
performance of the single layer solar drying processes of the shelled and unshelled pistachio samples.

ANALYSIS
To parametrically investigate the environmental and sustainability aspects of the single layer solar drying processes in
terms of exergy, Solar drying process was considered to be a steady-flow process. The experimental set up, shown in
Fig. 1, consists of a drying cabinet, solar air collector, and auxiliary heater and circulation fan. Drying cabinet is 55 ×
92 × 117 cm3 and has 16 shelves, which were made of wire-mesh-bottomed trays. The 4.5 m2 solar air collector
(collectors area =2 × (1.25 × 1.8 m2)) was connected to solar drying cabinet, and was used when solar radiation was
intense. The circulation fan to supply fresh air has a power of 0.37 kW (Midilli, 2001; Midilli and Kucuk, 2003).

Experimental Data
In order to estimate the exergetic sustainability indicators, the required experimental data have been taken from the
literature (Midilli, 2001; Midilli and Kucuk, 2003). Moreover, all details related to the experiments can be found in the
literature (Midilli and Kucuk, 2003)

648
Fig.1. Solar assisted drying cupboard (1. Fan, 2.Valve, 3.Connection pipe, 4.Orifice, 5.Auxiliary heater, 6.Temperature controller, 7.Inlet of solar air
collector, 8.Solar air collector, 8-1.Glass cover, 8-2.Glass cover, 8-3.Absorber plate, 8-4.Insulation, 8-5.Wood cover, Pyranometer, 10.Outlet of solar air
collector, 11.Chimney, 12.Outlet of drying cupboard, Inter section, 14.Shelves, 15.Inlet of drying cupboard, 16.Support of drying cupboard, 17.Flexible
connection pipe, 18.Support of solar air collector, 19. Manometer) (Midilli, 2001; Midilli and Kucuk, 2003)

Parameters
General exergy balance of solar drying processescan be written as
(Exergy Input) = (Exergy output) + (Irreversibility) (1)
However, if total exergy output and irreversibilities are clearly taken into consideration, the exergy balance for the
solar drying processes can be written as
(Total exergy input) = (Total exergyused in shelves) + (Total exergy loss) + (Total exergy destruction) (2)
which can be arranged as

∑ E x SDP
in = ∑ E x LSDP + ∑ E x dSDP + ∑ E x eus
SDP
(3)

where

∑ E x SDP
d = ∑ E xinput
SDP
, f + ∑ Ex input , sol − ∑ Ex dci
 SDP  SDP (4)

∑ E x SDP
in = ∑ E xinput
SDP
, f + ∑ Ex input , sol
 SDP (5)

∑ E x SDP
L = ∑ E x SDP dco (6)

where ∑ E x SDP
input , f describes inlet exergy from the fan, ∑ E x SDP
input , sol ; gained exergy from solar air collector..
If Eq.(4-6) are replaced into Eq.(3)

∑ E x SDP
eus = ∑ E x dci
SDP
− ∑ E x dco
SDP
(7)

where ∑ E x SDP
dci describes inlet exergy to the drying cabinet, ∑ E x SDP
dco ; outlet exergy from drying cabinet.
Under this solid theoretical background, the exergy-based environmental and sustainability parameters previously
presented in the literature (Midilli and Dincer, 2009; Midilli et al., 2012) are considered and rearranged for the solar
drying processes as
i) Exergetic efficiency
ii) Waste exergy ratio
iii) Exergy recoverability ratio
iv) Exergy destruction ratio
v) Environmental impact factor

649
vi) Exergetic sustainability index

Exergetic efficiency (η )
SDP
ex
In order to define the exergetic efficiency for a solar drying processes, total exergy entering from fan and solar air
collector and totalexergy used in shelves should be taken into consideration. If so, the exergetic efficiency of the solar
drying processes can be rearranged as in the following algebraic form, ranging from 0 to 1 theoretically (Midilli and
Dincer, 2009; Midilli et al., 2012).

η exSDP = ∑ E xeus
SDP
/ E xinSDP = (∑ E x dci
SDP
− ∑ E x dco
SDP
)/ (∑ E xinput
SDP
, f + ∑ Ex input , sol )
 SDP (8)

Waste exergy ratio (r ) SDP

wex
During the solar drying processes, exergy losses occur in the solar air collector and drying cabinet. Under these
considerations, waste exergy ratio can be written as follows, ranging from 0 to 1 (Midilli and Dincer, 2009; Midilli et al.,
2012).

SDP
rwex = (∑ E x SDP
d )
+ ∑ E x LSDP / ∑ E xinSDP
= (∑ E x ) (∑ E x )
(9)
input , f + ∑ Exinput , sol − ∑ Ex dci + ∑ Ex dco / + ∑ E xinput
SDP  SDP  SDP  SDP SDP SDP
input , f , sol

Exergy Recoverability Ratio (r ) SDP

exr
Exergy recoverability ratio indicates the exergy potential that is possible to be recovered in the system. Exergy
recoverability ratio of a solar drying process can be assumed as the ratio of the exergy recovery potential for the
system to total exergy input. The exergy recoverability ratio can be written as in the following form, ranging from 0 to
1 (Midilli and Dincer, 2009; Midilli et al., 2012).

SDP
rexr = ∑ E x rSDP / ∑ E xinSDP = E x dco
SDP
/ (∑ E x SDP
input , f + ∑ E xinput
SDP
, sol ) (10)

where E x r
SDP
describes the quantity of exergy recovery potential for the system.

Exergy destruction ratio (r )

SDP
exd
Exergy destruction ratio is an important parameter to indicate the decrease of the positive impact of the solar drying
processes on exergetic sustainability. The reference value of this factor should be ‘‘zero’’ for better
exergyeticsustainability of the solar drying processes. Exergy destruction ratio is given in the following form, ranging
from 0 to 1 (Midilli and Dincer, 2009; Midilli et al., 2012).

SDP
rexd = ∑ E x dSDP / ∑ E xinSDP
= (∑E x SDP + E x SDP
input , f ∑ input , sol − ∑ E x dci
SDP
/ ) (∑ E x SDP
input , f + ∑ E xinput
SDP
, sol ) (11)

= 1 − [∑ E x SDP
dci / (∑ E x SDP
input , f + ∑ E xinput
SDP
, sol )]
Environmental Impact Factor (f SDP
envi )
Environmental impact factor of solar drying processes is an important parameter to indicate whether or not it
damages the environment because of its unusable waste exergy output and exergy destruction. Environmental
impact factor (Midilli and Dincer, 2009; Midilli et al., 2012) can be rearranged as in the following form, ranging from 0
to +∞ (Midilli and Dincer, 2009; Midilli et al., 2012.)

1
SDP
f envi = rwex
SDP
×
η SDP
ex
(12)

Exergetic sustainability index (Θ ) SDP

esi
Exergetic sustainability index is an important parameter for exergetic sustainability of the solar drying processes in
terms of the second-law of thermodynamics. The exergetic sustainability index can be written to be reverse of the
environmental impact factor, ranging from 0 to 1 theoretically (Midilli and Dincer, 2009; Midilli et al., 2012).
1
Θ esi
SDP
= SDP
f envi (13)

650
RESULTS AND DISCUSSION
This paper presents the effects of some exergetic sustainability indicators on the performance of a single layer solar
drying process by using the experimental data in the literature (Midilli and Kucuk, 2003). For this purpose, the
following parameters which are investigated based on the constant characteristics of a solar drying process are taken
into consideration: i) Exergetic efficiency, ii) waste exergy ratio, iii) exergy recoverability ratio, iv)exergy destruction
ratio, v) environmental impact factor vi) exergetic sustainability index.
Fig. 2 presents the variation of exergetic efficiency and waste exergy ratio as a function of time (h) based on the
characteristics of a single layer solar drying process. The exergetic efficiency which is one of the most important
parameters for exergetic sustainability of the solar drying process is mainly based on the total exergy input (total
exergy entering from fan and solar air collector)and the total exergy used in the drying cabinet. It is noticed that the
exergetic efficiency highly affected by the amount of solar radiation. In such a way that, when the amount of solar
radiation increase total exergy used in the drying cabinet increases thus this causes the increase of exergetic
efficiency of the solar drying process. It can be assumed that waste exergy ratio is an important parameter to refer to
the necessity of waste energy management of the solar drying process in order to increase its exergetic performance
and its role on exergetic sustainability. As shown in Fig. 2, exergetic efficiency decreases while waste exergy ratio
increases for single layer solar drying process. The waste exergy ratio of solar drying process is function of exergy
losses occur in the solar air collector and drying cabinet and total exergy entering from fan and solar air collector. The
results show that exergetic efficiency and waste exergy ratio range from 0 to 0.410 and from 0.590 to 1, respectively
(their theoretical values range from 0 to 1 (Midilli and Dincer, 2009; Midilli et al., 2012).

1,0 0,5

0,9
0,4

Exergetic efficiency
0,8
Waste exergy ratio

0,3

0,7

0,2
0,6

Waste exergy ratio 0,1

0,5 Exergetic efficiency

0,4 0,0
0 1 2 3 4 5 6
Time (h)
Fig.2. Variation of exergetic efficiency and waste exergy ratio with time.

Fig. 3 shows the variation of exergetic efficiency and exergy destruction ratio as a function of time based on the
constant characteristics of the solar drying process. Exergy destruction ratio that is an important parameter to affect
the exergetic efficiency and waste exergy ratio of the single layer solar drying process should be taken into
consideration to indicate its role on exergetic sustainability. As shown in Fig. 3, when the exergy destruction goes up,
the exergetic efficiency decreases. Therefore, in order to minimize the exergy destruction ratio and increase the
contribution to exergetic sustainability of the solar drying process exergy loses occurred in the solar air collector
should be minimized. Thus, the contribution to exergetic sustainability of the solar drying process will be remarkable.
In terms of the second-law of thermodynamics, all outputs become meaningful for the single layer solar drying
process because they have potential to generate entropy in the environment. In terms of the first-law of
thermodynamics, the environmental effect of the heat loses from solar air collector and drying cabinet outlet is not
taken into account. It is estimated that exergy destruction ratio ranges from 0.510 to 0.922 (its theoretical values
range from 0 to 1 (Midilli and Dincer, 2009; Midilli et al., 2012).

651
 0,5 1,0

0,4 0,9

exergy destruction ratio

Exergetic efficiency

0,3 0,8

Exergetic efficiency
0,2 Exergy destruction ratio 0,7

0,1 0,6

0,0 0,5
0 1 2 3 4 5 6
Time (h)
Fig.3. Variation of exergetic efficiency and exergy destruction ratio with time.

0,18 1,0

0,16
Exergy recoverability ratio 0,9
Exergy destruction ratio
Exergy recoverability ratio

Exergy destruction ratio

0,14
0,8

0,12

0,7
0,10

0,6
0,08

0,06 0,5
0 1 2 3 4 5 6
Time (h)
Fig.4. Variation of exergy recoverability ratio and exergy destruction ratio with time.

Fig. 4 presents the variation of exergy recoverability and exergy destructionratios as a function of time based on the
constant characteristics of the single layer solar drying process. Exergy recoverability ratio indicates the exergy
potential that is possible to be recovered in the solar drying process. Exergy recoverability ratio of a solar drying
process can be assumed as the ratio of the exergy recovery potential for the system to total exergy input (total exergy
entering from fan and solar air collector). Outlet exergy of the drying cabinet is a potential exergy which can be used
for preheating of drying air for this system. It is determined that exergy recoverability ratio ranges from 0.074 to 0.171
(its theoretical values range from 0 to 1 (Midilli and Dincer, 2009; Midilli et al., 2012). Exergy destruction occurs in the
solar air collector which it is relating with structure of collector. It is noticed that the exergy destruction ratio highly
affected by the amount of solar radiation. It is seen that, when the amount of solar radiation increases exergy
destruction ratio decreases.

652
 Exergetic sustainability index 0,8 100

0,6

Environmental impact factor

0,4 10
Exergetic sustainability index
Environmental impact factor

0,2

0,0 1
0 1 2 3 4 5 6
Time (h)
Fig.5. Variation of exergetic sustainability index and environmental impact factor with time.

Fig. 5 presents quantitatively the relationships between exergetic sustainability index and environmental impact factor
as function of time. Exergetic sustainability index and the environmental impact factor are a function of waste exergy
ratio and exergetic efficiency. Environmental impact factor of single solar drying process is an important parameter to
indicate whether or not it damages the environment because of its waste exergy output and exergy destruction. As
seen in Fig. 5, the exergetic sustainability index and the environmental impact factor are highly affected by the
amount of solar radiation. When the amount of solar radiation increases exergetic sustainability index increases and
environmental impact factor decreases. It is estimated that exergetic sustainability index and environmental impact
factor range from 0 to 0.694 and from 1.441 to ∞, respectively (their theoretical values range from 0 to 1 and from 0 to
∞, respectively (Midilli and Dincer, 2009; Midilli et al., 2012).
Fig. 6 shows the variation of exergetic sustainability index and environmental impact factor as function of exergetic
efficiency. As shown in Fig. 6, as the exergetic efficiency increases the exergetic sustainability index increases
and the environmental impact factor decreases. In other words, increased exergy efficiency reduces environmental
damage for a single layer solar drying process. If total waste exergy output from the solar drying process increases,
exergy-based sustainability ability of the system decreases and thus environmental impact factor goes up, whereas
exergetic sustainability index decreases. As exergetic efficiency is zero value, the exergetic sustainability index has a
∞ value, whereas environmental impact factor is zero. Accordingly, it can be said that at the end of the single layer
solar drying process which amount of solar radiation has minimum value this system will not be an exergetic and
environmental benign system. However, as amount of solar radiation has maximum value environmental impact
factor has minimum value and exergetic sustainability index has maximum value. As a result, it can be said that for
higher amount of solar radiation this system is an exergetic and environmental benign system. Thus, to decrease the
environmental impact factor and to increase exergetic sustainability index of the solar drying process, exergy
destruction occurred in the solar air collector should be minimized and the exergy recovery potential from the system
should be optimized.

0,7 100

0,6 Exergetic sustainability index

Environmental impact factor
Exergetic sustainability index

Environmental impact factor

0,5

0,4
10
0,3

0,2

0,1

0,0 1
0,0 0,1 0,2 0,3 0,4
Exergetic efficiency
Fig.6. Variation of exergetic sustainability index and environmental impact factor with exergetic efficiency.

653
CONCLUSIONS
Environmental and sustainability aspects of a single layer solar drying process have been parametrically studied
based on the actual data. Some concluding remarks are given as follows:
• The exergetic efficiency and the exergy recoverability ratios range from 0to 0.410 and from 0.074 to 0.171,
respectively while exergy destruction and waste exergy ratios range from 0.510 to 0.922 and from 0.590 to
1, respectively.
• The single layer solar drying process has some high values of waste exergy ratio and exergy destruction
ratio, and it needs improvement for better environment. In this regard, exergy destruction occurred in the
solar air collector should be minimized and the exergy recovery potential from the system should be
optimized.

NOMENCLATURE
E x Exergy, kW
Greek Letters
η efficiency
Subscripts
d destruction
dci drying cabinet inlet
dco drying cabinet outlet
envi environmental
esd exergy destruction
esi exergetic sustainability index
eus exergy used in shelves
ex exergy
exr exergy recoverability
f fan
L loss
sol solar
wex waste exergy
Superscripts
SDP solar drying process

REFERENCES
Dincer, I. 2007. Environmental and sustainability aspects of hydrogen and fuel cell systems. International Journal
of Energy Research 31:29-55.
Dincer, I. 1999. Environmental impacts of energy. Energy Policy 27:845-854.
Midilli, A., M. Ay, I. Dincer and M. A. Rosen. 2005a. On hydrogen and hydrogen energy strategies: I: current status
and needs. Renewable and Sustainable Energy Reviews, 9(3):255-271.
Midilli, A., M. Ay, I. Dincer and M. A. Rosen. 2005b. On hydrogen and hydrogen energy strategies II: future
projections affecting global stability and unrest. Renewable and Sustainable Energy Reviews 9(3):273-287.
Dincer, I. and M. A. Rosen. 2007. Exergy, Energy, Environment and Sustainable Development, UK: Elsevier.
Midilli, A. and I. Dincer. 2009. Development of some exergetic parameters for PEM fuel cells for measuring
environmental impact and sustainability. International Journal of Hydrogen Energy 34(9):3858-3872.
Midilli, A. and H. Kucuk. 2003. Energy and exergyanalyses of solar drying process of pistachio. Energy 28(6):539-
556.
Midilli, A. 2001. Determination of pistachio drying behavior and conditions in a solar drying system. International
Journal of Energy Research 25:715-725.
Midilli, A., H. Kucuk and I. Dincer. 2012. Environmental and sustainability aspects of a recirculating aquaculture
system. Environmental Progress & Sustainable Energy 31(4):604-611.

654
 EXERGY ANALYSIS OF WIND TURBINE SYSTEM: A CASE STUDY OF BERGAMA, IZMIR

1 2
I. Kökey , A.Turgut
1
Dokuz Eylul University, The Graduate School of Natural and Applied Sciences,
Mechanical Engineering Department, Tinaztepe Campus, Buca, Izmir-Turkey
2
Dokuz Eylul University, Engineering Faculty, Mechanical Engineering Department,
Tinaztepe Campus, Buca, Izmir-Turkey
iskenderkokey@hotmail.com
alpaslan.turgut@deu.edu.tr

ABSTRACT

In this paper, energetic and exergetic efficiencies of a wind turbine system are investigated. Selected wind turbine
(GE’s 2.5MW) is located in Bergama, 110 km far from Izmir. The data used for calculations are taken from Scada
system of the turbine and from the local weather station, during the month of October 2011 till September 2012.
Energy and exergy efficiencies exhibit relative differences between %60 and %80 at studied wind speeds. It is
clear from the results that the exergy analysis is more preferable tool for planning wind power systems.

Keywords: Energy, exergy, energy efficiency, exergy efficiency, second law efficiency

1. INTRODUCTION
Energy is an essential factor for development and economic growth of the countries. It is expected that global
energy demand will grow -by 36% between 2011 and 2030- even with the significant efficiency gains. For the
developing countries energy demand will rise 61% by 2030 compared to 2011. World primary energy consumption
by fuel (Mtoe) types in 2011 is given in Table1 [1].
According to the table, renewable energy consumption will grow more than 3 times from 1.5% in 2011 to 5.2% in
2030 out of the total consumption.

As one of the developing countries, Turkey's energy demand has increased rapidly over the last few years
(Figure1) [2]

Table 1. Energy consumption by fuel types on the world

Sources 2011(MTOE) 2020(MTOE) 2030(MTOE)

Liquid 4,059.1 4,407.0 4,760.9
Natural Gas 2,905,6 3,633.2 4,251.6
Coal 3,724.3 4,454.1 4,701.0
Nuclear 599.3 733.7 9,67.9
Hydro 791.5 995.7 1,156.5
Renewables 194.8 470.6 878.1
Total 12,274.6 14,694.3 16716.0
MTOE: Million tonnes oil equivalent

655
Fig 1. Energy supply and demand in Turkey

It is expected that installed power capacity of Turkey will reach to 110,000-130,000 MW and 500 billion kwh
electricity consumption until 2023. It is planned to supply 30% of this from renewable energy and with 20,000 MW
installed capacity wind energy will be the most important part of it.

By the end of 2012 total installed wind power capacity of Turkey is 2,312 MW and it is the 4% percent of the total
installed power capacity (Table 2) [3]. The evaluation of this capacity by years is given in Figure2.

Table 2. Installed sources in Turkey

Sources Installed power (MW)

Hydraulic 19,619.7
Natural gas 18,317.8
Coal 14,853.6
Wind 2,312.0
Liquid 1,229.2
Others 455.7
Total 56,788.0

656
Fig 2. Total installed capacity in Turkey

The growth potential of investing in the renewable energy market attracts both commercial and scientific research.
One of these research areas is an exergy analysis of the renewable energy systems. Exergy analysis is a tool that
can provide a clearer view of the energy losses and thereby help us to improve the efficiency of the systems [4].

Koreneos et al.[5] was one of the firsts who applied exergy analysis to wind energy by only taking into account the
electricity generation of wind turbines. Sahin et al.[6], developed a new exergy formulation for wind energy by
considering the enthalpy and entropy. They showed the differences between energy and exergy efficiencies a
function of wind speed and concluded that differences between efficiencies could be significant. Redha et al.[7]
studied the energy and exergy analysis of a wind power turbine according to the local weather condition of
Sharjah, UAE. They pointed out that exergy efficiency is more precise to describe the irreversibility. Baskut et al.[8]
and Xydis [9] explained the effects of various meteorological parameters and air psychometrics on the exergy
analysis.

In this study, the energetic and exergetic efficiencies of a wind turbine system is investigated. Selected wind
turbine (GE’s 2.5MW) is located in Bergama, 110 km away from Izmir. The data used for calculations are taken
from Scada system of the turbine and from the local weather station, during the month of October 2011 till
September 2012.

2. GENERAL ENERGETIC AND EXERGETIC RELATIONS FOR WIND ENERGY SYSTEMS

a. Mass balance relations

A general expression for a conservation of mass of a control volume is given by

dmcv
= ∑ m in −∑ m out (2.1.a)
dt in out
where the left side represents the time rate of change of mass contained within the control volume, and the
subscripts in and out stand for inlet and outlet states, respectively. [10]
dmcv
In steady state, = 0 . Therefore Eq. (2.1.a) reduces to
dt
∑ m =∑ m
in
in
out
out (2.1.b)

657
 b. Energetic relations

The energy balance equation for a flow of matter through a system can be expressed as
dE cv   V2   V2 
= Qcv − Wcv + ∑ m in  h + + gz  −∑ m out  h + + gz  (2.2.a)
dt in  2 in out  2 out
dEcv
Where the is the time rate of change of energy, Q cv is the heat transfer rate over the boundary of the
dt
control volume, Wcv includes work effect, such as those associated with rotating shafts, displacement of the
boundary, and electrical effects, h is the specific enthalpy, V is the velocity and z is the elevation. [10]
dEcv
In steady state, = 0 . Therefore Eq. (2.2.a) reduces to
dt
 V2   V2 
Q cv − Wcv = ∑ m out  h + + gz  − ∑ m in  h + + gz  (2.2.b)
out  2 out in  2 in

Wind energy E is the kinetic energy of a flow of air mass m at a speed V. The mass of air which flows through the
studied cross section area is not easy to measure but it can be described in terms of volume V through its density
ρ = m / V . The volume (V) can be expressed as V=AL where A is cross section area perpendicular to the flow
and L is the horizontal distance. Physically, L=Vt and wind energy can be defined as [11]
1
E= ρAtV 3 (2.2.c)
2
 while the rate of momentum change is m (V1 − V2 ) ,
The mass of air passing through the rotor per unit time is m
V1 and V2 represent upwind and downwind speeds respectively at a considerable distance from the rotor. This
momentum change creates the resulting thrust of the wind turbine. The rate of extracted kinetic energy can be
expressed as [12]

W = m (V1 − V2 ).Vavg (2.2.d)

Vavg = V1 + V2
where
( ) while the rate of change in kinetic energy is
2
1
E = m (V1 − V2 ) = W
2 2
(2.2.e)
2
Eq. (2.2.d) and Eq. (2.2.e) are equal so Eq.(2.2.f) can be written as
V1 − Vavg = Vavg − V2 (2.2.f)
Assuming that the direction of the wind across the rotor is axial and the velocity is uniform throughout the swept
area A. Then, the power extracted by the rotor is: P = ρA(V1 − V2 )Vavg
2
and this equation can be reduced

ρAV13
P= (α + 1)(1 − α 2 ) (2.2.g)
4
where α = V2 V and the term (α + 1)(1 − α 2 ) is called “power coefficient” (C ). Differentiation shows that the
p
1

power P is maximum when α = 13 , when the final wind velocity V2 is equal to one third of the upwind velocity V1.
An ideal windmill could extract 0.593 of the power in the wind when the α = 1 . This result is known as “Betz
3
Law” [4].

658
Figure 3. Power coefficient distributions as a function of α = V2 V
1

c. Exergetic relations

The exergy of a system is defined as the maximum shaft work can be done by the composite of the system and a
specified reference environment[13]. The reference environment is assumed to be infinite, in equilibrium, and to
enclose all other systems. Typically, environment is defined by stating its temperature, pressure and chemical
composition. In this paper, the environment is regarded to be a simple compressible that is large in extent and
( ) ( )
uniform in temperature T 0 and pressure P 0 .The values for both T 0 and P 0 used throughout a particular
analysis are taken as typical ambient conditions such as 1 atm and 25oC (298 K) [14].
When a system of interest is at T 0 , P 0 and at rest relative to the environment, we say the system is at the dead
state. At dead state, there can be no interaction between system and environment, and thus no potential for
developing work.
The exergy balance equation for a flow of matter through a system in steady-state can be expressed as [10]
∑ E x −∑ E x
in out= ∑ E xdest (2.3.a)

E xheat − E xwork + E x − E x
mass , in mass , out = E xdest (2.3.b)
where
 T 
E xheat = ∑ 1 − 0 Q k (2.3.c)
 Tk 
E xwork = W (2.3.d)

E x = ∑ m ψ
mass , in in in (2.3.e)

E xmass , out = ∑ m outψ out (2.3.f)

ψ = ex = (h − h0 ) − T0 ( s − s0 ) (2.3.g)
where Q k is the heat transfer rate through the boundary at temperature Tk at location k and W is the work rate.
The exergy of fluid currents is entirely due to the kinetic energy of fluid. The maximum work is extracted from a
moving fluid when the velocity is brought to zero relative to the reference state.

659
 1
ψ f = ρV 3 (2.3.f)
2
where ρ is the density of the wind and V is the wind speed.

d. Energy and exergy efficiency relations

The expressions for energy (η) and exergy (ε) efficiencies for the wind turbines can be defined as [15]

useful work output Wout

η= = (2.4.a)
total energy input Wwind ,in

useful work output Wout

ε= = (2.4.b)
total exergy input E xin

3. CASE STUDY

Based on the theory above and in order to evaluate the wind energy and exergy potential, a database taken
between November 2011 and November 2012 from SCADA system of a wind turbine and local weather station is
considered. The wind turbine and the weather station are both located in Bergama that is 110km away from Izmir.

A wind turbine of 2.5MW rating power from GE is selected for this study to evaluate its second law efficiency for
climatic conditions logged by SCADA system and weather station. Selected wind turbine has 106m. rotor diameter
and 85m hub height. The cut in and cut out speeds are 3m/s and 25m/s respectively. Considered data of wind
speed and generated power are taken from SCADA system of wind turbine as daily mean value.

A weather station of Turkish State Meteorological Service located in Bergama is selected for taking data of
ambience temperature and relative humidity for considered time period.

4. RESULTS AND DISCUSSIONS

In this paper, the measured data of generated power taken from SCADA system are used to obtain a power curve.
The distribution of measured power data (kW) is shown in Figure 4. as a function of wind speed (m/s). The power
distribution is seen to exhibit two main types of behavior, depending on wind speed. At low wind speeds, the
power increases with wind speed until the rated power wind speed is reached.

Figure 4. Measured output power distribution as a function of wind speed

660
For the first case, at low wind speeds, a three point exponential function can be obtained by using nonlinear
regression model. A curve is fitted between 3m/s and 12m/s regarding the cut in and rated power wind speed
respectively and its coefficient of determination (R2) is estimated as 0.9529. Fitted curve is obtained as a three
a + bx + cx 2
point exponential function can be described as y = e . Calculated parameters are obtained as a=3.60185,
b=0.67912, c=-0.02756 and standard errors are 0.11327, 0.0268, 0.00154 respectively as shown in Figure 5.

At high wind speeds (above 12 m/s), the power generation tends to stay flat with increasing wind speed. In the
rated wind speed region, an exponential function can be obtained as y = y0 + Ae R0 x and its R2 value is estimated
as 0.23452. Calculated parameters are obtained as y0=2507.42674, A=-1.93411e10, R0=-1.55971 as shown in
Figure 6.

Figure 5. Fitted curve as a function of wind speed between 3m/s and 12m/s

Figure 6. Fitted curve as a function of wind speed above 12m/s

661
Power curve as a function of measured wind speed above 3m/s, cut-in speed, can be obtained by considering
Figure 5 and Figure 6 as shown in Figure 7. Power curve can be described as a function of wind speed in two
cases which are between 3-12m/s and above 12m/s, also curves can be obtained as a three point exponential
equation and exponential equation respectively as shown in Figure 7.

Figure 7. Fitted power curve as a function of measured wind speed

All exergy efficiencies are calculated by using Equation (2.4.b) for a selected point and shown in Figure 8 is
obtained as 22 moving average value for Savitzky - Golay method[16]. The Savitzky - Golay is a statistical method
performs a polynomial regression to the data points in the moving average values for smoothing highly fluctuating
variables. In our cases, 22 moving average is considered to illustrate variability. The first month of studied period is
October and shows that efficiency tends to stay flat in followed 6 month. Same figure shows large variations and
large fluctuations in spring and summer season because of climatic conditions at that period of year in Bergama.

Figure 8. Mean daily exergy efficiencies as a function of time

Measured wind speed and ambient temperature are plotted as a function of time after obtained 10 and 35 moving
average value respectively for Savitzky - Golay method shown in Fig. 9 and Fig. 10.

662
Figure 9. Mean daily wind speed as a function of time

Figure 10. Mean daily ambient temperature as a function of time

Daily mean output power and wind speeds are logged from SCADA system while ambient temperatures are taken
from meteorological station for the same period. Then, the enthalpy and entropy parts of Equation (2.3.g) are
calculated, and the energy and exergy efficiencies are evaluated with Equation (2.4.b). Totally 6935 data values
are employed for whole analysis. Then, polynomial regression analysis as 3rd order is applied to wind speeds
above 5m/s level after estimated mean values. Energy and exergy efficiencies are calculated as a function of wind
speed as shown in Figure 11. Lowest values are observed at the minimum and maximum wind speeds while
highest efficiency values are observed around 6.5m/s. Energy and exergy efficiencies exhibit relative differences
between %60 and %80 at studied wind speeds. There, it is shown that relative differences between energy and
exergy efficiencies increase at high wind speeds. According to these results, exergy efficiency studies include
more information about environment and wind energy capacity and they are seen to be more useful for qualified
wind energy assessment projects then classical energy efficiency studies.

663
Figure 11. Variation of energy and exergy efficiencies, and percent difference between these values as a function of wind speed

5. CONCLUSIONS

In this paper, the energy and exergy efficiency of the GE2.5 wind turbine located in Bergama was studied. It is
clear from the results that exergy efficiency is a more precise tool for wind assessment studies. The formula which
includes thermodynamic components such as enthalpy and entropy is used to illustrate difference between energy
and exergy efficiency as a function of wind speed. It is suggested that exergy efficiency should be used for more
realistic and significant wind assessment and planning studies to develop wind farm projects.

6. REFERENCES

[1] British Petroleum (BP) Energy Outlook 2030, January 2013, BP

[2] World Bank, World Development Indicators, http://data.worldbank.org/data-catalog/world-development-indicators
[3] Turkish Electricity Transmission Co. Ankara (http://www.teias.gov.tr/Eng/StatisticalReports.aspx)
[4] Dincer I, Rosen MA. 2012. EXERGY Energy, Environment and Sustainable Development, Elsevier Science
[5] Koroneos C, Spachos N, Moussiopoulos N. 2003. Exergy analysis of renewable energy sources. Renewable
Energy 28:295–310.
[6] Sahin A.D, Dincer I, Rosen M.A, Thermodynamic analysis of wind energy, Int. J. Energy Res. 2006;30:553–566
[7] Baskut O, Ozgener O, Ozgener L. Effects of meteorological variables on exergetic efficiency of wind turbine
power plants. Renewable and Sustainable Energy Reviews 2010;14(9):3237-41
[8] Xydis G. Effects of air psychrometrics on the exergetic efficiency of a wind farm at a coastal mountainous site -
An experimental study , Energy 37 (2012) 632-638
[9] Adel Mohammed Redha, Ibrahim Dincer, Mohamed Gadalla, Thermodynamic performance assessment of
wind energy systems: An application, Energy 36 (2011) 4002-4010
[10] Bejan A, Tsatsaronis G, Moran M. Thermal design & optimization. New York: John Wiley & Sons, Inc.; 1996.
[11] Sahin AD, Dincer I, Rosen MA. New spatio-temporal wind exergy maps. Trans ASME J Energy Resour
Technol 2006;128(3):194–202
[12] Patel R. M., Wind and solar power systems design, analysis, and operation, second edition, Newyork:
Taylor&Francis, 2006
[13] Xiang JY, Cali M, Santarelli M. Calculation for physical and chemical exergy of flows in systems elaborating
mixed-phase flows and a case study in an IRSOFC plant. Int J Energy Res 2004;28:101–15
[14] Dincer I, Cengel YA, Energy, Entropy and Exergy Concepts and Their Roles in Thermal Engineering, Entropy
2001, 3: 116-149.
[15] Dincer I, Rosen MA. Thermodynamic aspects of renewables and sustainable development. Renew Sustain
Energy Rev 2005;9:169–189.
[16]Savitzky, A.; Golay, M. J. E., Smoothing and differentiation of data by simplified least squares procedures,
Anal. Chem., 1964, 36: 1627-1639.

664
 ENERGY AND EXERGY ANALYSIS OF A PERLITE EXPANSION FURNACE

Mert Gürtürk , Hakan F. Oztop , Arif Hepbaslı

1 2 3

Department of Mechanical Education, Technical Education Faculty, Fırat University, 23100 Elazig,
1

Turkey.
Department of Mechanical Engineering,Technology Faculty, Fırat University, 23100 Elazig, Turkey.
2

Department of Energy Systems Engineering, Faculty of Engineering,Yaşar University, 35100 Izmir,

3

Turkey.
m.gurturk@gmail.com, hfoztop1@gmail.com , arifhepbasli@gmail.com

ABSTRACT

Utilization opportunities of perlite have made the product adaptable to numerous applications in the
construction, industrial, chemical, horticultural and petrochemical industries. Applications of the perlite include
filler, high and low temperature insulation, concrete aggregate, textured coatings, absorbent and carrier etc. Thus,
perlite seems to be a very important material. In this paper, thermodynamics analysis of a perlite expansion
furnace was performed. Both energy and exergy relations were derived for the considered system. Some
parameters, such as losses, irreversibility and design etc., were used for determining the energy and exergy
efficiencies. Based on the results of the analysis of the system, the energy and exergy efficiencies are calculated
to be 66% and 26%, respectively. Some recommendations were also made towards increasing the efficiency in
the Turkish perlite industry.

Keywords: Energy, Exergy, Perlite, Perlite expansion furnace.

INTRODUCTION
Energy consumers have to responsibility cognitively energy efficiency, energy saving, environmental and
sustainable concepts etc. Many processes in engineering applications occur under ideal conditions according to
Thermodynamics Laws, but some of the processes are not efficient due to insufficient thermal design. The First
Law of Thermodynamics implies the conservation of energy (Cengel and Boles, 2006). The energy audit and
saving were applied to dry type cement rotary kiln systems consuming intensive energy. Cement production is an
energy intensive process, consuming about 4 GJ per ton of cement product. Energy loss was found from hot flue
gas, kiln shell and cooler stack, and some possible ways to recover the heat losses were introduced and
discussed. Finding results showed that 15.6% of the total input energy (4 MW) could be recovered (Engin and Ari.
2005). Additionally, various energy efficiency studies have been conducted at industries that consume particularly
intensive energy. Liu et al. (1995) studied a cement industry in China to determine the prospects renovation and
they improved its energy efficiency. Various scenarios that are high-cost, state-of-the-art precalciner kilns,
moderate-cost advanced vertical kilns, and low-cost vertical kilns without advanced technology, were considered.
They discussed the cost, energy intensities and environmental implications of these scenarios. In another study,
an energy intensive industry in Turkey was investigated (Ates and Durakbasa. 2012). Gutierrez et al. (2013)
studied the energy and exergy assessments of a lime shaft kiln. In their study, the energy efficiency was found to
be higher than the exergy efficiency, as 71.6%. Their results showed that the most irreversible processes took
place in the kiln. Çamdali et al. (2004) performed energy and exergy analyses of a rotary burner with pre-
calcination in cement production. Although the energy efficiency was found to be about %97, the exergy efficiency
was 64.4%. Utlu et al. (2006) made energy and exergy analyses of a raw mill and raw materials preparation unit
in a cement plant in Turkey using the actual operation data. Both energy and exergy efficiencies of the raw mill
were investigated for the plant performance analysis and improvement, and were determined to be 84.3% and
25.2%, respectively. Hepbasli and Ozalp (2003) investigated the development of industrial energy efficiency and
management studies in Turkey. They reported that the Turkish industrial sector had an annual energy saving
potential of approximately 30% and some efforts towards improving the energy efficiency in the country should be
made. Jegla et al. (2000) studied on a plant energy saving through efficient retrofit of the furnaces. The furnace
retrofit procedure was based on an advanced furnace integration approach using some principles of pinch
analysis and considering furnace limitations. This method combined principles of an effective design of both
processes and equipment. They applied an efficient methodology for furnaces retrofit, using optimization of both
stack temperature and air preheating system. They indicated that an advantage of this approach was
demonstrated through a case study - retrofit of furnace in petrol hydrogenation refining plant for energy efficiency
improvement. Sakamoto et al. (1999) studied on estimating the energy consumption for each process in the
Japanese steel industry. Their paper exhibited that the energy consumption was estimated by a statistical process
to evaluate the possibility of reducing the energy consumption. The specific energy consumption for each product
was estimated and also for crude steel produced from an integrated steel plant route and an electric arc furnace
route. The specific energy consumption was also compared. Madlool et al. (2013) reviewed energy utilization and
savings in cement industries. Their categorization for cement manufacturing included four key processes, namely
dry, semi-dry, semi-wet and wet processes. Their study covered energy saving, carbon dioxide emission
reductions and various technologies used to improve the energy efficiency in the cement industry. Energy
efficiency measures for raw materials preparation, clinker production, products and feedstock changes, general

665
energy efficiency measures, and finish grinding were surveyed. The largest recorded amounts of thermal energy
savings, electrical energy savings and emission reductions to date were 3.4 GJ/t, 35 kWh/t and 212.54 kgCO2/t,
respectively.

The main objective of this study is determine energy and exergy efficiencies of a perlite expansion
furnace. The physical properties of perlite, data gathering and perlite expansion system will be described first. A
thermodynamics analysis of the system considered will be then made. Finally, the results obtained will be
presented in this paper.

A brief description of the perlite

Perlite is a vitreous substance that contains 2–6% water. It has gray, silver gray, dark brown and black colors.
Perlite is a lightweight material because it can expand 10–30 times of its volume when heated between 800 and
o
1150 C. Density of perlite decreases as its volume increases. In Turkey, 8 billion tons of perlite exist and equal to
70% of the world reserves. This shows that the perlite is a very important material for Turkish economy, and it can
be used as insulator owing to its low heat conductivity (Topçu and Işıkdağ. 2007). Physical properties of the
expanded perlite are shown in Table 1. Perlite used in the furnace includes moisture less than 0.5% and water
5%.

Table 1. Physical properties of expanded perlite (Topçu and Işıkdağ, 2007).

Color White
Melting point 1300 oC
Specific heat 0.8371 kJ/kg∙oC
Unit weight 2.2-2.4 g/cm3
Rough density 30-190 kg/m3
Heat conductivity 0.039-0.046 W/m o∙C
Sound insulating 18 db (125 Hz)

Energy policy of Turkey is to improve the energy efficiency in buildings and many policies have been implemented
in this context. One of these policies is to take some measures towards reducing energy losses in buildings. The
insulation materials have been a very important point in insulation applications. Various particle sizes are
available to meet specific customer requirements. Various utilization areas of the perlite make the product
adaptable to numerous applications in the construction, industrial, chemical, horticultural and petrochemical
industries. Applications of the perlite include filler, high and low temperature insulation, concrete aggregate,
textured coatings, absorbent and carrier etc. The mentioned perlite expansion furnace is used in the plaster
industry for application of the plaster in this paper. The produced perlite can be used in plaster applications, so
that it must be white colored and not hold products of combustion. There are many applications of the plaster, of
which one is decoration of building in architecture. Color quality and health are the most important parameters in
applications of decoration. Therefore, in the production of the perlite and the plaster, natural gas with relatively
higher combustion efficiency is used. This energy source is expensive per unit volume compared to many other
countries due to its importation. Because perlite producing companies aim at a sustainable production in Turkey,
they make some efforts to reduce fuel consumption and protect the environment.

Description of the perlite expanded furnace and data gathering

Energy analysis and fuel-air ratio were determined. The gases, air and natural gas, are burned in a combustor,
and the occurring flame in the combustor rises to up the heating zone. The mass flow rate of the perlite is 0.121
kg/s. Four holes are placed on the cylindrical furnace. The perlite and the flame mix in the heating zone occurring
expansion, so the perlite expansion process starts in the heating zone. As shown in Fig. 1, the heating zone
consists of four parts, which are the upper and lower cones, the cylindrical body and the combustor. The upper
and lower cones, which were made from the sheet metal, have the similar dimensions, with a total area of 1.24
2 2
m . There are four holes on the surface of the cylindrical body and its area is 2.53 m , also made from the sheet
metal. The chimney of the furnace was considered up to the cyclone, and its length and diameter are 20 m and
o
0.11 m, respectively. The temperature value inside the heating zone of the furnace is higher than 850 C and the
o
perlite expansion process starts from this section. It continues up to 850 C in the chimney. The air, the perlite and
o
the natural gas enter the control volume under the environmental conditions (1 atm and 0 C) and all sections from
the combustor to the cyclone are considered as a control volume, as shown in Fig. 1. The temperature surface of
o
the furnace was measured as 800 C and the temperatures of the expanded perlite and the exhaust gas were
o o
measured as 600 C at the outlet of the control volume. This temperature value is 140 C in the filter system, while
the temperature values of the filter and the cyclone systems are not included. The temperature values for the
expanded perlite, the exhaust gases, the air and the natural gas were taken from the control unit or the
automatisation unit of the factory, but the surface temperatures were measured by an infrared thermometer with a
o o o
range from -50 C to 1000 C. The surface temperature of the chimney was taken approximately 550 C. The
average daily production capacity is 7 ton of expansion perlite. This value can change depending on the
production capacity, Samples are taken from the cyclone tests in the laboratory of the factory to determine the
rate of the expanded perlite. There is no insulation anywhere in the control volume, thus heat losses from the
control volume are very high. A flow diagram along with mass balance and control volume is also illustrated in Fig.
1.

666
Fig. 1. The control volume and whole of the system.

The furnace is called as a vertical cylindrical furnace or named according to the intended use, which is the perlite
expansion furnace. When the perlite expansion process is considered to be as a system, the main components of
this system are the combustor, the heating zone, the chimney, the cyclone and the filter group. The combustor,
the heating zone, the cones and the chimney were chosen as the control volume and these components of the
control volume were studied using energy and exergy analysis. The flow rates of the air and the natural gas are
0.84 (kg/s) and 0.025 (kg/s), respectively. In the analysis, the air is assumed that it consists of 79% nitrogen (N2)
and 21% oxygen (O2) and the moisture in the air was neglected in the calculations. The composition of the natural
gas includes methane (CH4), ethane (C2H6), propane (C3H8), i-butane (C4H10), n-butane (C4H10), i-pentane
(C5H12), n-pentane (C5H12) and hexane (C6H14), but only methane that consists of 95% of the mass of the natural
gas, was considered as the combustor in the combustion equations for determining the products of the
combustion. Methane is the principal component of the natural gas. The control volume operates at steady state.
Combustion is complete. N2 is inert. Kinetic and potential energy effects are ignored. For a control volume at
steady state, the identity of the matter within the control volume changes continuously, but the total amount of
mass remains constant. At steady state, mass balance is given by
∑𝑖 𝑚̇𝑖 = ∑𝑒 𝑚̇𝑒 (1)

Air-fuel ratio is determined from Eq. (2) for the analysis of the combustion products (Bejan et al. 1996).
𝑚̇𝑓𝑢𝑒𝑙 𝑚̇𝑎𝑖𝑟 𝑛̇ 𝑝
� = 𝜆 𝑜𝑟 =1+𝜆 (2)
𝑀𝑓𝑢𝑒𝑙 𝑀𝑎𝑖𝑟 𝑛̇ 𝑎𝑖𝑟

where 𝜆 is found as 0.05. The molar analysis of the combustion products is shown in Table 2, which was obtained
from Eqs. (3) and (4) (Bejan, A., et al. 1996).
0.05𝐶𝐻4 + (0.21𝑂2 + 0.79𝑁2 ) ⇒ (1 + 0.05)�𝑥𝐶𝑂2 𝐶𝑂2 + 𝑥𝐻2𝑂 𝐻2 𝑂 + 𝑥𝑂2 𝑂2 + 𝑥𝑁2 𝑁2 � (3)
𝜆 2𝜆 0.21−2𝜆 0.79
𝑥𝐶𝑂2 = , 𝑥𝐻2𝑂 = , 𝑥𝑂2 = , 𝑥𝑁2 = (4)
1+𝜆 1+𝜆 1+𝜆 1+𝜆

Table 2. Mole fractions of the products of combustion.

Component N2 O2 CO2 H2O
Mole fraction 0.75 0.1 0.04 0.09

After applying the mass balance to the control volume, the energy balance is applied to the control volume. The
energy rate balance at steady state conditions is given by (Kotas, 1995).
1 1
0 = 𝑄̇𝑙 − 𝑊̇𝑐𝑣 + ∑𝑖 𝑚̇𝑖 �ℎ𝑖 + 𝑉𝑖2 + 𝑔𝑧𝑖 � − ∑𝑒 𝑚̇𝑒 �ℎ𝑒 + 𝑉𝑒2 + 𝑔𝑧𝑒 � (5)
2 2
After rearranging Eq. 6, one can write
0 = −𝑄̇𝑙 + 𝑚̇𝑎 ℎ𝑎 + 𝑚̇𝑓 ℎ𝑓 + 𝑚̇𝑝𝑒𝑟 ℎ𝑝𝑒𝑟 − 𝑚̇𝑝 ℎ𝑝 − 𝑚̇𝐸.𝑝𝑒𝑟 ℎ𝐸.𝑝𝑒𝑟 (6)

667
The enthalpies of the air, the fuel and the products of the combustion are calculated from Eqs. (7)- (9). The values
in these equations can be found in Refs. (Kotas, 1995; Bejan et al., 1996; Cengel and Boles, 2006).
ℎ𝑁2 ℎ𝑂2
ℎ𝑎 = �0.79 + 0.21 � (7)
𝑀𝑁2 𝑀𝑂2
ℎ𝑓 = 𝐿𝐶𝑉 (8)
Only methane that consists of 95% of the mass of natural gas was considered as the combustor in the
combustion equations for determining the mole fractions of the combustion products. But energy balance
included the Lower Heating Value of the natural gas with 45000 kJ/kg (Cengel and Boles, 2006).
ℎ𝑁2 ℎ𝑂2 ℎ𝐶𝑂2 ℎ𝐻2𝑂
ℎ𝑝 = �0.75 + 0.1 + 0.04 + 0.09 � (9)
𝑀𝑁2 𝑀𝑂2 𝑀𝐶𝑂2 𝑀𝐻2𝑂

The enthalpies of the air and the combustion products are determined at 273 K and 873 K. The entropy and
enthalpy values of O2, CO2, H2O (gas) and N2 in kJ/kmol are taken from (Cengel and Boles, 2006). Mk is the molar
mass in kg/kmol. The Entropies, enthalpies and molar masses are used to determine the entropies and enthalpies
in kJ/kg. The entropy difference of the perlite is calculated from Eq. (10).
𝑇
∆𝑠 = 𝐶𝑝 𝑙𝑛 � 𝑒 � (10)
𝑇𝑖
The values obtained are illustrated in Table 3.

Table 3. Enthalpies and entropies of the components entering and leaving the control volume.
Descriptions h(kJ/kg) h(kJ/kg) s(kJkg∙K) s(kJ/kg∙K)
(273K) (873K) (273K) (873K)
Natural gas 45000 - 0 -
N2 280.19 925.57 6.73 7.98
O2 245.56 840.59 6.31 7.45
CO2 189.38 813.92 4.77 5.94
H2O 497.41 1700.52 10.29 12.59
Perlite 226.863 725.463 ∆𝑠 =0.96

Heat losses occur by radiation and natural convection from the control volume while they are calculated from the
following relations (Cengel and Boles, 2006).
𝑔𝛽(𝑇 −𝑇 )𝐿3
𝑠 ∞
𝐺𝑟 = (11)
𝜐2
𝑅𝑎 = 𝐺𝑟𝑃𝑟 (12)
1� 2
0.387𝑅𝑎 6
𝑁𝑢 = �0.6 + 8 � (13)
�1+(0.559⁄𝑃𝑟 ) �27 �
𝑘𝑁𝑢
𝐻= (14)
𝐿
𝑄̇𝑙,𝑐 = 𝐻𝐴(𝑇𝑠 − 𝑇∞ ) (15)
𝑄̇𝑙,𝑟 = 𝜀𝜎𝐴(𝑇𝑠4 − 𝑇∞4 ) (16)
The total heat loss rate (𝑄̇𝑙 ) is determined as 472.83 kW. The value of 99 kW unaccounted in the previous study
of the researchers was added to the total heat loss in this study. The results of the energy balance are shown in
Table 4.

Table 4. Energy data for the control volume.

Description Result (kW) Description Result (kW)
Perlite 27.45 Expanded perlite 82.70
Air 230.46 Products of combustion 840.44
Natural gas 1125
Sensible heat of natural 13.65 Heat loss 472.83
gas
Total Input 1396 Total Output 1396

The total energy input rate is determined as 1396 kW and the total energy output rate with the expanded perlite
and the products of the combustion are determined as 928.3 kW. Energy analysis is concerned with the amount
of energy, but does not include the quality of energy and the direction of change of state. Energy analysis is
inadequate in this respect. The Second Law of Thermodynamics overcomes with concepts of entropy and exergy.
Exergy analysis of systems allows determining irreversibility and available energy (exergy) in the system. These
analyses produce efficiency of systems in terms of First and Second Law of Thermodynamics. At steady state,
the exergy balance is shown in the rate form as follows (Kotas,1995).
𝑇
0 = ∑𝑗 �1 − ∞� 𝑄̇𝑙 − 𝑊̇𝑐𝑣 + 𝐸̇ 𝑥𝑖 − 𝐸̇ 𝑥𝑒 − 𝐸̇ 𝑥𝐷 (17)
𝑇𝑠
Exergy rate of heat energy is calculated by (Cengel and Boles, 2006).
𝑇
�1 − ∞� 𝑄̇𝑙 = 321.88 𝑘𝑊 (18)
𝑇𝑠
Exergy of any material consists of physical and chemical exergies given by (Cengel and Boles, 2006).

668
𝐸̇ 𝑥 = 𝐸̇ 𝑥 𝑝ℎ + 𝐸̇ 𝑥 𝑐ℎ (19)
Natural gas, air and perlite enter into the control volume under environmental condition (ℎ𝑖 = ℎ∞ , 𝑠𝑖 = 𝑠∞ ), so their
physical exergies are zero (Bejan et al. 1996). Indeed, enthalpy consists of internal energy and flow work, which
occurs with the pressure difference, which is neglected for air and natural gas due because the pressure
difference could not be measured in this study.
𝑝ℎ
𝐸̇ 𝑥𝑘 = 𝑚̇((ℎ𝑖 − ℎ∞ ) − 𝑇∞ (𝑠𝑖 − 𝑠∞ )) (20)
𝑝ℎ 𝑝ℎ 𝑝ℎ 𝑝ℎ
𝐸̇ 𝑥𝑁𝐺 = 0, 𝐸̇ 𝑥𝑁2 = 0, 𝐸̇ 𝑥𝑂2 = 0, 𝐸̇ 𝑥𝑝𝑒𝑟 = 0 (21)
Exergy enters into the control volume with the chemical exergy of natural gas
𝐸̇ 𝑥𝑁𝑐ℎ2 = 0, 𝐸̇ 𝑥𝑂𝑐ℎ2 = 0 (22)
The chemical exergy rate of natural gas is determined as 1170 kW from Eqs. (22) and (23) (Kotas,1995; Bejan et
al., 1996).
𝐸̇ 𝑥𝑁𝐺
𝑐ℎ
= 𝑚̇𝑓 𝑒𝑁𝐺𝑐ℎ
(23)
𝑐ℎ
𝑒𝑁𝐺 = 𝜑(𝑁𝐶𝑉) (24)
where 𝜑 is 1.04 for natural gas (Kotas 1995). The chemical exergy of every component of combustion products is
evaluated using Eq. (25) (Bejan et al. 1996).
∑ 𝑥𝑘′ 𝑒̅𝑘𝑐ℎ + 𝑅�𝑇∞ ∑ 𝑥𝑘′ ln 𝑥𝑘′ (25)
Chemical and physical exergy rates of the control volume are shown in Table 5.

Table 5. Exergy data for the control volume.

Description Exergy rate (kW)
Substance 𝐸̇ 𝑥 𝑝ℎ 𝐸̇ 𝑥 𝑐ℎ 𝐸̇ 𝑥
Sate Inlet Exit Inlet Exit
Perlite 0 26.96 0 0 26.96
Natural gas 0 - 1170 - 1170
N2 0 197.3 0 1.17 198.47
O2 0 24.54 0 -0.34 24.2
CO2 - 10.55 - 0.39 10.94
H2O - 45.14 - 1.58 46.72

Exergy destruction rate is determined as 540.83 kW from the exergy balance rate. This value is very high due to
bad working conditions. Exergy destruction is 46.22% of the total exergy input. Energy efficiency of the control
volume is calculated from
𝐸̇𝑒
𝜂= (26)
𝐸̇𝑖
which is found as 66%. Exergy efficiency of the control volume is obtained from
𝐸̇𝑥𝑒
𝜂𝑒𝑥 = (27)
𝐸𝑥̇𝑖
which is determined to be 26%.

RESULTS AND DISCUSSION

Mass and energy balance relations are applied to the control volume for energy auditing, and the results obtained
are shown in Table 4. Heat losses are very high due to non-insulation. This value is equal to the amount of 37.46
kg of natural gas in one hour. The system runs 16 hours per day and 599.47 kg of natural gas is wasted daily. The
maximum energy output from the control volume occurs with the combustion gases. These gases having energy
are used for pre-heat application, namely for input parameters, which are perlite, air and natural gas. Assessment
of the current situation according to energy efficiency can be evaluated from different perspectives. Perlite
producer does not manufacture the perlite expansion furnaces, but the responsibility for the operation of the
furnaces belongs to them. This situation shows the inadequacy of the producer according to the energy
management. When the furnace manufacture is evaluated based on the obtained results, its technical
infrastructure is not sufficient. Exergy rate balance is shown in Table 5. Exergy destruction rate is determined
from the exergy rate balance as 540.83 kW. Natural gas, air and perlite are inlet parameters for the control
volume under environmental conditions (ℎ𝑖 = ℎ∞ , 𝑠𝑖 = 𝑠∞ ). Thus, their physical exergies are calculated as zero.
Exergy entering the control volume are due to the chemical exergy of the natural gas. Exergy efficiency value
shows that energy losses have very high values. When an insulation application is made on the control volume, it
leads to increasing both energy and exergy efficiencies. Energy and exergy analyses show that this case
increases the cost of production. The company should consider energy and exergy parameters in order to
achieve a sustainable production.

CONCLUSIONS
In this paper, we have performed thermodynamics analysis of a perlite expansion furnace. We have derived both
energy and exergy relations for the system considered. We have also included some parameters, such as losses,
irreversibility and design etc., for determining the energy and exergy efficiencies.
We may list the following concluding remarks:

669
 1. The values for energy and exergy losses are very high due to not applying insulation. Energy and exergy
loss rates are 468.34 and 321.88 kW, respectively. These values showed that if the insulation would be
made, a significant fuel saving could be achieved, as expected.
2. The energy and exergy efficiencies are found as 66% and 26%, respectively. Energy efficiency can be
increased by pre-heating applications of the primary air, natural gas or perlite. Energy and exergy losses
must be prevented by insulation applications. Additionally, a good controlling strategy should be applied
to the system.
3. Exergy analysis shows that exergy destruction rate is determined as 540.83 kW, corresponding to
46.22% of the total exergy input. The exergy destruction is reduced by increasing the energy and exergy
efficiencies.

Acknowledgement

Authors thank to Fırat University Project Support Unit with the project number TEKF.13.01 and Aralçi Company in
Elazig, Turkey for their valuable contribution to this work.

NOMENCLATURE
2
A Area, m
𝐶𝑝 Specific heat, kJ/kg∙K
𝐸̇ 𝑥 Exergy rate, kW
e Specific energy, kJ/kg
𝐸̇ Energy rate, kW
2
g Gravity, m/s
Gr Grashof number
h Specific enthalpy, kJ/kmol, kJ/kg
Heat convection coefficient, W/m ∙K
2
H
k Thermal conductivity, W/mK
LCV Low calorific value, kJ/kg
L Length, m
M Molar mass, kg/kmol
𝑚̇ Mass flow rate, kg/s
Nu Nusselt number
Pr Prandtl number
𝑄̇ Heat flow rate, kW
Ra Rayleigh number
𝑅� Universal gas constant, kJ/kmol∙K
s Specific entropy, kJ/kg∙K
V Velocity, m/s
𝑊̇ Work rate, kW
x Mole fraction
2 4
𝜎 Stefan-Boltzmann, W/m K
𝜀 Thermal diffusion

Greek Letters
𝜆 Air-fuel ratio
𝛽 Thermal expansion coefficient, 1/K
2
𝜐 Kinematic viscosity, m /s
𝜑 Ratio for industrial fuels
𝜂 Energy efficiency
𝜂𝑒𝑥 Exergy efficiency

Subscripts
a Air
cv Control volume
c Convection
D Destruction
e Exit
e.per Expanded perlite
f Fuel
i Inlet
k k-th product
l Loss
NG Natural gas
per Perlite
p Combustion products

670
r Radiation
s Surface
∞ Environment

Superscripts
ph Physical
ch Chemical

REFERENCES

Cengel, Y.A. and M.A. Boles. 2006. Thermodynamics: An Engineering Approach. New York: 5th ed., McGraw-
Hill.
Engin, T. and V. Ari. 2005. Energy auditing and recovery for dry type cement rotary kiln systems-A case study.
Energy Conversion & Management 46: 551–562.
Liu, F., M. Ross, S. Wang. 1995. Energy efficiency of Chine’s cement industry. Energy 20:669–681.
Ates, S.A. and N.M. Durakbasa. 2012. Evaluation of corporate energy management practices of energy intensive
in Turkey. Energy 45:81–91.
Gutierrez, A.S., J.B.C. Martinez and C. Vandecasteel. 2013. Energy and exergy assessments of a lime shaft kiln.
Applied Thermal Engineering 51:273–280.
Çamdali, Ü., A. Erişen, F. Çelen. 2004. Energy and exergy analyses in a rotary burner with pre-calcinations in
cement production. Energy Conversion & Management 45:3017–3031.
Utlu, Z., Z. Sogut, A. Hepbasli, Z. Oktay. 2006. Energy and exergy analyses of a raw mill in a cement production.
Applied Thermal Engineering 26:2479–2489.
Hepbasli A, Ozalp N. 2003. Development of energy efficiency and management implementation in the Turkish
industrial sector. Energy Conversion & Management 44:231–249.
Jegla Z, Stehlik P, Kohoutek J. 2000. Plant energy saving through efficient retrofit of furnaces. Applied Thermal
Engineering 20:1545–1560.
Sakamotoa Y, Tonookab Y, Yanagisawa Y. 1999. Estimation of energy consumption for each process in the
Japanese steel industry: a process analysis. Energy Conversion & Management 40:1129–1140.
Madlool NA, Saidur R, Rahim NA, Kamalisarvestani M. 2013. An overview of energy savings measures for
cement industries. Renewable and Sustainable Energy Rev 19:18–29.
Topçu, I.B. and B. Işıkdağ. 2007. Manufacture of high heat conductivity resistant clay bricks containing perlite.
Building and Environment 42:3540–3546.
Bejan, A., G. Tsatsaronis and M. Moran. 1996. Thermal Design and Optimization. New York: John Wiley & Sons,
Inc.
Kotas, T.J. 1995. The Exergy Method of Thermal Plant Analysis. Florida: Malabar, Krieger Publishing.

671
 MEASURING ENERGY EFFICIENCY OF A RESIDENTIAL AREA USING EXERGY ANALYSIS:
A COMPARISON OF MAVIŞEHIR I AND MAVIŞEHIR II, İZMIR

Yelda MERT *, Nicel SAYGIN

Department City and Regional Planning/ Yüzüncü Yıl University, Van, Turkey
Department City and Regional Planning/Izmir Institute of Technology, Urla, Izmir, Turkey
yeldamert@yyu.edu.tr, nicelsaygin@iyte.edu.tr

ABSTRACT

In this study, to understand how important the design strategies are with an urban planning and design point of
view, an exergy analysis of Mavişehir I and Mavişehir II is carried out based on buildings’ spatial properties,
shadow and wind effects and local climate. The exergy analysis is applied with a broad consideration covering the
heat loads of the separate buildings including the effects of the heating system, passive heating effects, and other
climatic properties. When we compare the exergy results of Mavişehir I and Mavişehir II, we found out that
Mavisehir II has less exergy by fuel and exergy load values. This may result from different design strategies as in
Mavisehir I, the same buildings and the same villas between residential buildings were located according to sea
view, whereas in Mavişehir II, there is a heterogeneous pattern with various forms and heights of the residential
buildings. It is also found that exergy efficiency of the area is 4.20% and the exergy flexibility factor is 19.3%.
Therefore, the planners and architects should understand and internalize the energy concern in making designs to
overcome such low efficiency values and to realize the potential which is indicated with high flexibility factor.

672
INTRODUCTION

The relating between the energy usage and city has an enormous importance since almost 30% of the total energy
consumption in the world arises from the residential areas [Balocco, 2004]. Also increasing sensitivity for the
global warming and pressure for reducing energy consumption because of the increasing energy prices. In
addition to these when the power of city planning on the land use decisions and applications it is easy to relate the
energy consumption and sustainability to the city planning. Besides, spatial planning evaluates the values/assets
of the city and the region from protection and use point of views.

The new arisen habits of the populations in the cities bring larger problems about energy and environment. These
problems cause some irreversible changes globally. For the solution of these problems urban design and planning
may play an important role. Design of urban areas may include ecological concern while having naturally
compatible and clean renewable energy oriented solutions.

Design of energy efficient residential unit, design of energy efficient residential blocks, design of energy efficient
districts and design of energy efficient cities are important steps that have been taken for the sake of energy
conservation. So, it is crucial that energy effectiveness must be taken into account from starting a single housing
design to the regional scale and all sub scales for sure with considering physical factors such as morphology,
density and microclimate properties. By considering energy effective design criteria in housing block design, the
energy efficiency citywide will be improved automatically (Hisarlıgil, 2009).

The planning decision process in Turkey is done mostly depending on the profitability and luxury points of views.
Whereas most of the important parameters like protection of the environment, lighting, orientation, air flows,
natural sources of energies are neglected. There are legislations ruling the system but these are ineffective and
insufficient. Such rules in the law are meaningless like the height of the building and road width relations from
natural climate conditions point of view. Some rules ignore the local properties or some general rules such as the
orientation of the buildings with respect to the solar energy in a country like Turkey in which there is a vast
opportunity for the solar energy.

The legal structure must be revised considering ecological and sustainable views and the national policies must
have tight bending rules and auditing. To control the success and progress of each plan, some local commissions
and governmental organizations must be formed and these organizations must educate the public and decision-
makers by preparing some guide books (Erçoşkun, 2007).

In Turkey after July 2000 with the implication of the TS825 Heat Insulation legislation the heating and cooling loads
of the buildings are started to be calculated and taken into consideration in the constructions. But, this is not
enough since to increase the energy efficiency and effectiveness in urban areas, (beside the heating and cooling
loads), the effect of the orientation and microclimatic effects has also to be taken into consideration.

As being an effective tool for identifying the energy consumption sources and the areas where the designer has to
strike in for better efficiency and effectiveness of energy usage, exergy analysis is the best choice for investigation
as Dinçer and Rosen pointed out (2007).

There are various studies in the global literature about energy efficiency and effectiveness of the heating, cooling,
lighting and transportation sectors. However, there are not any studies in the literature about the energy (exergy)
effectiveness and efficiency that covers a residential area or residential pattern completely.

When the literature is investigated it is seen that there are few studies about these considerations with a
comprehensive point of view. There are some studies only having exergy and /or energy analysis of the buildings
like the studies of (Balta et al. ,2011 and Depecker et al. 2001) in which different energy sources for a building and
the effect of shape for energy consumption is investigated. In another study (Barrerio et al., 2009) an investigation
about the energy efficiency of an area and proposes an integrated energy supply methodology is done that
investigates the relations between planning and efficiency. In some studies the exergy analysis is used for
investigating the urban area such as (Utlu and Hepbaşlı, 2007, Utlu and Hepbaşlı, 2005, Ozturk et al., 2004, Liu et
al., 2010, Balocco et al., 2004).

Local planning design fundamentals and energy effective planning decisions are effective and important on
regional scale decisions. The different properties of the spatial structure at different scales are important from
energy effectiveness point of view. Beside properties like orientation and microclimate at local scale is important in
wider spatial properties are also important in regional scale. At micro scales, direct forward changes bring
considerable improvements such as adjusting proper orientation for the sake of energy saving without having extra

673
cost to the construction. The physical structure of the settlement, the altitude and the closeness to the sea shore
directly affect the microclimates mainly since the land can absorb and desorb heat in a faster rate than the body of
water and the altitude directly affects the climate. The combination, location and texture that the houses form affect
the climate on regional base beside the orientation, geometric and physical properties. Moreover, the density of
the structures and building block especially dense and high buildings affect the climate greatly in urban areas. The
geometric shapes and the distances of buildings from other structures effect the wind directions and speed as well
as the air temperature. The wind speed is important since it affects the heat transfer mechanism and increases the
heating loads of the structures in winter and cooling loads in summer.

Depending on all of these parameters the physical properties such as the orientation and geometry as well as the
texture of the settlement is significant from energy efficiency point of view since they affect the climate in the urban
area.

In a previous study the Mavişehir I area is investigated from an exegetical point of view which is aimed to inform
planners about the importance of the energy/exergy concerns in the decision making phase of a planning process.
The effect of orientation, shadowing, insulation and heating systems are considered in the study. In order to reach
the main exergy loss points of a building and the urban area. Whereas, in this study a comparison for two different
stages of an urban area is done with an exergetic point of view.

CASE STUDY

Mavişehir area is in the boundaries of Karşıyaka Municipality at north of İzmir Bay (Figure 1-a). Metropolitan
Master Plan of İzmir defines this area as mass housing area with high-rises. Mavişehir mass housing area is
formed by three regions that were constructed in three stages by funding of the Housing Credit Bank and the area
is 270 ha.
The construction of the Mavişehir Stage-I was started in 1993 and took two years to complete. Mavişehir-I has two
urban blocks separated by the water canal; 12 apartment blocks at the western part, which are called Pamukkale
and eight apartment blocks at the eastern part, which are called Selçuk (Figure 1-b). In total there are 20
residential high-rises (2784 flats) and 88 dublex villas located in between the residential high-rises. Project has a
locational density of 174 houses/ha and population density of 664 person/ha [24]. It is aimed that every apartment
will have the sea view and the fresh air coming from the sea is nt blocked by adjusting the blocks perpendicular to
the sea shore. But the dublex villas that have positioned between the rectangular blocks are get caught between
them and have some troubles about privacy visually. Besides the villas have a great shadow problem because of
the high rise blocks .There are 2876 independent flats in total including 4 kindergartens in residential high-rises. 3
of the residential high-rises have 16 story, 10 of them have 18 story and 7 of them have 19 story. Each storey has
8 flats. 341 flats in the residential high-rises are one-bedroom with 69,6 m2 gross area, 341 flats in the residential
high-rises are two-bedrooms with 115,4 m2 gross area, 341 flats in the residential high-rises are three-bedrooms
with 148,5 m2 gross area, 1079 flats in the residential high-rises are three-bedrooms with 152 m2 gross area, 341
flats in the residential high-rises are four-bedrooms with 170,9 m2 gross area, 341 flats in the residential high-rises
are four-rooms with 176,3 m2 gross area and 88 dublex villas are three-rooms with 231,5 m2 gross area [24].

Fig. 1 a-Location of the Mavişehir in İzmir Bay [20] b- Location of the area of investigation [20]

Differing from Mavişehir I, Mavişehir II has different types of blocks. The various designs and the combination of
the high-rises obtain the dynamism of the built environment. The construction of the stage was started in 1995 and
was completed in 1997. Totally there are 3446 residential units in 109 residential high-rises. The names of high-
rises are Albatros, Flamingo, Kuğu, Kırlangıç and Turna. The Albatros consists of 10 high-rises with 22 storey.

674
Fig. 2 Plan of the Mavişehir area [Karşıyaka Municipality]

Flamingo blocks are 23 floors and there are 10 blocks. There is 1 Kuğu apartment building with 20 and 22 floors.
Kırlangıç and Turna buildings are combination of eight different plant ups with various storeys up to 13 floors.
Mavişehir III is located by the sea and designed as triplex villas totally 122. The thesis study concerns the first and
second stages of Mavişehir area. Every accommodation in the area can be assumed luxury residential high-rise
with having heating system, double glazing, sun blind, decorative coated doors, and double bathrooms.

EXERGY ANALYSIS OF A RESIDENTIAL AREA

The calculations for the exergy load of the urban areas is complicated and starts with the data handling which
covers five steps, that are;
• Size of the buildings
• Insulation of the buildings
• Orientation of the buildings
• Resident information
• Heating system properties
After completing the data handling exergy calculations are takes place;
• Calculation of the Shadow Effect Factor (SEF)
• Calculation of the exergy load of each building
• Calculation of the exergy loads for the urban area

Shadow Effect Factor (SEF) is an indication of a buildings blockage with the shadow of another building that is the
effect of the overlapping shadow on the building behind one another .

Some assumptions that will be applied for the analysis are;

• indoor air temperature of 21° C
• outdoor temperature is taken as 0° C.
• The heat transfer coefficients of the walls and doors will be taken from the real values depending on the
plan details of the buildings
• Also the number of the inhabitants at an average values of 4 are assumed during the calculations
• 70°/50° heating system (inlet/outlet temperature of the radiators) is used in calculations in accordance
with the reality and the condensing boiler systems.
• Heat loss from the exterior walls, roofs and from ground is calculated depending on the values taken from
the Karşıyaka Municipality depending on the reported data by the construction company.
The aim of the calculation procedure is to calculate the E xdemand values which show us the exergy need of building
that is calculated via the assumptions given above.

675
 E xdemand = E xloss − E xgain (10)

E xloss is calculated by the transmission losses through the doors, walls, windows and roofs. E x gain is a result of
Solar heat gains through windows which is a function of SEF also other gains that are arising from the axillary
equipment’s in the houses are taken in to consideration such as lighting ( 1 W/m2)
E xinput is calculated depending on the E xdemand by considering the efficiency of the heat production and heat
distribution systems which are used as 0.95 and 0.93 respectively
Exergy flexibility factor is calculated by Eq. 10, which is an indication of the possibility of replacing a given system
by another for meeting the exergy demand.
E xdemand
EFF = (11)
E xinput
In order to underline the effect of this calculations to the planners and residents the costs of this exergy demand is
also calculated. Exergetic Total Cost is the cost of the exergy input to the system by fuel.

In the calculations, the first and second stages of Mavişehir area are investigated. As in accordance with the
calculation algorithm first the size and various properties of the buildings and the area is gathered and tabulated in
the Table 1 given below.
Table 1 The surface area data for the buildings
Villa Villa
Building Albatros Flamingo Kuğu Kırlangıç Turna Selçuk Pamukkale
Selçuk Pamukkale

Exterior
2
wall [m ]
7832 4235 10003 18554 18554 7239 7239 1493 1493
Window
2
[m ] 822 504 1306 2378 2378 1463 1463 135.72 135.72
2
Door [m ] 602 726 1309 1584 1584 1178 1178 109.68 109.68
2
Roof [m ] 667 594 1044 2793 2793 1312 1312 1200 1200

Floors to
ground
2
[m ]
667 594 1044 2793 2793 1312 1312 1200 1200

For the calculation of the shadow effect factors 3-D models were developed and the previews for the summer and
winter periods for selected times are generated. Also a daily animation is created separately for each block group
and the shadow effect factors determined according to these visuals and the results are tabulated in the Table 2
and Figure 3.

676
Fig. 3 Some sample views from the model

The determination of the sides of the buildings in order to make the passive solar effect is very important. It shows
the effect of the orientation on the energy and exergy performance of the buildings. Every building in the area
evaluated separately and the sides that faces southwest to southeast and northwest to northeast is calculated and
taken in to consideration correspondingly. The data for these values are also shown in the Table 2.

677
Table 2 Data for SEF and the lateral areas with respect to the orientation
2 2 2
Building South area [m ] North area [m ] Other area [m ] SEF [-]
A1 285.8 358.3 178.6 0
A2 285.8 358.3 821.7 5
A3 408.1 250.1 57.4 8
A4 321.6 214.4 286.9 30
A5 357.3 465.5 0.0 30
A6 285.8 429.8 107.2 8
A7 285.8 358.3 821.7 5
A8 408.1 250.1 57.4 35
A9 321.6 214.4 286.9 5
A10 357.3 465.5 0.0 3
F1 170.0 107.3 277.2 8
F2 154.5 97.5 252.0 8
F3 170.0 107.3 277.2 10
F4 170.0 107.3 277.2 10
F5 170.0 107.3 277.2 5
F6 192.0 78.0 234.0 20
F7 211.2 85.8 257.4 8
F8 170.0 107.3 277.2 10
F9 170.0 107.3 277.2 8
F10 170.0 107.3 277.2 5
F11 154.5 97.5 252.0 5
F12 170.0 107.3 277.2 3
Kırlangıç 1062.3 548.0 768.6 3
Turna 938.0 117.4 577.9 3
Kuğu 407.9 407.9 491.0 3
P1 862.4 862.4 1267.2 3
P2 862.4 862.4 1267.2 3
P3 862.4 862.4 1267.2 5
P4 862.4 862.4 1267.2 5
P5 862.4 862.4 1267.2 5
P6 862.4 862.4 1267.2 5
P7 862.4 862.4 1267.2 0
P8 862.4 862.4 1267.2 3
P9 862.4 862.4 1267.2 3
P10 862.4 862.4 1267.2 3
P11 862.4 862.4 1267.2 3
P12 862.4 862.4 1267.2 3
PV1 61 61 0 50
PV2 61 61 0 50
PV3 61 61 0 70
PV4 61 61 0 80
PV5 61 61 0 80

678
Table 2 Data for SEF and the lateral areas with respect to the orientation (Continued)
2 2 2
Building South area [m ] North area [m ] Other area [m ] SEF [-]
PV6 61 61 0 80
PV7 61 61 0 75
PV8 61 61 0 70
PV9 61 61 0 60
PV10 61 61 0 60
PV11 61 61 0 75
PV12 61 61 0 80
PV13 61 61 0 80
PV14 61 61 0 80
PV15 61 61 0 70
PV16 61.0 61.0 0.0 60
S1 665 665 1292 0
S2 665 665 1292 10
S3 665 665 1292 30
S4 665 665 1292 20
S5 665 665 1292 0
S6 665 665 1292 10
S7 665 665 1292 30
S8 665 665 1292 20
SV1 20 13 33 50
SV2 20 13 33 60

RESULTS AND DISCUSSION

The results of the analysis is tried to be tabulated in the Table 3 with having the values of the exergy load, exergy
by fuel, exergy efficiency and exergy flexibility factor (Figure 4).

The insulation properties, the passive heating with solar energy properties, windowing and orientation properties
and total solids of each building are taken into account in the exergy analysis. An Exergy load of the buildings
defines the amount of exergy needed by the building including the passive effects and the effects of the
surrounding. On the other hand, the exergy by fuel is the definition for the amount of fuel exergy must be supplied
to the building.

The results of the exergy analysis are given in Table 3. It is seen that the efficiency values are around 3-5 % that is
an expected value since a natural gas fired heater is used for a domestic heating system and when the factors of
heat transmission, heat loss and exergy destruction in the combustion process are taken into account. These
values are in accordance with the literature and expected. It is seen that the small efficiency values relies on the
buildings with bad orientation and SEF values. Moreover, the exergy loads of those buildings are considerable
higher than the better oriented and small SEF buildings.

679
Table 3 Results for the Exergy Analysis
Building No: Exergy Load [W] Exergy by Fuel [W] Exergy Efficiency [%] Exergy Flexibility Factor [%]

A1 415586.6 300228.4 4.3499 19.97

A2 415736.1 300343.0 4.3501 19.97
A3 415862.3 300439.6 4.3502 19.98
A4 416143.8 300655.2 4.3505 19.98
A5 416632.4 301029.6 4.3511 19.98
A6 415974.8 300525.8 4.3503 19.98
A7 411699.7 297250.9 4.3453 19.95
A8 417014.9 301322.6 4.3516 19.98
A9 415302.7 300011.0 4.3496 19.97
A10 415623.2 300256.5 4.3499 19.97
F1 280481.6 198710.9 4.1941 19.26
F2 255942.3 181306.4 4.1934 19.26
F3 280162.4 198466.3 4.1934 19.26
F4 280162.4 198466.3 4.1934 19.26
F5 280073.5 198398.2 4.1932 19.25
F6 256012.2 181360.0 4.1935 19.26
F7 279943.9 198298.9 4.1929 19.25
F8 280162.4 198466.3 4.1934 19.26
F9 280126.8 198439.1 4.1933 19.25
F10 280073.5 198398.2 4.1932 19.25
F11 252110.6 178371.2 4.1836 19.21
F12 280037.9 198371.0 4.1931 19.25
Kırlangıç 863827.5 655387.2 4.7552 21.83
Turna 863827.5 655387.2 4.7552 21.83
Kuğu 477371.5 350347.8 4.4783 20.56
P1 330956.5 247729.8 4.6419 21.31
P2 330956.5 247729.8 4.6419 21.31
P3 331136.8 247867.9 4.6420 21.31
P4 331136.8 247867.9 4.6420 21.31
P5 331136.8 247867.9 4.6420 21.31
P6 331136.8 247867.9 4.6420 21.31
P7 330686.0 247522.5 4.6417 21.31
P8 330956.5 247729.8 4.6419 21.31
P9 330956.5 247729.8 4.6419 21.31
P10 330956.5 247729.8 4.6419 21.31
P11 330956.5 247729.8 4.6419 21.31
P12 330956.5 247729.8 4.6419 21.31
PV1 36967.2 24039.9 3.5493 16.30
PV2 36967.2 24039.9 3.5493 16.30
PV3 37094.8 24137.7 3.5537 16.32
PV4 37158.7 24186.6 3.5559 16.33

680
Table 3 Results for the Exergy Analysis (Continued)

Building Exergy Load Exergy by Exergy Efficiency Exergy Flexibility Factor

No: [W] Fuel [W] [%] [%]

PV5 37158.7 24186.6 3.5559 16.33

PV6 37158.7 24186.6 3.5559 16.33
PV8 37094.8 24137.7 3.5537 16.32
PV9 37031.0 24088.8 3.5515 16.31
PV10 37031.0 24088.8 3.5515 16.31
PV11 37126.7 24162.1 3.5548 16.32
PV12 37158.7 24186.6 3.5559 16.33
PV13 37158.7 24186.6 3.5559 16.33
PV14 37158.7 24186.6 3.5559 16.33
PV15 37094.8 24137.7 3.5537 16.32
PV16 37031.0 24088.8 3.5515 16.31
S1 415322.1 327226.0 4.6396 21.30
S2 350184.1 276730.0 4.6404 21.31
S3 351062.6 279107.0 4.6421 21.32
S4 416365.2 330009.0 4.6413 21.31
S5 349744.9 275558.0 4.6396 21.30
S6 415843.7 328617.0 4.6404 21.31
S7 416886.8 331440.0 4.6421 21.32
S8 416365.2 330009.0 4.6413 21.31
SV1 40012.1 26372.0 3.6460 16.74
SV2 39926.3 26306.0 3.6473 16.75

The exergy flexibility factors of the urban area are around %16-20. This is the indication of any replacement in the
system will clearly increase the efficiency and support effective and renewable energy systems.

Table 4 Exergy results for the Mavişehir I, II Area

Exergy Load [kW] 17423.51

Exergy Efficiency [-] 4.20

Exergy Flexibility Factor [-] 19.30

Table 4 shows the total exergy load of the case urban area with the total efficiency and the flexibility factor. It is
seen that this value is very large and proves the estimations about the residential energy consumption. This brings
us to a position where the planners should understand and internalize the energy concern in making designs since
the energy demand and cost will be the prevailing cost of the building and with keeping in mind that the increasing
price and decreasing supply of the energy by fossils.

681
Fig. 4 The Exergy by fuel values of the various buildings in the area

CONCLUSION

With high construction standards and materials the case study area is very effective in fact from energy point of
view. The positive effect of quality insulation elements, the proper orientation and building geometry, and climate –
friendly designs can be easily seen in the exergy analysis, although there is still room for improvement in the
buildings that can be seen from EEF values.

When the Mavişehir 1 and Mavişehir 2 stages are compared it is seen that in general the high-rises have similar
efficiency and EEF values. But on the other hand the villas in the Mavişehir-1 have bas SEF values since they are
located in between the high-rises the passive solar energy input is very limited. That situation decrease the
efficiency of the buildings and the area.. When the bad orientation and solar capacity of the villas taken into
account the energy effective planning and design decision’s importance can easily seen.

So for proper energy efficient planning decisions the parameters taken into account must be varied and the actors
which are in the decision making process must have knowledge and sensitivity about the energy conservation and
sustainability.

On the other hand the low exergy efficiency value of the total area brings us to a position to think about the
importance on the research about better heating/cooling systems, transmission systems, insulation materials,
construction materials and renewable energy.

Furthermore this study will continue by proposing a new exergy efficient plan to the case area and try to make
comparison between the exergy efficient plan and the current situation from efficiency, effectiveness and
sustainability points of views.

682
REFERENCES

Balocco, C., Papeschi, S., Grazzini, G., Basosi, R., Using Exergy to Analyze the Sustainability
of an Urban Area, Ecological Economics 48, 2004, pp.231-244.
Balta, T.M., Dinçer, I., Hepbaşlı, A., Development of Sustainable Energy Options for Buildings in
a Sustainable Society, Sustainable Cities and Society 1, 2011, pp.72-80
Barreiro, E., Belausteguigoitia, J., Perea, E., Rodriguez, R., Romero, A., Turienzo, E., City
Planning and Energy Efficiency: Towards an Integrated Urban Design and Planning
Process , Sustainable Development and Planning IV, Vol.1, 2009, pp.43-51.
Depecker, P., Menezo, C., Virgone, J., Lepers, S., Design of Buildings Shape and Energetic
Consumption, Building and Environment 36, 2001, pp.627-635.
Dinçer İ., Rosen M., Exergy: Energy, Environment, and Sustainable Development, 2007, pp 11-
14
Ercoşkun, Ö.,Y., Sürdürülebilir Kent İçin Ekolojik-Teknolojik (Eko-Tek) Tasarım: Ankara-Güdül
Örneği. PhD Thesis. Gazi University, Ankara, 2007
Hisarlıgil, H., Enerji Etkin Planlamada Konut Adası Tasarımı: Hipotetik Konut Adalarının Ankara
Örneğinde Mikroklimatik Analizi. PhD Thesis. Gazi University, Ankara, 2009
Liu, G., Yang, Z., Chen, B., Extended Exergy-Based Urban Ecosystem Network Analysis: A
Case Study of Beijing, China, Procedia Environmental Sciences 2, 2010, pp.243-251.
Ozturk, H.K., Canyurt, O.E., Hepbasli, A., Utlu, Z., Three Different Genetic Algorithm
Approaches to the Estimation of Residential Exergy Input/ Output Values, Building and
Environment 39, 2004, pp.807-816.
Utlu, Z. and Hepbasli, A., A Review on Analyzing and Evaluating the Energy Utilization
Efficiency of Countries, Renewable and Sustainable Energy Reviews 11, 2007, pp.1-29.
Utlu, Z. and Hepbasli, A., Analysis of Energy and Exergy Use of the Turkish Residential-
Commercial Sector, Building and Environment 40, 2005, pp.641-655

683
 PERFORMANCE ASSESSMENT OF A SMALL SOLAR POND STRATIFIED WITH MAGNESIUM
CHLORIDE WATER
1 2 1 3
Sibel Deniz , Ismail Bozkurt , Mehmet Karakilcik and Ibrahim Dincer
1
Department of Physics, Faculty of Sciences and Letters, University of Cukurova, Adana 01330, Turkey
2
Department of Mechanical Engineering, Faculty of Engineering, University of Adiyaman, Adiyaman, 02040, Turkey
3
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology (UOIT)
2000 Simcoe Street, North Oshawa, Ontario, L1H 7K4, Canada
fzksibel.deniz@hotmail.com, ismail_bozkurt44@yahoo.com, kkilcik@cu.edu.tr, ibrahim.dincer@uoit.ca

ABSTRACT
In this study, we present an experimental investigation of energy and exergy efficiency of a small solar pond
stratified with magnesium chloride water. The solar pond system is filled with varying density magnesium
chloride water in order to form layers of the respective inner zones. A data acquisition device is employed to
measure record the hourly temperatures at various locations in the inner zones. It is found that the gradient
of the different density layers considerably affect the energy storage performance of the solar pond. Thus,
the energy and exergy efficiencies of the solar pond are decreased with the increasing the erosion of layers.
The efficiencies of the inner zones are defined in terms of temperatures as the average representative solar
energy from August to November of the year. As a result, the maximum and the minimum energy efficiencies
of the solar pond are observed in August as 27.41% and 12.64% in November, respectively. Also, the
maximum and the minimum exergy efficiencies of the solar pond are observed in August as 26.04% and
12.62% in November, respectively.

Key words: Solar energy, Solar pond, Energy and Exergy efficiency

INTRODUCTION
Due to a wide range of global problems, a wide range of potential energy solutions are then needed to
reduce environmental impact and increase sustainable development. Many of the renewable energy
resources are of limited applicability due to their reliability, quality, quantity and density (Orhan et al., 2012).
For these reasons, renewable energy sources, solar energy in particular, have gained greater importance.
Solar energy have several uses like heating water, warming greenhouses, drying, water desalination, power
generation and so on. Therefore, research in the development and application of solar energy systems is
essential. One of these systems is solar ponds. In this regard, a lot of studies have been undertaken on solar
ponds by various researchers (Karakilcik et al. 2006, 2008; Leblanc et al., 2011; Nie et al., 2011; El-Sebaii et
al., 2011; Date et al., 2012; Sakhrieh et al., 2013). Bozkurt and Karakilcik (2012) investigated an integrated
system that is a combination of a solar pond and collectors. Heat energy obtained from solar collector is
transferred to the solar pond storage zone through a heat exchanger system that is placed in storage zone.
The integrated solar pond efficiencies were calculated experimentally and theoretically according to the
number of collectors. Karakilcik et al. (2013) investigated a solar pond performance with and without shading
effect and comparison of the energy efficiencies for the upper convective zone, non-convective zone and
lower convective zones, as well as determination of the shading coefficients each zones of the experimental
solar pond. Thus, the shading effect and its ratios, and losses work out to determine energetic performance
of system components.
Sodium chloride salt has been used in most of these studies. In few studies, however, magnesium
chloride salt has been used. Subhakar and Murthy (1991) constructed an experimental saturated solar pond
by using magnesium chloride salt. Subhakar and Murthy (1993) studied magnesium chloride and potassium
nitrate salt ponds located at Madras (India) for the experimental and parametric assessment. A comparison
was also made with an unsaturated solar pond. Subhakar and Murthy (1994) constructed an experimental
saturated solar pond by using magnesium chloride salt. The temperature and concentration gradients were
developed by heating the pond from the bottom and adding finely powdered salt from the top. The
development of a temperature distribution in the pond depends on the direct sunlight and its daily variation.
Many experimental and theoretical works have generally focused on to investigate the thermal
efficiencies of various types of solar ponds, especially, salt gradient solar pond. The present work becomes
the first work in the area dealing with the investigation of energetic and exergetic performance analyses of
the magnesium chloride water solar pond and comparison with the corresponding energy efficiencies during
the months (for summer season) of the year.

684
EXPERIMENTAL APPARATUS AND PROCEDURE
A solar pond is stored heat energy in salty water for a long period of time by preventing heat loss by
convection. They consist of different density zones, Upper Convective Zone (UCZ), Non-Convective Zone
(NCZ) and Heat Storage Zone (HSZ). UCZ is the fresh water layer at the top of the pond. This zone is fed
with fresh water to maintain its density as close as possible to the density of fresh water in the upper part
and to meet the water loss due to evaporation. NCZ is composed of salty water layers whose brine density
gradually increases toward bottom of the pond. The density gradient of NCZ is very important to prevent heat
loss by convection. In order to maintain the density gradient of NCZ an auxiliary operating system on a
feedback principle was developed. The system is first proposed by Akbarzadeh and MacDonalds (1982).
HSZ is composed of salty water with highest density. The high amount of the solar energy is absorbed and
2
stored in this zone. An experimental model solar pond with the area of 0.72 m and a depth of 1.1 m was
built in University of Cukurova, Turkey (i.e., 35°18′ E longitude, 37°05′ N latitude). The pond’s bottom and
side wall was insulated by using 0.10 m thick glass wool. The pond temperature was measured at 7 points,
starting from the bottom, at 0.25, 0.40, 0.55, 0.65, 0.75, 0.85 and 1.05 m heights by using thermocouples
with an accuracy of about ±1 °C. Density distributions was also measured and analyzed by taking samples
from at the same point of the temperature sensors. The pond was filled in August 2012 and worked. The
thicknesses of the UCZ, NCZ and HSZ are 0.1, 0.5, 0.40 m, respectively. The range of salt gradient in the
3 3
inner zones is such that the density is 1000-1020 kg/m in UCZ, 1030-1200 kg/m in NCZ and 1170-1200
3
kg/m in HSZ. In the present study, we report the experimental data that were measured from August to
December. Figure 1 shows a schematic representation of the experimental solar pond system by using
magnesium chloride salt.

Figure 1. A schematic representation of the magnesium chloride layers of the solar pond.

ENERGY AND EXERGY ANALYSES

The energy and exergy values of the solar pond are calculated using conservation of mass and energy
principles as well as second law analysis. Therefore, in this section, energy and exergy analyses of the solar
pond are presented. The energy efficiency of the magnesium chloride solar pond (MCSP) can be defined as

ηMCSP =
Q stored
= 1−
{Q down + Qup + Q side } (1)
Q solar Q solar

where Qstored is stored heat energy in the HSZ of the solar pond, Qsolar is the amount of net solar energy
which is coming from the sun, Qdown is the total heat loss to the down wall from HSZ, Qup is the heat loss from
HSZ to the above zone. Qside is the total heat loss to the side walls of the solar pond. Substituting equations
for each parameter in Eq. 1 provides us with the following energy efficiency of MCSP:

685
  k sw A 
 (Tdown − Ta ) + k s A (THSZ − TNCZ ) + k sw 2 π rLHSZ (THSZ − Ta )
ηMCSP = 1−
 Δx down Δx HSZ-NCZ Δx side  (2)
[
βEA HSZ (1 − F ) h(x − δ ) ]
where E is the total solar energy reaching the solar pond surface, AHSZ is the area of the HSZ which is
subjected to solar insolation, F is the fraction of energy absorbed at a region of δ-thickness, h is solar
radiation ratio; A is the surface area of the solar pond, Ta is the ambient air temperature, ksw is the thermal
conductivity of the side and bottom walls, ks is the thermal conductivity of the salty water, LHSZ is the
thickness of the HSZ (m), r is inner radius of the cylindirical solar pond, Δxdown is the thickness of the down
wall, Δxside is the thickness of the side wall, ΔxHSZ-NCZ is the thickness of the HSZ’s middle point and the
NCZ’s middle point, and β is the fraction of the incident solar radiation that actually enters the pond and is
given by Hawlader (1980) as follows:

2 2
 sinθi − sinθr   tanθi − tanθr 
β = 1 − 0.6   − 0.4   (3)
 sinθi + sinθr   tanθi + tanθr 

Here, θi and θr are the incidence and refraction angles. h represents the ratio of the solar energy reaching

[ ]
the depth in the layer I to the total solar incident falling on to the surface of the pond and is given by Bryant
(x −δ)
and Colbeck (1977) as hI =0.727 −0.056ln I (here xI is the thickness of the layer, δ; thickness of the layer in the
cosθr

UCZ where long-wave solar energy is absorbed).

Figure 2. Exergy flows in the inner zones of MCSP

As shown in Fig. 2, the exergy flows in UCZ can be illustrated. We first write the exergy balance
equation for HSZ as

[
Ex r,NCZ − Ex d,HSZ − Ex l,HSZ + Ex sw,HSZ + Ex b,HSZ = ΔE st ] (4)

686
Here, Exr,NCZ is the recovered exergy from NCZ for HSZ, Exd,HSZ is the exergy destruction in HSZ, Exl,HSZ is
the exergy loss from HSZ to NCZ, Exsw,HSZ is the exergy loss through side walls. Exb,HSZ is the exergy loss
through bottom wall and also, Exst is the exergy stored in HSZ. Here Exd,HSZ is the exergy destruction in HSZ
which can be written as:

(
Ex d,HSZ = T0 ΔS net,HSZ ) (5)

where ΔSnet is the net entropy change of NCZ which is defined ΔSnet=ΔSsys+ΔSsurr. Then, the exergy losses,
including exergy destructions, within NCZ can be derived as follows:

 THSZ  Q g,HSZ Q sw,HSZ   Qb 

Ex d,HSZ = T0 mHSZ Cp,HSZln −  +  +   (6)
 T0  THSZ T0   T0 

  THSZ 
(
Ex l,HSZ = mHSZ Cp,HSZ  THSZ − Tm,NCZ − T0  ln )  T

 (7)
  m,NCZ 

Here, mHSZ=ρHSZVHSZ is the mass of salty water in HSZ; ρHSZ is the averaged density as given earlier []; and
VHSZ is the volume of the salty water in HSZ. Cp,HSZ is the specific heat of HSZ; T0 is the reference
environment temperature; THSZ is temperature of HSZ and Tm,NCZ denote the average temperature of NCZ.
We can now define the exergy efficiency for HSZ as the ratio of the exergy recovered from UCZ to the total
exergy input to HSZ:

  THSZ 
(
Ex sw,HSZ = mHSZ Cp,HSZ  THSZ − Tsw,HSZ − T0  ln )  T

 (8)
  sw,HSZ 

where, Tsw,HSZ denote side wall of the average temperature for HSZ.
Furthermore, Exb,HSZ=Exsw,HSZ, due to the fact that both side wall and bottom layer have the same
insulating materials and are surrounded by the ambient air. We can now define the exergy efficiency for HSZ
as the ratio of the exergy stored in HSZ to the total exergy input to HSZ which is essentially the exergy
recovered from NCZ:

ΔEx st Ex d,HSZ + Ex l,HSZ + Ex sw,HSZ + Ex b,HSZ

ψHSZ = = 1− (9)
Ex r,NCZ Ex r,NCZ

RESULTS AND DISCUSSION

The density gradient is an important part of the maintenance of the solar pond as it affects the heat storage
performance. Fig. 3 shows the alteration of the experimental salty solution densities with height from bottom
to surface of the pond, throughout the four months. Here, density differences are observed between the
density variation measured for these four different months, due essentially to increase of the inner zone
temperature and to diffusion of the salt molecules. As seen in Fig. 3, the density gradient in the solar pond is
kept approximately stable. Some erosion is observed at the top of HSZ. The density erosion is prevented by
using salt gradient protection system.
Regarding the experimental temperature distributions in Fig. 4, the average temperature of HSZ is
observed to be a maximum of 52.42 ºC in August, a minimum of 24.46 ºC in November. Similarly, the
average temperature of NCZ is observed to be a maximum of 44.29 ºC in August, a minimum of 23.17 ºC in
November. The average temperature of UCZ is observed to be a maximum of 30.31 ºC in August, a
minimum of 15.16 ºC in November.
Figure 5 shows both averaged energy and exergy distribution from August to November. As seen in
the figure, the energy and exergy contents are determined to be a maximum of 114.77 MJ and 106.88 MJ in
August, a minimum of 54.90 MJ and 51.33 MJ in November, respectively. Difference of the energy and
exergy is due to the fact that energy is conserved, but not exergy.

687
 1200
Aug.
Sep.
1150
Oct.
Nov.

Density (kg/m3 )
1100

1050

1000

950

900
0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 1.05
Height from the bottom (m)
Figure 3. Density ditribution of MCSP.

60
Aug.
Sep.
50 Oct.
Nov.
Temperature (°C)

40

30

20

10

0
0.25 0.4 0.55 0.65 0.75 0.85 1.05
Height from the bottom (m)
Figure 4. Temperature distribution of MCSP.

140
Energy
120 Exergy
Energy and Exergy (MJ)

100

80

60

40

20

0
Aug. Sep. Oct. Nov.
Month
Figure 5. The amount of energy and exergy input to HSZ

Fig. 6 shows the variation of the exergy input, exergy stored and destruction and losses from August
to November. The exergy content distributions were calculated by using temperature distribution of the solar
pond and the reference air temperature. As seen in Fig. 6, the exergy inputs are equivalent to the summation
of exergy stored and exergy destruction and losses. The exergy stored appear to be maximum 27.84 MJ in
August and minimum 6.48 MJ in November, respectively. The exergy destruction and losses appear to be
maximum 79.05 MJ in August and minimum 44.84 MJ in November, respectively.

688
 120
Exergy input
Exergy stored
100
Exergy destruction and losses

80

Exergy (MJ)
60

40

20

0
Aug. Sep. Oct. Nov.
Month
Figure 6. Variations of the exergy input, stored, destruction and losses of HSZ

30
Energy
Exergy
25

20
Efficiency (%)

15

10

0
Aug. Sep. Oct. Nov.
Month
Figure 7. Energy and exergy efficiency of MCSP

The energy and exergy efficiency variations are given in Fig. 7 for comparison purposes. As seen in
the figure, the maximum and the minimum energy efficiencies of the solar pond are observed in August as
27.41 % and 12.64 % in November, respectively. Also, the maximum and the minimum exergy efficiencies of
the solar pond are observed in August as 26.04 % and 12.62 % in November, respectively.

CONCLUSION
In this paper, we have studied both energetic and exergetic performances of a magnesium chloride solar
pond through efficiency analysis. Exergy efficiencies of a magnesium choloride water solar pond were found.
The exergy efficiencies determined for storage zone using the experimental data are compared with the
corresponding energy efficiencies. As expected, the exergy efficiencies appear to be little less than the
energy efficiencies for sorage zone of the solar pond due to the small magnitudes of exergy destructions in
the storage zone and losses to surroundings. These results openly demonstrate the effect of the insulation
materials and ambient temperature on the performance of the solar pond. It is important that exergy is a
potential to help achieve better efficiency of the solar pond.

Acknowledgements
The authors are thankful to University of Cukurova for financial support of the present work (Grant No.
FEF2010BAP5 and FEF2012YL13).

NOMENCLATURE
2
A surface area, m
C specific heat, J/kg K
2
E total solar energy reaching to the pond, MJ/m
Ex exergy, J
F absorbed energy fraction at a region of δ-thickness
h solar radiation ratio

689
HSZ heat storage zone
k thermal conductivity, J/m°C
L thickness of the inner zones, m
m mass, kg
MCSP magnesium chloride solar pond
NCZ non-convective zone
Q heat, J
r inner radius, m
S entropy, J/Kmol
T temperature, °C
UCZ upper convective zone
3
V volume, m

Greek Letters
η energy efficiency
δ thickness where long wave solar energy is absorbed, m
β incident beam entering rate into water
θ angle, rad
ρ
3
density, kg/m
ψ exergy efficiency

Subscripts
a ambient
b bottom
d destruction
dw down wall
g gained
HSZ heat storage zone
i incident
in energy input
m mean
NCZ non-convective zone
net net irradiation
out energy output
r refraction
rec recovered
surr surrounding
sw side wall
sys system

References
Akbarzadeh, A. and Macdonald, R.W.G. 1982. Introduction of a passive method for salt replenishment in the
operation of solar ponds. Solar Energy 29:71-76.

Bozkurt, I. and Karakilcik, M. 2012.The daily performance of a solar pond integrated with solar collectors.
Solar Energy 86:1611-1620.

Bryant, H.C. and Colbeck, I. 1977. A solar pond for London. Solar Energy 19:321.

Date, A., Alam, F., Khaghani A., Akbarzadeh, A. 2012. Investigate the potential of using trilateral flash cycle
for combined desalination and power generation integrated with salinity gradient solar ponds. Procedia
Engineering 49:42-49.

El-Sebaii, A.A., Aboul-Enein, S., Ramadan, M.R.I., Khallaf, A.M. 2011. Thermal performance of an active
single basin solar still (ASBS) coupled to shallow solar pond (SSP). Desalination 280:183-190.

Hawlader, M.N.A. 1980. The influence of the extinction coefficient on the effectiveness of solar ponds. Solar
Energy 25:461-464.

690
Karakilcik, M., Dincer, I., Rosen, M. A. 2006. Performance investigation of a solar pond. Applied Thermal
Engineering 26:727–735.

Karakilcik, M. and Dincer, I. 2008. Exergetic performance analysis of a solar pond. International Journal of
Thermal Sciences 47:93–102.

Karakilcik, M., Dincer, I., Bozkurt, I. and Atiz, A. 2013. Performance assessment of a solar pond with and
without shading effect. Energy Conversion and Management 65:98-107.

Leblanc, J., Akbarzadeh, A., Andrews, J., Lu, H. Golding, P. 2011. Heat extraction methods from salinity-
gradient solar ponds and introduction of a novel system of heat extraction for improved efficiency. Solar
Energy 85:3103–3142.

Nie, Z., Bu, L., Zheng, M., Huang, W. 2011. Experimental study of natural brine solar ponds in Tibet. Solar
Energy 85:1537–1542.

Orhan, M.F., Dincer, I., Rosen, M.A., Kanoglu, M. 2012. Integrated hydrogen production options based on
renewable and nuclear energy sources. Renewable and Sustainable Energy Reviews16:6059–6082.

Sakhrieh, A., Al-Salaymeh, A. 2013. Experimental and numerical investigations of salt gradient solar pond
under Jordanian climate conditions, Energy Conversion and Management 65:725–728.

Subhakar, D. and Murthy S.S. 1991. Experiments on a magnesium chloride saturated solar pond.
Renewable Energy 5-6:655–660.

Subhakar, D. and Murthy, S.S. 1993. Saturated solar ponds: 2. Parametric studies, Solar Energy 50:307–
319.

Subhakar, D. and Murthy, S.S. 1994. Saturated solar ponds: 3. Experimental verification, Solar Energy 53:
469–472.

691
PERFORMANCE ASSESSMENT OF A RECUPERATIVE HELIUM GAS TURBINE SYSTEM
a,b,* a
R. S. El-Emam , I. Dincer
a
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology,
2000 Simcoe Street North, Oshawa, Ontario, L1H 7K4, Canada.
b
Faculty of Engineering, Mansoura University, Mansoura, Egypt.
rami.elemam@uoit.ca

ABSTRACT
Helium is considered an ideal working fluid for closed cycle gas turbines powered by the heat of nuclear reactors or sun
concentrators. Energetic and exergetic based thermodynamic analysis is applied to an actual 120 MW recuperative
closed gas turbine operating with helium with two compression stages. The maximum pressure and temperature of the
plant are 70 bar and 850oC, at the turbine inlet. Parametric studies are performed to investigate the effects of different
parameters on the plant performance and the irreversibilities associated with the system operation. In the exergy
analysis, internal and external irreversibilities are considered. At the optimal pressure ratio, the energy and exergy
efficiencies obtained here are found as 45.5% and 60.5% at the base case. Also, various exergetic parameters are
calculated for the exergy assessment of the gas turbine plant. Furthermore, gas turbine exergetic performance map is
introduced.

Keywords: Energy, exergy, efficiency, pressure ratio, closed cycle, gas turbine

INTRODUCTION
Several gas turbine plants, based on the well-established closed cycle gas turbine technologies, are in operation for
over hundreds of thousand hours. The main advantages of this technology is its adaptability to a wide range of fuels,
e.g., gas, oil, coal, solar and nuclear fuels and the possibility to use different working gases, e.g., air, helium, nitrogen,
carbon dioxide, argon, neon or their mixtures. Helium is the one of the most suitable working fluids for large closed
cycle gas turbine plants. Helium turbines, based on a closed Brayton gas cycle, are also considered proper alternative
for efficient power generation for nuclear reactors. Helium is radioactive stable, inert and non-corrosive gas (Conn and
Kuo 1976; Frutschi 2005; Horlock 2002). Helium gas offers a high thermal capacity as well.
In addition, helium gas turbine is a candidate as a multi-generation energy system for different purposes. The
heat rejected from the plant is available at high temperature. It can be utilized for combined cooling and heat
production, water desalination, process steam and industrial process heat in integrated chemical plants. In early 60’s,
Two helium closed cycle gas turbines were built for power generation and air liquefaction, heat was provided to the
gas at 650 - 660oC and 18 - 23 bar through a natural gas furnace. In 1974, a helium gas turbine achieved a 30 MW of
power production when turbine was fed with helium at 750oC and 27 bar (Frutschi 2005). Helium turbines are also
used integrated with nuclear reactors where heat is to be provided through molten salt blanket or heat exchanger that
transmits the heat from the reactor cooling system to the helium turbine plants. Conn et al. (Conn and Kuo 1976)
studied a conceptual design of 500 MW helium gas turbine units to be utilized with a nuclear reactor. The operating
turbine inlet temperature was set to 870oC at around 68 bar. Solar energy is a good source of heat to helium turbines
as well. The low density and high thermal capacity of helium facilitate a more efficient operation of the gas turbine at
more compact design of the heat exchangers of the system. Heat can be provided using heliostat solar towers or
parabolic dish technology and it can be concentrated to a high pressure helium heater or through a molten salt heat
exchanger with storage tank.
Various studies in the literature have been undertaken to investigate the performance of helium gas turbine
power plants. Some researchers (No, et al. 2007) performed a review study on helium turbines that works with the
high temperature gas cooler reactors. A model for calculating the design point performance of the gas turbine plant
was developed in this study. A transient analysis software for high temperature cooled reactor helium gas turbine was
developed (Wenlong, et al. 2012). A review on the helium gas turbine history and the operation experience from
different test and power facilities was performed (McDonald 2012), focusing on the helium turbines powered by the
nuclear heat as well. Multi-reheat helium gas turbine was investigated (Zhao and Peterson 2008) at different turbine
inlet temperature values and achieved a thermal efficiency of 39% to 47%, respectively.
The thermodynamic assessment of closed cycle gas turbine plant is performed at different operating
conditions at steady and transit operation. A recuperative closed cycle gas turbine of a scramjet was
thermodynamically investigated (Qin, et al. 2010). They presented a thermal management system for reducing the
hydrogen cooling flow and also studied the effect of different operating parameters on the system performance. A work
(Gandhidasan 1993) was performed to study the effects of different pressure ratio on a solar driven closed cycle gas
turbine.

The present work aims to investigate the system performance, through energetic and exergetic efficiencies, of
an actual helium closed cycle recuperated gas turbine with intercooling, through energy and exergy analyses. The

692
performance parameters are studied, and the effects of changing the operating parameters on the performance are
also considered. The pressure losses on the cycle are taking into consideration in the energy and irreversibility
analyses. Including the internal irreversibility calculations give a more realistic analysis of the gas turbine. The
performance of the gas turbine is defined in this work in a novel representation with respect to the exergy efficiency.
The exergetic performance map is introduced as a more descriptive approach with respect to the actual potential of
gas turbine performance improvement at certain operating condition. Exergetic loss map is also introduced as a new
exergetic assessment map. It relates the exergy destruction ratio, with respect to the total exergy of fuel streams, to
the specific work or the dimensionless work parameter.

SYSTEM DESCRIPTION
The enthalpy-entropy diagram for simple closed cycle gas turbine is shown in Fig. 1. The diagram shows a realistic
cycle with considering the pressure losses that causes internal irreversibilities in the cycle heat exchangers. Turbine
and compressor irreversibilities in form of the device adiabatic efficiency are considered in this diagram as well. PH and
PL on this diagram represent the highest and lowest actual operating pressure in the cycle, and the pressure ratio of
the gas turbine is defined in this study based on these two values.

PH + ΔPH
PH
3* 3
h & T = constant

PL + ΔPL
Specific Enthalpy

2
2is 4is
4*
2*

1*
h & T = constant
1
PL + ΔPL PL
Specific Entropy
Fig.1 Enthalpy-Entropy diagram of a realistic simple closed cycle gas turbine.

The recuperative closed cycle helium gas turbine considered in this study is shown in Fig. 2. A two-stage
compression with intercooling is considered for this system, while one turbine is used for power production and to drive
the compressors. After being heated, helium expands through the turbine. The low pressure gas from the turbine
enters a heat transfer recuperator to heat the cooler gas at the high pressure side of the cycle. The cooled helium
leaving the recuperator is directed to another heat exchanger that works as a precooler before the compressor. The
working fluid is further cooled in the precooler. Heat rejected from the cooler can be transferred to generator of an
absorption cooling system for cogeneration. Two stage compressions with intercooler are used to raise the helium
pressure to its highest value in the cycle. Then helium is heated by the turbine low pressure outlet gas through the
recuperator before it goes to the main heat source.
er
ng
ha

Precooler
xc
tE
ea
H

recuperator

C C T G

Intercooler

Fig. 2 Schematic of recuperative closed cycle gas turbine with intercooler.

693
THERMODYNAMIC MODELING
In a thermodynamic study, it is essential to make some assumptions for analysis and assessment. Here, we make the
following assumptions for energy and exergy analyses of the system:
− Turbine and compressors are adiabatic.
− The plant operates at steady state condition.
− Helium properties are taken for ideal gas assumption.
− Heat source is considered as molten salt at constant temperature in all the analyses.
− The pressure ratio is defined in terms of the higher pressure, PH, and lower pressure, PL, of the cycle after
considering the pressure losses.
− Ambient temperature and pressure are constant.
− No cogeneration and no use of the rejected heat of the cycle are considered in the exergy assessment
calculations.
− Changes in kinetic and potential energies and exergies are negligible.

Energy Analysis
The prescribed helium gas turbine energy performance is investigated and modeled in the following subsections. The
energy analysis of the components is performed and the components performance is described in form of the following
parameters:
− Turbine and compressor polytropic efficiency.
− Plant pressure ratio.
− Plant temperature ratio.
− Heat exchangers effectiveness.
− Heat exchangers pressure loss coefficients.
The first law of thermodynamics is applied to each of system components. The following equation represents the
concept of conservation of energy for an open system:
dE⁄dt = ∑ Q̇ − Ẇ + ∑in ṁ h − ∑out ṁ h (1)
where the changes in kinetic and potential energies across the system are considered negligible.

Exergy Analysis
The practical operation of any system is an irreversible operation, where work is destroyed based on the entropy
generation rate that is associated with the operating process. The amount of lost available work can be defined as the
difference between the reversible work when no irreversibilities are considered, i.e., entropy generation is zero, and
the system actual irreversible work.
𝑊̇𝑙𝑜𝑠𝑠 = 𝑊̇𝑟𝑒𝑣 − 𝑊̇ (2)
The lost available work is not an actual energy loss; it represents a lost opportunity to convert part of the energy
utilized into a useful work. The lost available work was defined (Kestin 1980) in a direct proportional relation with the
entropy generation rate. This is better described by the exergy destruction term which expresses the system
irreversibility as follows:
𝐸𝑥̇𝑑 = 𝑇𝑜 . 𝑆𝑔𝑒𝑛̇ (3)
The temperature To is the defined as the dead state temperature. It is the temperature at which the system cannot
undergo any state change through any interactions with the environment (Moran, et al. 2011). The entropy generation
for a general open system can be determined based on the second law of thermodynamics.
𝑑𝑆 𝑄 ̇
̇
𝑆𝑔𝑒𝑛 = 𝑑𝑡 − ∑ 𝑇 + ∑𝑜𝑢𝑡 𝑚̇ 𝑠 − ∑𝑖𝑛 𝑚̇ 𝑠 (4)
where T is the temperature of the boundaries at which the heat transfer occurs:
𝑇 𝑑(𝐸−𝑇𝑜 𝑆)
𝑊̇ = ∑ �1 − 𝑇𝑜 � 𝑄̇ + ∑𝑖𝑛 𝐸𝑥̇ + ∑𝑜𝑢𝑡 𝐸𝑥̇ − − 𝐸𝑥̇𝑑 (5)
𝑑𝑡
The term 𝐸𝑥 ̇ expresses the exergy flow rate or the exergy associated with the mass flow rate of the streams that enter
or exit from the system. It is defined as:
𝐸𝑥̇ = 𝑚̇ 𝑒𝑥 = 𝑚̇(ℎ − 𝑇𝑜 𝑠) (6)
The exergy destruction rate is applied on the prescribed gas turbine plant components as illustrated in the following
section, considering the ideal gas behaviour of the working fluid.
The exergy destruction is expressed taking external and internal irreversibilities into consideration. External
irreversibilities are associated with the heat transfer between the working fluid and the hot and cold reservoirs at the
source and dead state temperature values, respectively. Internal irreversibilities are the mainly accompanied with the
finite pressure drop across the system devices. In the following sections, energy and exergy analyses of each for
system components are presented.

694
Table 1: The helium turbine plant operating parameters.
Operating Parameters for base case
Net power output 120 MW
Turbine polytropic efficiency 92 %
Compressor polytropic efficiency 90 %
Mechanical efficiency 99 %
Recuperator effectiveness 88 %
Precooler effectiveness 90 %
Intercooler effectiveness 87 %
Pressure loss in recuperator, hot side 2.1 %
Pressure loss in recuperator, cold side 1.8 %
Pressure loss in Precooler, hot side 1.2 %
Pressure loss in Intercooler, hot side 1.3 %
o
Turbine inlet temperature 850 C
o
Compressor inlet temperature 40 C
Turbine inlet pressure 70 bar
Plant temperature ratio 313K/1123K
o
Ambient temperature 25 C
o
Hot source temperature 930 C
o
Cooling water inlet temperature 25 C

Compressor Modeling
The system in this study has two compression stages. The total pressure ratio is a design parameter as in Table 1. The
compressor is considered adiabatic and its polytropic efficiency is also provided as in Table 1. Practically, helium
compressor blades design is different than the air compressors. It has to account for the flow separation and
secondary flow occurrence to achieve a better performance (McDonald 2012; No, et al. 2007). Also, due to helium low
molecular weight, a more reliable sealing mechanism needs to be integrated with the compressor and the other
system components to avoid leakage (McDonald, et al. 1994). Based on the defined pressure ratio and polytropic
efficiency of the compressor, and the plant temperature ratio; the power required to drive the compressor can be
formulated as follows:
𝑟𝑐𝑘 −1
𝑊̇𝐶 = 𝑚̇ 𝑐𝑝 𝑇𝐻 𝛽 �𝜂 � (7)
𝐶,𝑎 𝜂𝑚𝑒𝑐ℎ
where 𝑇𝐻 is the highest temperature at the gas turbine plant, i.e., turbine inlet condition, 𝛽 is the operational
temperature ratio, and is defined as follows:
𝑇
𝛽 = 𝑇𝐿 (8)
𝐻
and k is function of the heat capacity ratio:
𝛾−1
𝑘= 𝛾
(9)
The pressure ratio is defined as the ratio between the higher and lower pressure across the compressor, which is
defined as follows, taking into consideration the pressure losses at the high and low pressure streams of the working
fluid.
𝑃𝐻
𝑟𝑐 = (10)
𝑃𝐿
The compressor adiabatic efficiency can be defined as a function of the polytropic efficiency as follows:
𝑟𝑐𝑘 −1
𝜂𝐶,𝑎 = 𝑘/𝜂𝐶,𝑝 (11)
𝑟𝑐 −1
The temperature of the stream exits the compressor is to be estimated based on the inlet temperature of the working
fluid:
𝑟𝑐𝑘 −1
𝑇𝐶,𝑒𝑥 = 𝑇𝐶,𝑖𝑛 �1 + 𝜂𝐶,𝑎
� (12)

Considering an adiabatic compression process, the exergy destruction rate associated with this process can be
formulated as a function of the total pressure ratio of the gas turbine plant and the compressor adiabatic efficiency as
follows:
1 1
𝐸𝑥̇𝑑,𝐶 = 𝑚̇ 𝑇𝑜 𝑐𝑝 ln ��𝑟 𝑘 − 1� �1 − 𝜂 � + 1� (13)
𝑐 𝐶,𝑎

695
For the calculations, identical adiabatic efficiency values are then chosen for the two operating compressors.

Intercooler Modeling
The high thermal conductivity and heat transfer coefficient of helium, compared with air, results in more compact
designs of the utilized heat exchangers. The use of intercooler in between compression stages is a standard feature of
efficient gas compressors. The total compression required power is formulated as a function of the total pressure ratio
and the compressors stage adiabatic efficiency as follows:
𝑚̇ 𝑐𝑝 𝑇𝐻 𝛽 𝑘/𝑁𝐶 1
𝑊̇𝐶 = �𝑟𝑐 − 1� ∑𝑁𝐶 �𝜂 � (14)
𝜂𝑚𝑒𝑐ℎ 𝐶,𝑎
For two compression stages with intercooling and the same adiabatic efficiency of the two stages, the total
compression work becomes
𝑘/2
𝑟 −1
𝑊̇𝐶 = 2. 𝑚̇ 𝑐𝑝 𝑇𝐻 𝛽 �𝜂 𝑐 𝜂 � (15)
𝐶,𝑎 𝑚𝑒𝑐ℎ
which does not count for the pressure loss in the intercooler. For the definition of the total pressure ratio in this study
as the final pressure after second compression stage over the lowest pressure at the inlet of the first compression
stage, after considering the total pressure loss in the cycle, the second stage of compression would count for the drop
in the pressure at the intercooler, which means a slightly higher pressure ratio would be required for the second
compression stage to achieve the design total pressure ratio of the plant. This results in more required work to drive
the high pressure compressor, even if the two compressors operate at the same adiabatic efficiency.
The heat transfer rate removed from the working fluid at the intercooling stage is formulated as follows with
the assumption that the flow exits the intercooler at the inlet temperature of the preceding compressor:
𝑄̇𝐼𝐶 = 𝑚̇ 𝑐𝑝 �𝑇𝐶,𝑒𝑥 − 𝑇𝐻 𝛽� (16)
where 𝑇𝐶,𝑒𝑥 is the temperature at the exit of the preceding compression stage, at the intercooler inlet.
In the intercooler, there are irreversibilities associated with the heat transfer within the temperature difference between
the stream flow and the ambient, and irreversibility associated with the flow due to the pressure drop across the
device. The exergy destruction in the intercooler can be described as follows, based on the output of the energy
analysis:
𝑄 ̇ ∆𝑃
𝐸𝑥̇𝑑,𝐼𝐶 = � 𝑇 𝐼𝐶
⁄𝑇
� − 𝑚̇ 𝑇𝑜 ��𝑠𝐼𝐶,𝑖𝑛 − 𝑠𝐼𝐶,𝑒𝑥 ��𝑃 − 𝑅 � 𝑃 � � (17)
𝐼𝐶 𝑜 𝐼𝐶 𝐼𝐶
Here, the destruction of exergy is expressed for an equivalent intercooler that works with no pressure loss, hence,
slightly warmer, and then the entropy increase because of the pressure drop is added as it is expressed in the
equation for an ideal gas assumption.

Gas Turbine Modeling

The efficiency of the helium turbine is highly affected by its design, which is a bit different from the conventional gas
turbine design. It has shorter blade height and this result in an increase of the blade tip clearance leakage (McDonald
2012; McDonald, et al. 1994; No, et al. 2007). On the other hand, compared with steam turbine, helium gas turbine is
more economic due to the less material requirements as the blades of the former are ten times larger in length (Boyce
2006; Frutschi 2005). The power output of the gas turbine is formulated in the following form, as function of the plant
pressure ratio:
∆𝑃
1+𝑘∙∑� �
𝑊̇𝑇 = 𝑚̇ 𝜂𝑇,𝑎 𝑐𝑝 𝑇𝐻 �1 − 𝑃
� (18)
𝑟𝑐𝑘
where the turbine adiabatic efficiency is calculated as a function of the polytropic efficiency of the turbine as follows:
𝑘
𝜂
1−(1⁄𝑟𝑇 ) 𝑇,𝑝
𝜂𝑇,𝑎 = (19)
1−(1⁄𝑟𝑇 )𝑘
The turbine pressure ratio is to be determined based on the defined pressure ratio of the compression stages,
considering the pressure losses in the working fluid streams across the different system component, as follows:
∆𝑃 ∆𝑃
𝑃𝑇,𝑖𝑛 1−�� � +� � �
𝑃 𝐻𝐸 𝑃 𝑅,𝑐
𝑟𝑇 = 𝑃 = 𝑟𝑐 ⋅ � ∆𝑃 ∆𝑃
� (20)
𝑇,𝑒𝑥 1+�� � +� � �
𝑃 𝑃𝐶 𝑃 𝑅,ℎ
The total pressure loss effect in the turbine work equation is the summation of the pressure loss coefficients in the hot
and cold streams of the recuperator, the hot streams of the precooler and intercooler, and the cold stream of the main
heat exchanger.
∆𝑃 ∆𝑃 ∆𝑃 ∆𝑃 ∆𝑃
∑� � = � � +� � +�𝑃� +�𝑃� (21)
𝑃 𝑃 𝑃 𝑅 𝑃𝐶 𝐼𝐶 𝐻𝐸

696
where the recuperator pressure loss is described as the summation of its hot and cold streams pressure losses.
∆𝑃 ∆𝑃 ∆𝑃
�𝑃� = �𝑃� +�𝑃� (22)
𝑅 𝑅,ℎ 𝑅,𝑐
The temperature value at the exit of the turbine is to be determined as follows:
1
𝑇𝑇,𝑒𝑥 = 𝑇𝐻 �1 − 𝜂𝑇,𝑎 �1 − 𝑟 𝑘 �� (23)
𝑇

as a function of the operating compressor pressure ratio.

The irreversibility associated with the turbine, amounts to the following form, as a function of the gas turbine
operating parameters:
𝐸𝑥̇𝑑,𝑇 = 𝑚̇ 𝑇𝑜 𝑐𝑝 ln�1 − (𝑟𝑇𝑘 − 1)(𝜂𝑇,𝑎 − 1)� (24)
Here, the working fluid is treated as an ideal gas with constant specific heat.

Recuperator Modeling
From the defined recuperator effectiveness, the temperature at the exit of the cold stream, at the high pressure side, is
determined:
𝑘 ∆𝑃
��𝑟𝑐 � −1 1+𝑘∙∑� �
𝑇𝑅,𝑐,𝑒𝑥 = 𝑇𝐻 �𝛽(1 − 𝜀𝑅 ) �1 + � + 𝜀𝑅 �1 − 𝑃
�� (25)
𝜂𝐻𝐶,𝑎 𝑟𝑐𝑘 / 𝜂𝑇,𝑎

This formula is derived based on the following recuperator effectiveness definition:

𝑇 −𝑇
𝜀𝑅 = 𝑇𝑅,𝑐,𝑒𝑥 −𝑇𝑅,𝑐,𝑖𝑛 (26)
𝑅,ℎ,𝑖𝑛 𝑅,𝑐,𝑖𝑛
The hot stream exit temperature, on the low pressure side, is to be determined by applying the energy balance on the
hot stream, at low pressure side, as follows:
𝑇𝑅,𝐻,𝑒𝑥 = 𝑇𝑅,𝐻,𝑖𝑛 − 𝑄̇𝑅 ��𝑚̇ 𝑐𝑝 � (27)
where 𝑄̇𝑅 is the heat transfer through the recuperator from the hot stream to cold stream. It is simply calculated from
the energy balance of the cold stream:
𝑄̇𝑅 = 𝑚̇ 𝑐𝑝 (𝑇𝑅,𝑐,𝑒𝑥 − 𝑇𝑅,𝑐,𝑖𝑛 ) (28)
By defining the exist streams of the recuperator, the analysis of the main heat exchanger and the precooler can be
performed. Based on the two working fluid streams passing through the recuperator, the exergy associated with the
recuperator can be formed as follows:
∆𝑃
𝐸𝑥̇𝑑,𝑅 = 𝑚̇ 𝑇𝑜 ��𝑠𝑅,𝑐,𝑒𝑥 − 𝑠𝑅,𝑐,𝑖𝑛 ��𝑃 + �𝑠𝑅,𝐻,𝑒𝑥 − 𝑠𝑅,𝐻,𝑖𝑛 ��𝑃 + 𝑅 � 𝑃 � � (29)
𝐻 𝐿 𝑅
Same as for the other heat exchangers in this analysis, the entropy generation analysis is performed and is formed in
a no pressure loss entropy differences, and then the internal irreversibility caused by the pressure losses in hot and
cold streams is added. The total pressure coefficient in the recuperator is equal to both sides’ pressure losses
coefficients. In the recuperator case, pressure loss at both sides need to be considered as the irreversibility analysis is
performed on both flow streams considering the heat transfer effect in form of sates difference between inlet and outlet
streams.

Heat Exchanger Modeling

Based on the defined highest temperature of the gas turbine plant, the heat rate required to be supplied to the working
fluid through the heat exchanger can be calculated as follows:
𝑇
𝑄̇𝐻𝐸 = 𝑚̇ 𝑐𝑝 𝑇𝐻 �1 − 𝑅,𝑐,𝑒𝑥 � (31)
𝑇 𝐻
The working fluid, after being heated by the hot, low pressure, stream at the recuperator, it passes through the main
heat exchanger which provides the required amount of heat to reach the turbine inlet temperature with enough
potential to gain the required power. The following equation expresses the destruction of exergy associated with this
process:
𝑒𝑥
− ∫𝑖𝑛 𝑇𝑑𝑠�
∆𝑃
𝐸𝑥̇𝑑,𝐻𝐸 = 𝑚̇ 𝑇𝑜 ��
𝑃𝐻
� + �𝑠𝐻𝐸,𝑒𝑥 − 𝑠𝐻𝐸,𝑖𝑛 ��𝑃 + 𝑅 � 𝑃 � � (32)
𝑇𝐻𝐸 𝐻 𝐻𝐸

The first term in the brackets represents the specific heat gained by the working fluid considering the heat transfer to
happen at constant pressure, where:
𝑒𝑥
∫𝑖𝑛 𝑇𝑑𝑠� = 𝑞𝐻𝐸 (33)
𝑃𝐻

697
In the exergy destruction equation, the external irreversibility in the heat exchanger is expressed by the first two terms
in the brackets. This part of irreversibility is associated with the heat transfer to the working fluid within the temperature
difference between the source temperature 𝑇𝐻𝐸 , and the working fluid bulk temperature through the heat exchanger,
considering no pressure loss to occur during the process. To add the effect of the pressure drop, the internal
irreversibility is added in terms of pressure loss across the heat exchanger for ideal gas case as shown in the last term
of the equation.
The temperature 𝑇𝐻𝐸 is the heating source temperature and is calculated, in analogous with the open gas
turbine engine with combustion, based on the definition of the optimal effective flame temperature of the combustion
products. It can be formulated as function of the heating molten salt temperature as follows:
𝑇 − 𝑇𝑜
𝑇𝐻𝐸 = 𝑙𝑛� 𝑇𝑠,𝑖𝑛 (34)
𝑠,𝑖𝑛 ⁄ 𝑇𝑜 �
More details regarding the calculation of the optimal flame temperature can be found elsewhere (Bejan 2006).

Precooler Modeling
The heat released from the low pressure stream at the intercooler is determined as follows, based on the energy
balance on this stream:
𝑄̇𝑃𝐶 = 𝑚̇ 𝑐𝑝 �𝑇𝑅,ℎ,𝑒𝑥 − 𝑇𝐻 𝛽� (35)
In the same way of calculating the exergy destruction in the intercooler, the exergy destruction in the precooler can be
formed as follows:
𝑄 ̇ ∆𝑃
𝐸𝑥̇𝑑,𝑃𝐶 = � 𝑇 𝑃𝐶
⁄𝑇
� − 𝑚̇ 𝑇𝑜 ��𝑠𝑃𝐶,𝑒𝑥 − 𝑠𝑃𝐶,𝑖𝑛 ��𝑃 − 𝑅 � 𝑃 � � (36)
𝑃𝐶 𝑜 𝐿 𝑃𝐶
Here, the term �𝑠𝑃𝐶,𝑒𝑥 − 𝑠𝑃𝐶,𝑖𝑛 ��𝑃𝑙 is the entropy difference considering the process through the intercooler occurs at
constant pressure with no pressure losses. 𝑅(∆𝑃⁄𝑃)𝑃𝐶 represents the entropy change as a result of the pressure
loss at the working fluid side, considering ideal gas behavior.

Energetic Assessment
The thermal energy efficiency can be formulated, based on the shown analysis, in form of the gas turbine thermal
design parameters for a closed recuperative gas turbine cycle with two compression stages as follows:
𝑘
𝑟𝑘 −1−𝑘∙∑(∆𝑃⁄𝑃) 1 1 𝛽���𝑟𝑐 � −1�
� 𝑐 𝑘 �− � + �� �
𝑟𝑐 / 𝜂𝑇,𝑎 𝜂𝐿𝐶,𝑎 𝜂𝐻𝐶,𝑎 𝜂𝑚𝑒𝑐ℎ
𝜂𝑡ℎ = 𝑘
(37)
��𝑟𝑐 � −1 1+𝑘∙∑(∆𝑃⁄𝑃)
1−�𝛽(1−𝜀𝑅 )�1+ �+𝜀𝑅 �1− 𝑘 ��
𝜂𝐻𝐶,𝑎 𝑟𝑐 / 𝜂𝑇,𝑎

The optimal pressure ratio of the plant is determined at maximum thermal efficiency, by differentiating the previous
formula with respect to pressure ratio, 𝑟𝑐 , setting 𝑑𝜂𝑡ℎ ⁄𝑑𝑟𝑐 to zero for maximum efficiency. The pressure ratio for
each compressor stage in this relation are assumed to be identical, however, with considering the pressure loss in the
intercooler, they will be slightly different as mentioned before, based on the given definition of the higher and lower
pressure values of the gas turbine cycle.
Another parameter used to describe the system performance is the dimensionless net work output of the gas
turbine plant. The dimensionless work is to be calculated by relating the actual specific work output to the total
potential of producing work expressed in terms of the temperature ratio and the turbine inlet temperature as follows:
𝑤𝑛𝑒𝑡
𝑊𝑑𝑖𝑚𝑙𝑒𝑠𝑠 = 𝑐 (38)
𝑝 𝑇𝐻 (1−𝛽)
The dimensionless work is used to express the results of the energetic performance of the system.

Exergetic Assessment
The helium turbine plant performance can be described in different exergy parameters. The exergy efficiency of the
overall plant performance can be expresses in terms of the useful exergy output of the plant over the exergy of the fuel
supplied in form of exergy of the heat added to the system:
𝐸𝑥̇𝑜𝑢𝑡𝑝𝑢𝑡
𝜓= (39)
𝐸𝑥̇𝑓
It can be also estimated based on the exergy destruction calculation as follows:
∑ 𝐸𝑥̇𝑑
𝜓 =1− (40)
𝐸𝑥̇𝑓
Each of the exergy destruction rates in system components is described as percentage of the total exergy destruction:

698
 𝐸𝑥̇𝑑
𝑦𝐷 = ∑ (41)
𝐸𝑥̇𝑑
The exergy destruction can be expressed by the exergy destruction ratio as it is defined by relating the exergy
destruction rate in each component to the total exergy rate of the fuel to the system.
𝐸𝑥̇𝑑
𝐸𝑥𝑅 = ∑ (42)
𝐸𝑥̇𝐹
Another parameter is the sustainability index, which helps in giving an indication of the improvement potential for more
efficient use of the available resources of the system as follows:
1
𝑆𝐼 = 1−𝜓 (43)

The improvement potential for each of the system components can be measured in energy units based on the
following relation (Gool 1997):
𝐼𝑃 = (1 − 𝜓)�𝐸𝑥̇𝑖𝑛 − 𝐸𝑥̇𝑒𝑥 � (44)
And it can be also related to the plant net power output in a percentage form.

Case Study
For verifying the modeling code of the helium closed cycle gas turbine presented in this study, the results for a case
study are compared in the following section with data of a helium gas turbine operating with power reactor as
presented in (Conn and Kuo 1976). The turbine inlet conditions are 871oC and 69 bar. The compressor inlet
temperature and pressure are 43.3oC at 26.2 bar. This gas turbine is of 500 MW power and operates at 46% of
thermal efficiency. The adiabatic efficiency of the turbine and the high and low pressure compressors are 91, 89 and
89%, respectively. The effectiveness of the recuperator, intercooler and precooler are 88, 88 and 92%, respectively.
Considering the same losses in the heat exchangers streams, Table 2 shows the comparison of the results of
thermodynamic properties and parameters of the current model compared with the data provided in the literature.

Table 2: Thermodynamic simulation model verification.

Parameter Current Model Literature data Difference
Specific Heat in Regenerator 1802 kJ/kg 1883 kJ/kg 4.30 %
Specific Heat in Intercooler 427.2 kJ/kg 428.6 kJ/kg 0.32 %
Specific Heat in Precooler 683.1 kJ/kg 685.9 kJ/kg 0.41 %
Overall Thermal Efficiency 45.68 % 46 % 0.70 %
Temperature after compression 127.5 oC 125.5 oC 1.59 %
Temperature after expansion 521.9 oC 537.7 oC 2.93 %

Table 3: Results of the energy and exergy analysis of the system components.

Exd yD ηex Exd yD ηex Exd yD ηex

[MW] [%] [%] [MW] [%] [%] [MW] [%] [%]
β = 313K/923K β = 313K/1023K β = 313K/1123K
Gas Turbine 9.067 7.56 96.8 7.535 7.91 97.1 6.467 8.23 97.3
LP Compressor 7.420 6.26 90.4 6.141 6.44 90.6 5.254 6.71 90.6
HP Compressor 7.644 6.39 90.5 6.308 6.63 90.5 5.386 6.87 90.6
Intercooler 12.24 10.4 97.6 10.80 11.4 97.2 9.701 12.4 96.8
Precooler 23.37 19.5 95.1 20.71 21.7 94.1 18.84 24.1 93.0
Recuperator 16.34 13.6 87.9 13.70 14.4 88.7 11.90 15.2 89.3
Heat Exchanger 43.41 36.2 81.9 30.01 31.5 86.0 20.76 26.5 89.5

RESULTS AND DISCUSSION

The energy and exergy analyses are performed on the prescribed system in the base case operating conditions as
shown in Table 1. The results of the exergy analysis of the system components are given in Table 3. Different plant
high-temperature values are studied, and the optimal operating pressure ratio is calculated based on the energy model
of the system. The results in Table 3 show the exergy destruction rates and their ratios as percentage of the total
exergy destruction of the system. The exergy efficiency calculations are also shown. Table 4 shows the results of the
energy and exergy analyses of the overall system performance. The results show that for the same compressor inlet
temperature, increasing the turbine inlet temperature results in a higher energy and exergy efficiency as a result of the
higher operating optimal pressure ratio of the gas turbine plant. In the same Table, the increase of the specific work
output, which reflects the enthalpy potential of producing useful work, is shown. This causes a decrease in the helium
flow rate demand, for the same power output. This highly affects the economic viability of the heat exchangers and
turbomachinery as the mass flow rate is to be increased by 40% for a decrease in the temperature ratio from 2.8 to 2.4
at the same compressor inlet temperature.

699
 Figures 3-5 show the results of the parametric study of the pressure ratio effect on different performance
parameters. The results in these figures are calculated at the base condition listed in Table 1. Fig. 3 shows the change
of the energy and exergy efficiencies with the pressure ratio, the energy efficiency shows a maximum of 45.48% at a
pressure ratio of 2.74, however, the exergy efficiency reaches a maximum of 60.6% at pressure ratio of 2.78. The
dimensionless work is also shown in the same figure; it is as an indication of the produced power with respect to the
theoretical maximum potential. It gives a peak at the pressure ratio of 6.2. At this value, the energy and exergy
efficiency drops to 39.1 and 52%, respectively.

Table 4: Overall performance assessment based on the thermal analyses

o o o
Turbine Inlet temperature 650 C 750 C 850 C
Optimal pressure ratio 2.407 2.648 2.74
Temperature ratio 0.34 0.306 0.278
Overall System
performance

Thermal efficiency 37.64% 41.91% 45.48%

Exergy Efficiency 50.07% 55.75% 60.51%
Helium flow rate 211 kg/s 164 kg/s 128.6 kg/s
Specific net work 545.7 kJ/kg 732 kJ/kg 933 kJ/kg
Sustainability index 2.0 2.3 2.5
Improvement potential 59.74 MW 42.14 MW 30.93 MW
Intercooler flow rate 251.0 kg/s 190.8 kg/s 152 kg/s
Cooling
water

Precooler flow rate 264.9 kg/s 200 kg/s 158.3 kg/s

o o o
Intercooler exit temp 98.4 C 105.7 C 112.4 C
o o o
Precooler exit temp 133.3 C 145 C 156 C

The effects of the pressure ratio on components exergy destruction rates are shown in Figs. 4 and 5 in
different representations. In Fig. 4, it shows the exergy destruction rate in MW. The recuperator exergy destruction rate
decreases with the increase in pressure ratio. The total exergy destruction reaches a minimum at the point of maximum
exergy efficiency. Fig. 5 shows the change of exergy destruction in the system components, as percentage of the total
exergy destruction, with increasing the pressure ratio. For a pressure ratio more than 3.4, the changes in the
percentages for the turbine, compressors, main heat exchanger and the precooler are very small compared with the
other heat exchangers of the system.
Figures 6-9 shows the results of the energy and exergy analyses at different temperature ratios with respect
to the turbine inlet temperature. All other operating parameters are kept constant at the base case condition shown in
Table 1. Fig. 6 shows the effect of changing the pressure ratio on the dimensionless work and the total exergy
destruction rate of the plant. Increasing the temperature ratio at constant turbine inlet temperature causes a decrease in
the work output as the enthalpy difference drops. The maximum work point occurs at higher pressure ratio the higher
turbine inlet temperature. The total exergy destruction rate has a fully reversed trend.

80
Exergy Efficiency
70 Energy Efficiency
Dimentionless Parameter, %

Dimless Net Work

60

50

40

30

20

10

0
1.2 2.2 3.2 4.2 5.2 6.2 7.2 8.2
Pressure Ratio
Fig. 3 Effect of pressure ratio on the overall system performance.

700
 80 130
Turbine
70 Compressor 120

Total Exergy Destruction Rate, MW

Heat Exchanger
Exergy Destruction Rate, MW

60 Intercooler 110
Precooler
50 Recuperator 100
Total
40 90

30 80

20 70

10 60

0 50
1.2 2.2 3.2 4.2 5.2 6.2 7.2 8.2
Pressure Ratio
Fig. 4 Effect of pressure ratio on exergy destruction rates in the system components and the overall plant.

50

45 Turbine Compressor
Exergy Destruction percentage, %

Heat Exchanger Intercooler

40 Precooler Recuperator
35

30

25

20

15

10

0
1.2 2.2 3.2 4.2 5.2 6.2 7.2 8.2
Pressure Ratio

Fig. 5 Effect of pressure ratio on exergetic performance in form of exergy destruction rates percentage of the total exergy
destruction.
35 500
650°C
450
Total Exergy Destruction Rate, MW

30 750°C
850°C 400
25
Dimless Work, %

350
20 300

15 250

200
10
150
5
100

0 50
1.2 2.2 3.2 4.2 5.2 6.2 7.2 8.2
Pressure Ratio
Fig. 6 Pressure ratio effect on the overall system performance at different turbine inlet temperature.

In Fig. 7, as a result of the increase in work output, the exergy efficiency is higher for higher turbine
temperature. But with the increase of pressure ratio, even with the increase in work output, the exergy efficiency drops
after the peak as a result of the increase in the exergy rate of streams provided to the system with respect to the
product exergy rates. Fig. 8 shows the exergetic performance of the studied helium turbine in a new method of
performance measurement. The figure is introduced as the exergetic performance map. The exergetic performance
map shows the effect of pressure ratio and temperature ratio on the performance of the gas turbine and its output. This
performance map measures the performance in terms of the exergy efficiency and the dimensionless work. It can be
also introduced with respect to the specific work output. This map gives an idea of the potential of enhancement in the

701
system performance. Fig. 9 also shows a new exergetic map of the gas turbine. This map is the exergetic loss map
which shows the exergy destruction ratio, with respect to the exergy of the fuel streams provided to the system
components, and the specific net work output. The gas turbine operation is described on the map at constant lines of
temperature or temperature ratio at different operating pressure ratio.
70
65 650°C
750°C
60
850°C
Exergy Efficiency, %

55
50
45
40
35
30
25
20
1.2 2.2 3.2 4.2 5.2 6.2 7.2 8.2
Pressure Ratio
Fig. 7 Change in overall exergy efficiency with pressure ratio at different turbine inlet temperature.

70

60 2.8
Exergy Efficiency, %

1.4
2.64 6.2
50 1.4
3.372 5.3
8.2
rc=1.4 4.47
40 8.2
650°C
30 750°C
8.2
850°C

20
10 15 20 25 30
Dimless Work, %
Fig. 8 Exergetic performance map at different turbine inlet temperature.

14
13 650°C 8.2
Exergy Destruction Ratio, %

12 750°C 8.2
850°C 8.2
11
10
9 4.47 5.3 6.2

8
7
6
rc =1.4
5 1.4 1.4
4
200 400 600 800 1000 1200
Specific Work, kJ/kg
Fig. 9 Exergetic loss map at different turbine inlet temperature.

CONCLUSIONS
Energy and exergy analyses are conducted for an actual recuperated gas turbine, working on a closed cycle with
helium as the working fluid. The energy and exergy efficiencies of the cycle at temperature ratio of 0.278, turbine inlet
condition of 850oC and 70 bar as a base case for a calculated optimal pressure ratio of 2.74 are 45.48% and 60.51%
respectively. Increasing the high-temperature of the cycle increases the optimal pressure ratio and also increases the
plant performance.

702
NOMENCLATURE
cp Specific heat capacity, kJ/kg.K
E Energy, kJ
𝐸𝑥 ̇ Exergy rate, kW
𝐸𝑥𝑑 ̇ Exergy destruction rate, kW
𝐸𝑥𝑅 Exergy destruction ratio, %
h Specific enthalpy, kJ/kg
IP Improvement potential, kW
k Constant
𝑚̇ Mass flow rate, kg/s
Nc Number of compression stages
P Pressure, kPa
(∆𝑃 ⁄𝑃 ) Pressure loss coefficient
𝑄̇ Heat rate, kJ/s
rc Pressure ratio
R Gas constant, kJ/kg.K
s Specific entropy, kJ/kg.K
𝑆̇ Entropy rate, kW/K
SI Sustainability index
t Time, s
T Temperature, K
w Specific work, kJ/kg
𝑊̇ Power, kW
yD Exergy destruction percentage, %
β Temperature ratio
ε Heat exchanger effectiveness, %
η Energy efficiency, %
ψ Exergy efficiency, %
ϒ Specific heat ratio
Subscript
a Adiabatic
c Cold steam
C Compressor
ex Exit flow
f Fuel
gen Generation
h Hot stream
H Highest
HE Heat exchanger
in Inlet flow
IC Intercooler
L lowest
p Polytropic
PC Precooler
R Recuperator
T Turbine

REFERENCES
Bejan A. 2006. Advanced Engineering Thermodynamics: Wiley.
Boyce MP. 2006. Gas Turbine Engineering Handbook: Gulf Professional Publishing.
Conn RW, Kuo SC 1976. An advanced conceptual tokamak fusion power reactor utilizing closed cycle helium gas
turbines. Nuclear Engineering and Design 39: 45-72.
Frutschi HU. 2005. Closed-Cycle Gas Turbines: ASME.
Gandhidasan P 1993. Thermodynamic Analysis of a Closed-Cycle, Solar Gas-Turbine Plant. Energy Conversion and
Management 34: 657-661.
Gool W. 1997. Energy Policy: Fairy Tales and Factualities. In: Soares OD, Cruz AM, Pereira GC, Soares IRT, Reis
APS, editors. Innovation and Technology — Strategies and Policies: Springer Netherlands. p. 93-105.

703
Horlock JH. 2002. Combined Power Plants: Krieger.
Kestin J 1980. Availability, the concept and associated terminology. Energy 5: 679-692.
McDonald CF 2012. Helium turbomachinery operating experience from gas turbine power plants and test facilities.
Applied Thermal Engineering 44: 108-142.
McDonald CF, Orlando RJ, Cotzas GM editors. International Joint Power Generation Conference. 1994 Phoenix,
Arizona.
Moran MJ, Shapiro HN, Boettner DD, Bailey MB. 2011. Fundamentals of Engineering Thermodynamics: Wiely.
No HC, Kim JH, Kim HM 2007. A Review of Helium Gas Turbine Technology for High-Temperature Gas-Cooled
Reactors. Nuclear Engineering and Technology 39: 21-30.
Qin J, Zhou W, Bao W, Yu D 2010. Thermodynamic analysis and parametric study of a closed Brayton cycle thermal
management system for scramjet. International Journal of Hydrogen Energy 35: 356-364.
Wenlong L, Heng X, Zuoyi Z, Yujie D 2012. Development and primary verification of a transient analysis software for
high temperature gas-cooled reactor helium turbine power system. Nuclear Engineering and Design 250: 219-228.
Zhao H, Peterson PF 2008. Multiple reheat helium Brayton cycles for sodium cooled fast reactors. Nuclear
Engineering and Design 238: 1535-1546.

704
 COMPARISON OF DIFFERENT TURBULENT MODELS IN TURBULENT FORCED
CONVECTIVE FLOW AND HEAT TRANSFER INSIDE RECTANGULAR CROSS-SECTIONED
DUCT HEATING AT THE BOTTOM WALL

1 2
Kamil ARSLAN , Nevzat ONUR
Çankırı Karatekin University, Technical and Business College, 18200, Çankırı, Turkey
1
2
Çankaya University, Faculty of Engineering, Mechanical Engineering Department, 06810, Ankara, Turkey
kamilarslan@karatekin.edu.tr, nevonur@cankaya.edu.tr

ABSTRACT
In this study, steady-state turbulent forced flow and heat transfer in a horizontal smooth rectangular cross-
sectioned duct was numerically investigated. The study was carried out in the turbulent flow condition where
Reynolds number ranges from 1x104 to 5x104. Flow was hydrodynamically and thermally developing, i.e.
simultaneously developing flow, under uniform bottom surface temperature condition. A commercial CFD program
Ansys Fluent 12.1 with different turbulent models was used to carry out the numerical study. Different turbulence
models (k-ɛ Standard, k-ɛ Realizable, k-ɛ RNG, k-ω Standard and k-ω SST) were used. Based on the present
numerical solutions, new engineering correlations were presented for the heat transfer and friction coefficients.
The numerical results for different turbulence models were compared with each other and the experimental data
available in the literature. It was obtained that k-ɛ turbulence models represented the turbulent flow condition very
well for the present study.

Keywords: Simultaneously developing flow, forced convection, heat transfer, rectangular cross-sectioned duct,
turbulent flow, CFD

INTRODUCTION
Predicting the pressure drop and heat transfer under hydrodynamically and thermally developing flow condition is
quite important in many applications such as compact heat exchangers where flow passages are typically short in
length. Flow and heat transfer in rectangular channels have received considerable attention due to their practical
importance. Especially, turbulent forced convection inside rectangular cross-sectioned ducts is of interest in the
design of a high-Reynolds number heat exchanger apparatus. Ducts having rectangular cross-sections are widely
used in industrial heat transfer equipment such as compact heat exchangers. They have been widely used in
various industries because of their high heat transfer area on the basis of unit volume and good mechanical
strength even with very thin foils. These devices are required to be compact as well as to transfer heat rapidly to
the environment. Among the various duct cross sections, rectangular cross-sectioned ducts are used in compact
heat exchanger extensively, because of its simplicity of construction. Compared to circular cross-sectioned ducts,
heat transfer and fluid flow in rectangular cross-sectioned ducts are complicated and dependent on several
parameters.

Several studies of flow in straight rectangular cross-sectioned ducts were presented in the past. An excellent
comprehensive review of forced convection flow in rectangular ducts was presented by Shah and London (1978),
Kakaç et al. (1987), and Kakaç and Liu (2002). Zhang (2007) numerically analyzed hydrodynamically developed
thermally developing forced convection heat transfer in rectangular ducts under uniform plate temperature. Bottom
and top plates of the rectangular duct were kept constant temperature, at the same time different conductance
materials were chosen for side walls. Various duct aspect ratios were considered, and the side walls had different
conductance parameters from zero to infinitely large. The forced convection heat transfer parameters for
rectangular ducts were determined. Muzychka and Yovanovich (2004) investigated laminar combined developing
flow in non-circular cross-sectioned ducts. New models were proposed which simplify the prediction of the Nusselt
number for combined entry region in most non-circular duct geometries. This new model predicts both local and
average Nusselt numbers and is valid for both isothermal and isoflux boundary conditions. Chung et al. (1992)
investigated numerically behavior of hydrodynamically fully developed and thermally developing laminar flow of
Newtonian fluids in rectangular ducts for the constant heat flux boundary condition. The developing Nusselt
number was obtained for a wide range of duct aspect ratios. Aparecido and Cotta (1990) analytically studied
laminar forced convection inside a rectangular duct by extending the generalized integral transform technique.
Fully developed and thermally developing regions were examined. Beavers et al. (1970) investigated
experimentally hydrodynamic flow development in rectangular ducts for different aspect ratio. The concern of the
research was the determination of the pressure field associated with the hydrodynamic development of the flow in
rectangular duct. Montgomery and Wibulswas (1967) developed an alternative numerical method which was used
to solve the heat transfer equations for laminar flow in ducts of rectangular cross-sectioned with simultaneously

705
developing temperature and velocity profiles, both for constant wall temperature and for constant heat input per
unit length of the duct. Comparisons were made with analytical solutions for circular ducts and parallel plates with
experimental data. It was obtained that Prandtl number had a strong effect on the Nusselt number for
simultaneously developing profiles. Sparrow (1955) studied laminar simultaneous development of temperature and
velocity profiles in the entrance region of a flat rectangular duct for different Prandtl numbers. Two temperature
conditions at the duct walls were considered: both duct walls had the same uniform temperature and one of the
duct walls was maintained at a uniform temperature while the other wall was insulated. The results were given
graphically for different boundary conditions, Reynolds number, and Prandtl numbers. Savino and Siegel (1964)
determined the temperature distributions analytically for fully developed heat transfer in rectangular ducts with
different aspect ratios. The channel walls were uniformly heated, but the heat flux on the short sides was an
arbitrary fraction of the heat flux on the broad sides. Chiu et al. (2007) numerically examined the mixed convection
heat transfer in rectangular cross-sectioned duct with radiation effects. The vorticity–velocity method was
employed to solve the three-dimensional Navier–Stokes and energy equation simultaneously. It was obtained that
radiation effects had a considerable impact on the heat transfer and would reduce the thermal buoyancy effects.
Naccache and Mendes (1996) examined the heat transfer to non-Newtonian fluids flowing laminarly through
rectangular cross-sectined ducts. It was observed in the numerical investigation that shear thinning yields
negligible heat transfer enhancement effect when compared with the secondary flow effect. Chong et al. (2011)
experimentally investigated the laminar and transition opposing mixed convection in the entrance region through a
rectangular cross-sectioned duct. The friction factors were calculated and analyzed under different Reynolds
numbers, Grashof numbers, and inclination angles. The results indicated that the friction factors showed different
characteristics when the Reynolds numbers were larger or smaller than 1500. Gnielinski (1976) modified the
widely accepted turbulent flow correlations which provide fully developed heat transfer coefficients for flows with
Reynolds numbers exceeding 1x104. It was noted that the modified equation purported to be applicable to
Reynolds numbers as low as 2300. The vast majority of the data were mean values of the heat transfer coefficient
rather than fully developed values, reflecting the early era in internal convection investigations. Kostic and Hartnett
(1986) reported an experimental investigation for the turbulent flow of water in a 2:1 asymmetrically heating
rectangular duct. The results of the study was compared with the literature results. Abraham et al. (2009)
presented a predictive theory which was capable of providing quantitative results for the heat transfer coefficients
in round pipes for the three possible flow regimes: laminar, transitional, and turbulent. The theory was based on a
model of laminar-to-turbulent transition which was also viable for purely laminar and purely turbulent flow.

Heat and momentum transfer processes in the simultaneously developing flow of rectangular cross-sectioned
channels are very complex. Basic knowledge on the flow and heat transfer of the three-dimensional turbulent
forced convection in rectangular cross-sectioned duct is needed for the design of thermal equipment. However, it
is seen from literature survey that in spite of being commonly used in engineering applications, the rectangular
geometry under constant bottom surface temperature and turbulent flow condition has not been adequately
studied in depth despite its importance.

The present study concerns three-dimensional numerical study of flow in an isothermally heated bottom surface
horizontal straight rectangular cross-sectioned duct under hydrodynamically and thermally developing turbulent
flow condition. Air (Pr≅0.7) was used as the heat transfer medium. The momentum, continuity, energy, and
turbulence equations for three-dimensional flow in the hydrodynamic and thermal entrance region of a rectangular
cross-sectioned duct were solved using finite volume based commercial software Ansys Fluent 12.1. Five
turbulence models (k-ɛ Standard, k-ɛ Realizable, k-ɛ RNG, k-ω Standard and k-ω SST) were used in numerical
calculations. Practical engineering correlations for the average Nusselt number and average Darcy friction factor
were determined. Local Nusselt number and local Darcy friction factor for different Reynolds numbers along the
flow direction were obtained. Furthermore, the results of the numerical investigation with different turbulence
models were compared with each other and results of the literature in this study.

THEORETICAL DESCRIPTION
A schematic diagram depicting the computational domain of the rectangular cross-sectioned duct along with the
coordinate system and flow geometry is presented in Fig. 1a. Rectangular cross-sectioned duct was
mathematically modeled for numerical computations. Since the flow field was symmetric with respect to y-x plane,
only one half of the channels was considered for computational domain to reduce the computational time. The
principle flow was in the x-direction. Turbulent flow enters the duct with a uniform velocity and temperature profile.
The three dimensional Navier-Stokes, energy, and turbulence equations were used to describe the flow and heat
transfer in the computational domain. The three-dimensional incompressible Newtonian flow with negligible
buoyancy effects and viscous dissipation was regarded as turbulent and steady; the physical properties of fluid,
taken at the bulk temperature, were considered to be constant in the duct.

706
 The continuity, momentum, energy, and turbulence equations were solved by Ansys Fluent 12.1 in the duct.
Turbulence intensity levels used at the inlet of the duct varied from 4 to 5 % depending on the Reynolds number.
Hydraulic diameter of the duct was used as appropriate length scale at the inlet. No slip boundary conditions were
employed on the duct walls. A uniform wall temperature boundary condition was employed on the bottom surface
of the duct. Insulated boundary condition was applied on the sides and top surface of the duct. At the outlet of the
duct, pressure outlet boundary condition of Ansys Fluent 12.1 was used. Symmetry boundary condition was
employed on the symmetry planes. All the boundary conditions applied on the duct were depicted in Fig. 1b.

a b
x
q′′w = 0

5 cm L = 58 cm symmetry
y x
Vi, Ti Tw Tw
15 cm z
z
c

Fig.1. (a) Computational domain of the rectangular cross-sectioned duct, (b) boundary condition of the duct, (c)
mesh distribution in y-z plane

The objective of the data reduction is to calculate the average Nusselt number and Reynolds number along with
average Darcy friction factor. The hydraulic diameter ( Dh = 4Ac / P ) is chosen as the characteristic length.

Re= U ⋅ Dh / ν (1)
Nu= h ⋅ Dh / k (2)
f = ∆ p( Dh / L ) /  ρU / 2 
2
(3)

In Eqs. (1)-(3), U (ms-1) is the mean velocity of the air flow in the rectangular cross-sectioned duct, Dh (m) is the
hydraulic diameter of the rectangular duct, ν (m2s-1) is the kinematic viscosity of air, h (Wm-2K-1) is the average
heat transfer coefficient, and k (Wm-1K-1) is the thermal conductivity of air. In addition, ∆p (Pa) is the pressure drop
along the duct, L (m) is the axial length of the duct, and ρ (kgm-3) is the density of air.

Average convective heat transfer coefficient of the air flow in the rectangular cross-sectioned duct is obtained
(Çengel, 1998) as follows

=  / [ A(T − T )]
h Q (4)
c w b

where, A (m2) is the surface area (i.e. in contact with the air) of the bottom surface of the duct and Tw (K) is the
surface temperature of the bottom surface of the duct. Tb (K) is the mean bulk temperature of the air flow in the
 (W) is the convective heat transfer from the duct’s bottom surface to the air flowing through the duct
duct. Qc

under steady-state condition and is calculated as follows (Çengel, 1998)

=
Q ρUA c c p (Tbo − Tbi ) (5)
c

707
where Ac (m2) is the cross sectional area of the rectangular cross-sectioned duct, cp (Jkg-1K-1) is the specific heat
of the air. Tbi (K) and Tbo (K) are bulk temperatures of the air at the inlet and outlet of the duct, respectively. All fluid
properties in the duct were evaluated at the bulk temperature; Tb=(Tbi+Tbo)/2. The air flow properties were taken
from Incropera and DeWitt (2002).

In this study, a general purpose finite-volume based commercial CFD software package Ansys Fluent 12.1 was
used to carry out the numerical study. The code provides mesh flexibility by structured and unstructured meshes.
Ansys Fluent 12.1 includes several turbulence models.

Computations were performed under turbulent flow condition. The energy equation was solved neglecting radiation
effects. Five different turbulence models (k-ɛ Standard, k-ɛ Realizable, k-ɛ RNG, k-ω Standard and k-ω SST) were
used in this study for solving the flow in this study. The Reynolds averaged Navier-Stokes equations were solved
numerically in conjunction with transport equations for turbulent flow. Near wall regions were fully resolved for
average y+ values between 0.5 and 1.1 in all the calculations, sufficiently resolving the laminar sub-layer (i.e. y+≤4-
5). In the present study, hexahedral cells were created with a fine mesh near the duct walls. A non-uniform grid
distribution was employed in the plane perpendicular to the main flow direction while uniform grid distribution was
employed in main flow direction as shown in Fig. 1c. Close to each wall, the number of grid cells or control
volumes was increased to enhance the resolution and accuracy. Steady segregated solver was used with second
order upwind scheme for convective terms in the mass, momentum, energy, and turbulence equations. For
pressure discretization, the standard scheme was employed while the SIMPLE-algorithm (Patankar, 1980) was
used for pressure-velocity coupling discretization.

The grid independence study was performed by refining the grid size until the variation in both average Nusselt
number and average Darcy friction factor are less than 0.18 % and 0.09 % respectively. To obtain the optimum
mesh size, a grid independence study was conducted using six different grid sizes changing from 1.8x105 to
17.25x105 for Re=5x104. Each mesh was processed using the k-ɛ Standard turbulence model. It was observed
that a further refinement of grids from 2.8x105 to 4.0x105 did not have a significant effect on the results in terms of
average Nusselt number and average Darcy friction factor as depicted in Table 1. Based on this observation, grid
size of 2.8x105 was used for all of the calculations. No convergence problems were observed during the
calculations. Same procedure was used for other turbulence models, and optimum grid size was obtained for each
turbulence model calculations. To obtain convergence, each equation for mass, momentum, and turbulence was
iterated until the residual falls below 1x10-5 while energy equation was iterated until the residual falls below 1x10-6.

Table 1. Grid size effect for average Nusselt number and average Darcy friction factor for k-ɛ Standard turbulence
model
Grid size Nu f
1.80x105 137.7626 0.03296
5
2.80x10 138.0095 0.03293
4.00x105 138.1041 0.03293
6.45x105 138.1762 0.03296
10.50x105 138.2819 0.03299
17.25x105 138.3207 0.03302

RESULTS AND DISCUSSION

In the study reported here, the convective heat transfer and fluid friction in an air-cooled rectangular cross-
sectioned duct under uniform bottom surface temperature were numerically investigated. The investigation was
carried out under hydrodynamically and thermally developing turbulent flow condition. The results were presented
in non-dimensional Nusselt number and Darcy friction factor. After the determination of temperature fields in the
fluid, the average Nusselt number was calculated. In addition, average Darcy friction factor was estimated with the
determination of pressure drop in the duct. Numerical results obtained under steady-state conditions are presented
in Fig. 2 through Fig. 6 for different turbulence models. Plotted in these figures are the best-fit lines. The flow and
temperature distributions, numerically obtained average Nusselt numbers and average Darcy friction factors
presented in this study highlight the influence of duct geometry and wall boundary conditions on thermal
performance of rectangular cross-sectioned duct and provide additional useful design data.

Air at the ambient temperature entered the duct with uniform velocity profile. The velocity fields for different
turbulence models at x=0.2 m of the duct entrance were considered first. Typical velocity magnitude contours for
different turbulence models at x=0.2 m are presented in Fig. 2 for Re=5x104. It is seen in Fig. 2 that there are small
differences for velocity contours at the entrance region of the duct for different turbulence models.

708
Fig.2. Isovel contours of velocity magnitude for different turbulence models at the entrance region of the duct

Velocity contours at the outlet of the duct is presented in Fig. 3 for Re=5x104 and different turbulence models. It is
obtained in the figure that velocity contours for k-ɛ turbulence models have almost same profile; however, velocity
profiles for k-ω turbulence models are slightly different from k-ɛ turbulence models.

Fig.3. Isovel contours of velocity magnitude at the outlet of the duct for different turbulence models

In order to compare the velocity distributions of different turbulence models, typical velocity profiles are obtained
for each turbulence model and these are plotted on symmetry plane at the entrance region for Re=5x104. As can
be seen from Fig. 4a, the velocity profile for different turbulence models in the rectangular duct is plotted as a
function of the dimensionless height (y/H) of the duct. It is obtained in this figure that velocity magnitudes of k-ɛ
and k-ω turbulence models have nearly same profiles; however, velocity profiles of k-ω turbulence models are
slightly different from velocity profiles of k-ɛ turbulence models. Dimensionless temperature distribution is plotted
as function of the dimensionless height (y/H) of the duct for various turbulence models on symmetry plane at the
entrance region as shown in Fig. 4b. Dimensionless temperature profile is defined as
Θ= (Tw − T ( x,y ,z ) ) (Tw − Ti ) where Tw (K) is the bottom wall temperature of the duct and Ti (K) is the inlet
temperature (Kakaç et al., 1987). It is seen in the figure that dimensionless temperature profile of k-ɛ turbulence
models are a bit different than dimensionless temperature profile of k-ω turbulence models.

709
 a
1.2 b

1.0
1.0

0.8 0.8

k-ε Standard k-ε Standard

0.6 k-ε Realizable 0.6 k-ε Realizable
y/H

y/H
k-ε RNG k-ε RNG
0.4 k-ω Standard 0.4 k-ω Standard
k-ω SST k-ω SST
0.2
0.2

0.0
0.0
-0.2
0 2 4 6 8 10 12 0.0 0.2 0.4 0.6 0.8 1.0

Velocity Magnitude (m/s) Dimensionless Temperature (Θ)

Fig.4. (a) Velocity and (b) dimensionless temperature distributions along the duct on symmetry plane

One of the key parameters of interest in the hydrodynamic and thermal entrance region is the variation of
numerically obtained local Darcy friction factor and local heat transfer coefficient for different turbulence models
along the axial position. Fig. 5a and Fig. 5b show the variation of local Darcy friction factor and local heat transfer
coefficient as function of dimensionless position (x/Dh) for the channel, respectively. Fig. 5a shows the local Darcy
friction factor as a function of dimensionless position (x/Dh) for different turbulence models. An examination of Fig.
5a reveals that local Darcy friction factor in the entrance is very large and decreasing along the duct for all
turbulence models. An inspection of Fig. 5b reveals that the values of local heat transfer coefficient in the inlet of
the duct is very large and is decreasing along the duct for various turbulence models. It is observed from an
inspection of Fig. 5a that local Darcy friction factor of each k-ɛ turbulence model is larger than local Darcy friction
factor of each k-ω turbulence model. Local heat transfer coefficient of each k-ɛ turbulence model is larger than
local heat transfer coefficient of each k-ω turbulence model. On the other hand, k-ɛ RNG turbulence model
predicts the largest value for local heat transfer coefficient.

a b
40x10-3
k-ε Standard k-ε Standard
k-ε Realizable 80 k-ε Realizable
-3
35x10 k-ε RNG k-ε RNG
k-ω Standard k-ω Standard
k-ω SST k-ω SST
30x10-3
60
fx

hx

25x10-3

40
20x10-3

15x10-3 20
0 2 4 6 0 2 4 6
x/Dh x/Dh

Fig.5. (a) Local Darcy friction factor and (b) local heat transfer coefficient, as function of dimensionless position
(x/Dh)

Fig. 6a and Fig. 6b display the average Nusselt number and the average Darcy friction factor in the duct along with
the available experimental data from the literature for Reynolds numbers ranging from 1x104 to 5x104. It is
observed that as the Reynolds number increases, the average Nusselt Number increases. In addition, average
Darcy friction factor decreases with increasing the Reynolds number. It is also seen from an inspection of the Fig.
6 that the Nusselt numbers obtained using the k-ɛ turbulence models are closer to each other and larger than the
Nusselt numbers obtained using k-ω turbulence models. At the same time, the average Darcy friction factors
obtained using k-ɛ turbulence models are slightly larger than those obtained using the k-ω turbulence models.

710
Average Nusselt numbers obtained with the present study are compared with the experimental correlations of the
Gnielinski (1976) and Hausen (1959) in Fig. 6a.

The average Darcy friction factors obtained with present study are compared with the experimental correlation of
smooth rectangular cross-sectioned duct given by Kakaç et al. (1987) as;

f=[1.0875-0.1125(H/W)]fc (6)

where fc is the average Darcy friction factor for circular cross-sectioned duct, suggested by Petukhov (1970) in Fig.
6b. H (m) and W (m) are the duct height and width, respectively.

A comparison of the average Nusselt numbers of present study with the experimental data available in literature
was carried out. It is observed that the results of k-ɛ turbulence models are very close the correlation of Gnielinski
(1976) and Hausen (1959) especially for small Reynolds numbers. On the other hand, the k-ω turbulence models
predict lower Nusselt numbers compared to experimental data available in the literature. The difference is nearly
16%. Average Darcy friction factors obtained using the k-ɛ turbulence models are nearly harmonious with the
experimental correlation of Kakaç et al. (1987). However, average Darcy friction factors obtained using of k-ω
turbulence models are approximately 6% lower than experimentally obtained average Darcy friction factors.

a b
250 Nuk-ε Standard fk-ε Standard
0.15
200 Nuk-ε Realizable fk-ε Realizable
Nuk-ε RNG fk-ε RNG
150 Nuk-ω Standard 0.10 fk-ω Standard
Nuk-ω SST 0.08 fk-ω SST
NuGnielinski (1976) fKakaç et al. (1987)
100 0.06
Nu

90 NuHausen (1959)
f

80 0.05
0.04
60
50 0.03

40
0.02
30
10000 15000 20000 25000 30000 40000 50000 60000 10000 15000 20000 25000 30000 40000 50000 60000
Re Re
Fig.6. (a) The changing of average Nusselt number with Reynolds number; (b) The changing of average Darcy
friction factor with Reynolds number

Numerically obtained average Nusselt numbers and average Darcy friction factors were expressed as a power law
variation with Reynolds number. In other words, results were presented in the forms of Nu=aReb and f=cRed. The
average Nusselt number, average Darcy friction factor, and Reynolds numbers for flow in this duct are based on
the hydraulic diameter Dh. The equations were obtained with least-square method for average Nusselt number and
average Darcy friction factor. The values of a, b, c, and d for Nu=aReb and f=cRed were given in Table 2.

Table 2. Values of a, b, c, and d for Nu=aReb and f=cRed

k-ɛ k-ɛ k-ɛ k-ω k-ω
Standard Realizable RNG Standard SST

a 0.0242 0.0243 0.0256 0.0213 0.0205

Nu
b 0.8 0.8 0.8 0.8 0.8
c 0.5190 0.5192 0.5190 0.5003 0.4905
f
d -0.25 -0.25 -0.25 -0.25 -0.25

CONCLUSIONS
In this study hydrodynamically and thermally developing three dimensional steady turbulent flow in a horizontal
rectangular cross-sectioned duct with constant bottom surface temperature and with three insulated sides were
numerically investigated for different turbulence models. Numerical study was carried out for Reynolds number
ranging from 1x104 to 5x104 with Pr=0.7. The turbulence models used in numerical simulations were k-ɛ Standard,
k-ɛ Realizable, k-ɛ RNG, k-ω Standard and k-ω SST. The results of numerical computations are presented in
terms of average Nusselt numbers and average Darcy friction factors. Results are given in Table 2 and in Fig. 2

711
through Fig. 6. It is shown that increasing the Reynolds number increases the average Nusselt number. On the
other hand, average Darcy friction factor decreases with increasing Reynolds number. Based on the present
numerical solutions of full 3D governing equations, new engineering correlations were obtained for the average
Nusselt number and average Darcy friction factor in the form of Nu=aReb and f=cRed, respectively. It is obtained
that k-ɛ turbulence models represent the turbulent region very well for the present study.

REFERENCES
Abraham, J. P., E. M. Sparrow, and J. C. K. Tong. 2009. Heat transfer in all pipe flow regimes: laminar,
transitional/intermittent, and turbulent. International Journal of Heat and Mass Transfer 52:557–563.

Aparecido, J. B., and R. M. Cotta. 1990. Thermally developing laminar flow inside rectangular ducts. International
Journal of Heat Mass Transfer 33(2):341-347.

Beavers, G. S., E. M. Sparrow, and R. A. Magnuson. 1970. Experiments on hydrodynamically developing flow in
rectangular ducts of arbitrary aspect ratio. International Journal of Heat and Mass Transfer 13:689-702.

Chiu, H. C., J. H. Jang, and W. M. Yan. 2007. Mixed convection heat transfer in horizontal rectangular ducts with
radiation effects. International Journal of Heat and Mass Transfer 50(15-16):2874-2882.

Chong, D. T., J. Liu, J. Yan. 2011. An experimental study of pressure drop and friction factor for laminar and
transition opposing mixed convection in entrance region of rectangular duct. Energy Conversion and Management
52:2272–2281.

Chung, B. T. F., Z. J. Zhang, and G. Li. 1992. Thermally developing convection from newtonian flow in rectangular
ducts with uniform heating. Journal of Thermophysics and Heat Transfer 7(3):534-536.

Çengel, Y. A. 1998. Heat Transfer a Practical Approach. USA:McGraw-Hill.

Gnielinski, V. 1976. New equations for heat and mass transfer in turbulent pipe and channel flow. International
Chemical Engineering 16:359-368.

Hausen, H. 1959. Neue gleichungen für die wärmeübertragung bei freier oder erzwungener stromüng (New
equations for heat transfer in free or forced flow). Allg. Wärmetechn 9(4/5):75-79.

Incropera, F. P., and D. P. DeWitt. 2002. Fundamentals of Heat and Mass Transfer. USA:John Wiley and Sons.

Kakaç, S., and H. Liu. 2002. Heat Exchangers Selection, Rating, And Thermal Design. USA: CRC Press.

Kakaç, S., R. K. Shah, and W. Aung. 1987. Handbook of Single-Phase Convective Heat Transfer. USA: John
Wiley and Sons.

Kostic, M., and J. P. Hartnett. 1986. Heat transfer to water flowing turbulently through a rectangular duct with
asymmetric heating. International Journal of Heat and Mass Transfer 29(8):1283-1291.

Montgomery, S. R., and P. Wibulswas. 1967. Laminar flow heat transfer for simultaneously developing velocity
and temperature profiles in ducts of rectangular cross section. Applied Science Research 18:247-259.

Muzychka, Y. S., and M. M. Yovanovich. 2004. Laminar forced convection heat transfer in the combined entry
region of non-circular ducts. Journal of Heat Transfer 126:54-61.

Naccache, M. F., and P. R. S. Mendes. 1996. Heat transfer to non-newtonian fluids in laminar flow through
rectangular ducts. International Journal of Heat and Fluid Flow 17(6):613-620.

Patankar, S. V. 1980. Numerical Heat Transfer and Fluid Flow. USA:Hemisphere Publishing Corporation.

Petukhov, B. S. 1970. Heat transfer and friction factor in turbulent pipe flow with variable physical properties.
Advances in Heat Transfer 6:503-565.

Savino, J. M., and R. Siegel. 1964. Laminar forced convection in rectangular channels with unequal heat addition
on adjacent sides. International Journal of Heat and Mass Transfer 7:733-741.

Shah, R. K. and A. L. London. 1978. Laminar Flow Forced Convection in Ducts. New York: Academic Press Inc.

712
Sparrow, E. M. 1955. Analysis of laminar forced convection heat transfer in entrance region of flat rectangular
ducts. NACA Technical Note 3331.

Zhang, L. 2007. Thermally developing forced convection and heat transfer in rectangular plate-fin passages under
uniform plate temperature. Numerical Heat Transfer Part A. 52:549-564.

713
 PERFROMANCE EVALUATION OF ECO-FRIENDLY HUDMIDIFYING MATERIAL USING
HWANGTO
1 2 3
Hyeon Ku Park , Seong Seok Go , Myeong Yeol Ryu
1. Sustainable Building Research Center, Hanyang University
55 Hanyangdaehak Ro, Sangnok Gu, Ansan Shi, Kyeonggi Do, 426-791, Korea
soundpark@cricmail.net
2. School of Architecture, Chonnam National University
300 Yongbong dong, Buk gu, Gwangju Shi, 500-757, Korea
3. Jigu Co.
985-16 Jagndeok dong, Gwangsan Gu, Gwangju Shi, 506-251, Korea

ABSTRACT
Indoor air in the winter season is very dry because of heating especially in Korea. When indoor air is dry, the problems
are arisen not only in the furniture and static electricity, but also in the residents’ health due to contaminated indoor air.
This is resulted in the use of artificial humidifier to humidify dry indoor air easily and continuously. Recently there were
accidents because of humidifier used sterilizing agent, and became social problem. This study aimed to develop
humidifying material using Hwangto called loess to substitute the artificial humidifier. The material developed is used
in the evaporating humidifier, therefore the effective as humidifier and environmental performances were examined by
test.

INTRODUCTION
Dry air effects on the human body a potential health hazard, skin irritations, eye itching and static electricity.
These problems can be reduced by increasing the indoor humidity by humidifiers, vaporizers, steam generators
and other simple methods such boiling water or wet towels. In the winter season the room is heated to keep the
temperature warm, and the ventilation is decreased, that is why the air in the room becomes dry and gives bad
effects to resident and furniture[Aldous et. al, 1996]]. Particularly the impacts of dry air on people are immunity
decrease on the disease due to the dry respiratory organs, snivel, soar throat and headache. Dry air also makes
problem skin oversensitiveness.
To humidify room plant or small artificial water fountain is used, and the most general way of humidifying room is
to use electrical device. The electrical device can add moisture in room effectively, while there found a side effect
that use of the sterilizing agent to prohibit the fungi and bacteria from growing inside the humidifier was
appeared to cause lung disease when breathing. This put enfant to death and became social problem.
The aim of this study is to seek for the alternatives to avoid this kind of problem, thus to develop humidifying pad
to be used in evaporating humidifier with eco-friendly performance and anti-bacteria performance. Contrary to
the general electrical humidifier, the evaporating humidifier can reduce electric energy, furthermore can heat
room while humidifying. To develop pad material, the tests on the anti-bacteria performance and resistance
performance to fungi were carried out to give good effect to the indoor air quality together with the performance
of absorbing and evaporating moisture.
The most popular humidifying pads are two categories depending on their behavior of flowing wind : aspen
excelsior and rigid cellulose media. The aspen pad doesn’t have uniform direction of wind flow, but draws the
wind to every direction. On the contrary the rigid media pad are made of special wettable cellulose in corrugated
sheets bonded together at opposing angles[Paschold H. et. al, 2003]. The angles of the corrugated cellulose are
intended to maximize air contact and evaporation[Watt J. R. et. al, 1997]. The main material used in this study is
hwangto, which is never been used for the evaporative pad in the existing research or industry. The shape of the
pad using hwangto is considered to be well fitted to make corrugated sheet, and the work to make the
corrugated sheet will be carried out using hwangto with hardening agents and function materials.

EVAPORATIVE HUMIDIFICATION
Evaporation is to transmit moisture into the air invisibly by using a fan to blow air through a moistened absorbent
material, such as a belt, wick, or filter[EPA, 1991]. When the fluid turns into vapor, it needs heat from outside.
The heat absorbed is called evaporative heat. In case of evaporative humidifier it uses the heat of warm air from
air conditioner, thus no extra energy is needed. The evaporative humidifier is a humidifier applying such
phenomenon that the heated air blows on the surface of wet material and the room becomes humid. The ways
of wetting a material are mainly known two types; a spraying method above the material through water pipe and
a capillary method that makes material absorb water from the water stored at the bottom.

DEVELOPMENT OF HUMIDIFYING PAD

The objective of this study is to develop a humidifying pad which is composed of various eco-friendly or natural
material such as Hwangto(loess), traditional paper, silica and charcoal etc. The existing pads used for
evaporative humidifier are made from silica coated ceramic mixed with glass wool or rubber type for a spraying

714
method, and fiber for a capillary method to absorb water well. Hwangto, the main raw material to develop
evaporating pad, is abundant in Korea and has been used in the traditional housing. Thus the research has done
to use hwangto as pozzolanic admixtures(Go S. S. et. al, 2009, 2010). So it is expected to provide environment
reservation if the development of evaporating pad is successful. Especially the pad is wet which is likely to be
contaminated by fungi or bacteria, it is very important to develop a pad free of contamination. In this point, mixed
rates between natural materials like pulp, hwangto, elvan or silica, including fiber material, were considered.
The development was proceeded in the point of getting strength of the specimen, then the attempt to make
standardized size was done. And finally the representative specimen was used for the evaluation of
environmental performance. Firstly to get enough strength of specimen it should act well with water. Whether it
absorbs water or not, the specimen should keep its strength. Hardening agent that consists of natural material
and vegetable adhesive material were applied for the test, and fiber and pulp were used for the reinforcement of
the strength to prevent leakage.
Table 1 shows the test for the strength of pad carried out in three steps. In the 1st step, the pulp law material of
paper was used to check if it has enough strength when mixed with hwangto. Because there is no specific
standard for the test of pad strength, it was observed by eye on the surface and on the resistance to water if it
nd
the specimen maintain the strength without any deformation or separation. The 2 test was carried out using
natural bondage made from see plant which is generally used for the adhesion of wall paper. In addition, the pulp
rd
and the fiber were used to reinforce the strength. In the 3 step, the strength of mixed material was examined.
The materials mixed were MGO powder and fluid, pulp, fiber with activated hwangto.

Table 1. Test for the strength of hwangto paste with adhesives

Test Main ingredients Process Result

Mixing hwangto with

st pulp
1 test Pulp Not good
Examine the strength
after hardening

Use adhesive from

sea plant
nd
2 test Starch from sea plant Mixing hwangto with Not good
pulp and fiber
Examine the strength
Use hardening
powder and fluid from
rd Powder and fluid from
3 test MGO Enough strength
MGO
Mixing hwangto with
pulp, fiber

As results from the tests during three times, the strength of the hwangto mixture with pulp was not good.
Practically the pulp needs some specific mould with high temperature to have enough strength. In this test the
mould was not applied, then it is considered not to have enough strength. The disadvantage is resulted that the
pulp can be applied only in order to improve the strength as a subsidiary material. When the adhesive from sea
nd
plant in the 2 test was used, the shape was good but there was separation in the materials. The strength was
judged to be good, but appeared to be contracted. In addition the specimen when absorbing water was
deformed so that it can’t be applied to humidifying pad. The result when powder and fluid extracted from
MGO(oxygenized magnesium) were mixed with hwangto was good that the specimen have enough strength.
Moreover the pulp and fiber were shown to give additional strength to the specimen. The specimen with thin
cloth was best, and there was no deformation of its shape when it absorbs water. The detail about the test and
result are represented in the Table 2.

715
Table 2. The result from the test to examine the strength of hwangto mixture depending on the mixing ratio.

Appearance Separation
Hardening
of material
1(bad)-
Mixing ratio Good- (none- Photo Evaluation.
2(fair)-
Crack normal-
fair-bad 3(good)
severe)

Hwangto:hardening
powder:hardening
fluid:fiber:water = Good None Good None Good
1:1:1: 0.01:0.6

Hwangto:Hardenign
powder:Hardening
Bad Crack Good None Bad
fluid:fiber:Water=
1:0.5:0.5:0.01:0.6

The test result showed that the hwangto mixture mixed with hardening powder, fluid from MGO and fiber at the
same time was best in strength and has good performance of both absorbing and discharging water. The mixing
ratio was when hwangto is mixed with hardening powder and fluid with same amount. Therefore the mixing ratio
which shown best result was decided to apply to the humidifying pad, then the size fit to the evaporating
humidifier and the shape were taken into consideration.
The concept of shape design was based on the type of evaporating humidifier. The humidifier considered to this
study is that the water is sprayed above the pad and discharged at bottom through small pipe. The types
considered were 1) vertical type, 2) horizontal type, 3) bending type and 4) sponge type. The horizontal type has
some problems in keeping water balanced that the water is supplied excessively and consumed. The bending
type was considered to be effective but the water flowing was not well balanced. The sponge type has limitation
to compose the mould. Eventually the vertical type was chosen to be applied to the humidifier by making
specimen. The specimen would be used for the test of eco-friendly characteristics and humidifying performance.

PERFORMANCE EVALUATION
The performance of specimen was evaluated on the eco-friendly characteristics and humidity control
performance. The test of eco-friendly performance is done for the anti-bacteria performance and resistance
performance to fungi to provide good effect on the indoor air quality. The humidity control performance is tested
on the performance of absorbing and evaporating moisture by testing the room humidity when operating
evaporating humidifier in practical. The test method on the anti-bacteria was in accordance with KS J 4206:2008
and on resistance on the performance in accordance with KS J 3201:1980(2006 reexamined). The humidity
control performance was tested and evaluated for the specimen manufactured with the corrugated form fit to be
equipped to the evaporative humidifier in two same sized test rooms.

Eco-friendly performance
To evaluate the eco-friendly performance, some materials known to be good in the air were used for the
mixing. The ingredients were hwangto including coal powder, elvan and silica. These materials were mixed
with hardening materials from MGO with fiber reinforcing strength which was approved to give enough
strength to the specimen through earlier test. Table 3 and 4 illustrate the mixing ratio of each material and the
specimen, respectively. For the test of anti-bacteria and resistance to fungi, the specimen no. 1 and no. 5
were used. The specimen no. 4 was excluded in the test because it took a relatively long time to dry and
harden by using adhesive from sea plant, and no. 2 and no. 3 were also not used in the test but examined if
the strength and shape were good because the ingredients used were included in the specimen no. 5 as well.

716
Table 3. Mixing ratio of the specimen for the test of eco-friendly performance.
Adhesive
Hardening agent
Charcoal Hardening
no Hwangto Elvan Silica agent (from Water Remarks
powder agent (fluid)
(powder) sea
plant)
1 300 300 300 180

2 300 15 300 300 180

3 300 15 15 15 300 300 180

4 300 15 15 15 300 300

5 300 15 15 15 300 300 180 PVA 5g

Table 4. Test specimen for the test of eco-friendly performance.

No.
1 2 3 4 5
Specimen

Anti-bacteria
test

Resistance to
fungi

The appearance after hardening, the specimen were good for their strength and its shape except no. 4 used
sea plant adhesive. The result of anti-bacteria test showed that the rate of bacteria reduction was above
99.9%, then there have not shown any of bacteria growth. The test result of the resistance to fungi showed
that there was no any fungi growth until 4 weeks were passed.

Humidifying performance
For the evaluation of humidity control performance, only specimen no. 1 was used among those specimen.
The shape is corrugated like in the fig. 1, which was manufactured using corrugated mould. The corrugated
mould was made from the punched iron plate flexible to be bent. The easy separation between mould and
specimen could be able to use oil paper between them. Finally the specimen made one by one were installed
to the frame of evaporating humidifier to measure indoor air temperature and humidity. The test rooms are
composed of two same rooms that one room, the reference room doesn’t have anything in it, and the other
room, test room, has been installed the humidifier. The two humidifiers were tested to compare the
performance of the specimen with the existing humidifier. The air temperature and the humidity were
measured by self writing thermohygrometer manufactured by SATO.

717
 Fig. 1. The shape of the corrugated specimen and installation to the case fit to evaporating humidifier.

A B

Reference room Test room

A B

Existing Specimen

Fig. 2. Test room for the test of humidity control performance.

The air temperature outdoor was 22 to 30 ℃, and the humidity 20 to 70 % fluctuated by time. The air
temperature in the reference room was 22 to 27 ℃, and the humidity 45 to 60 % less fluctuated than that in
outdoor. On the contrary while the air temperature in the test room was 22 to 25 ℃ similar to the reference
room, the humidity was approached to 78 % close to the humidity which was set up to 80 % which is quite
different with that in the reference room. It could be judged that the humidity control performance of the
specimen is quite good, because the evaporating humidifier installed specimen set the optimized indoor
humidity and keep the humidity targeted value. Fig. 3 is showing the trend of humidity in the test room how it
matches the targeted value well comparing with reference room and outdoor.

a. Outdoor b. Reference room c. Test room

Fig. 3. The humidity in the test room compared with outdoor and reference room.

718
CONCLUSIONS
According to the electrical humidifier has become a social problem due to the sterilizing agent to prevent the
device from contamination, the eco-friendly humidifier is necessary good for the indoor air quality. In addition, the
electrical devices more or less consume electrical energy to run it, then the device reducing the energy
consumption is another reason of developing humidifier. In this study, focusing on the evaporating humidifier, the
humidifying pad was developed and tested. Tests were carried out for the eco-friendly performance and humidity
control performance. The developed pad has shown good ability for the anti-bacteria and resistance to fungi,
which can be used as a good for the indoor air quality. The humidity control performance was evaluated good
that the humidity in the test room, where specimen was installed, meets the targeted value, even though the
humidity in the reference room was very low. This result tells that the evaporating humidifier installed specimen
developed can provide optimized room humidity together with good air quality, and be useful in the point of
energy saving.

Acknowledgement
This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea
government(MEST)(No. 2005-0049719, 2012006695) and supported by Business for Cooperative R&D
between Industry, Academy, and Research Institute funded Korea Small and Medium Business Administration in
2011.

REFERENCES
Aldous M. B., Holberg C. J., Wright A. L., Martinez F. D., Taussig L. M., 1996, Evaporative cooling and other home
factors and lower respiratory tract illness during the first year of life, American Journal of Epidemiology 143(5),
423-430.
EPA, 1991, Indoor Air Facts No.8 Use and Care of Home Humidifiers, Research and Development(MD-56)
Go S. S., Lee H. C., Lee J. Y., Kim J. H. and Chung C. W. 2009. Experimental investigation of mortars using
activated Hwangtoh, Construction and Building Materials 23: 1438-1445.
Go S. S., Chung C. W., Struble L. J. and Lee H. C. 2010. Pozzolanic activity of Hwangtoh clay, Construction and
Building Materials 24: 2638-2645.
Jin S. K., Jung S. Y., 2010, Production of hardened soil brick using liquid type or powder type stabilizing agent,
Geo-environmental and slope stability conference 2010, 195-203
Paschold H., Li W. W., Morales H., Walton J., 2003, Laboratory study of the impact of evaporative coolers on
indoor PM concentrations, Atmospheric Environment 37, 1075-1086.
KS J 4206:2008 Testing method for antibacterial activity of antibacterial functional products – Part 1 : Shake flask
method : Korean Standards Association.
KS J 3201:1980(2006) Methods of test for fungus resistance : Korean Standards Association.
Watt J. R., Brown W. K., 1997, Evaporative air conditioning Handbook, Third Edition, The Fairmont Press Inc.,
Lilburn, Georgia.

719
 THE DEVELOPMENT OF WALL SYSTEMS TO IMPROVE THE ENVIRONMENTAL
PERFORMANCE OF TRADITIONAL BUILDINGS
1 2 2 2 2 3 4
Hyeon Ku Park , Sun-Woo Kim , Bang Yeon Lee , Deuk Yum Cheon , Won Sik Kim , Eun Sun Baek , Sang Ku Yeo ,
5
Hang Kim
1. Sustainable Building Research Center, Hanyang University
55 Hanyangdaehak Ro, Sangnok Gu, Ansan Shi, Kyeonggi Do, 426-791, Korea
soundpark@cricmail.net
2. School of Architecture, Chonnam National University
300 Yongbong dong, Buk gu, Gwangju Shi, 500-757, Korea
3. Dept. of Fire Administration, Dongshin University
185 Geonjae ro, Naju shi, Chonnam, 520-714, Korea
4. The Prime Development Co., Ltd
885-7 (402) Donglim dong, Buk gu, Gwangju,500-804, Korea
5. Korea Marine Equipment Research Institute
24-20 Noksan sandan 335, Kangseo gu, Busan shi, 618-270, Korea

ABSTRACT
Korean traditional buildings are generally composed primarily of wood and soil. Gradually these have been replaced
by concrete due to advantages including strength and ease of construction. However, as “sick building syndrome” and
environmental concerns have come into additional focus, more research is being done on use and application
technologies. This study seeks a way to improve the typically weak environmental performance of traditional
buildings, with the goal of maintaining traditional appearances as well as eco-friendliness. The result showed that the
heat transfer coefficient of the designed wall system satisfied the requirement and the airborne sound insulation
performance was ranged between Rw+C 53 dB and 58 dB which is grade 2.

INTRODUCTION
Hanok, traditional Korean housing, is composed of wooden and soil materials, which have desirable aesthetic
qualities, but are limited in their residential convenience. It has been demonstrated (Park G. S. et. al, 2010) that
in Hanok houses, heat loss through walls is high compared with roofs and floors. According to a post-occupancy
evaluation of Hanok(Lee et. al, 2010), it was found that people feel negatively about the thermal and acoustic
environments of their homes, often outweighing aesthetic concerns. Hanok houses often cannot meet standards
concerning acoustic and temperature efficiency designed for modern-style houses. In addition construction
methods are labor intensive in traditional buildings because wood and soil materials are difficult to manufacture.
Ultimately these difficulties yield economic disadvantages compared to modern structures. Technical
improvement is required to increase the desirability of Hanok houses. Crucial residential factors include acoustic
performance, and thermal insulation performance, including airtightness. Other maintenance and management
factors include efficiency of weatherproofing and combustibility.
Important physical features concerning thermal and acoustic performance include the roof, walls, windows, and
floors. Among these, the walls and windows are most crucial. This study aims to improve the physical
performance of Hanok walls and windows in providing a warm and quiet residence. Windows are individually
developed regardless of other building elements, thus their profiles are separate from other Hanok construction
factors . Thus, walls are the main target of this study (Grimwood, 1997, Langdon et. al, 1981).
The main purpose of this study is to develop thermally efficient, acoustically resolute, traditionally appealing, and
eco-friendly walls in the Hanok style. Eco-friendliness is important not just because of the inherent value in
preserving the environment, but because surveys indicate that it is an increasingly important holistic
consideration for home-buyers (Jang M. S. et. al, 2011.8). The scope of this study is Hanok walls excluding
joints between columns and beams, which can already meet the requirements proposed based on residential
performance. A reasonable method of wall construction between wet and dry processes can be chosen
considering the constructability and economic efficiency.

FEATURES OF TRADITIONAL HOUSING WALLS AND DEVELOPMENT

The walls of Hanok are composed of a hearting structure in which the column and the beam are exposed on the
wall, and there is wall or window located between columns(photo 1). This is the main feature that should exist in
the façade. One of the main ingredients used for walls is soil that was constructed via wet process. It takes a
long time to build using soil-based materials, and these walls are vulnerable to water damage. Moreover soil
walls crack over time due to shrinkage, resulting in decreased thermal insulation performance. Therefore the dry
process is actively considered to complement this weak point, increasing airtightness by fabricating wall
elements.

720
 Photo 1. Traditional Hanok(Sunbichon, Sosu seowon, Kyongbuk)

Modern buildings are constructed with heavyweight concrete walls as well as fabricated lightweight walls using
wood and steel studs (Park H. K., 2010). The lightweight wall system could potentially apply well to Hanok. The
lightweight wall structure is generally fabricated with some building elements in the field. It could be classified
according to the type of studs used. Building elements could be applied in the same way. They can be classified
according to their building materials such as plaster board walls, MGO (magnesium oxide) board walls,
CRC(calcium silicate reinforced concrete) board walls, etc. Even though most generally plaster boards are
attached on both sides, both MGO and CRC boards are applied to the wall under considerations of strength,
environment, acoustic performance etc.
This study proposes that walls of this style be applied to traditional housing, selecting efficient elements of
modern-style walls. There are three objectives for these walls. The first is to that they should be eco-friendly. The
second is that the airborne sound insulation performance should be above Rw+C 48 dB, grade 3 in the green
building certification regulations. The third goal is that the heat transmission coefficient should be under 0.36
2
W/m K, thermal insulation criteria in central South Korea.

WALL SYSTEM DESIGN AND DEVELOPMENT

The wall system was designed to meet the thermal performance requirement. Once thermal compliance was
established, modifications were made to increase acoustic performance. The acoustic performance was
calculated via computer simulation, and the data was referred to roughly because the computer simulations can
only calculate simple structures. Finally the sound insulation performance of the designed wall system was
examined through laboratory testing. In addition, tests for resistance to bacteria and fungi were conducted.
The wall, especially in case of dry process wall, is not composed of only one material, but assembled of many
different elements. Thus the acoustic performance is changed if the means of assembly, such as order or
construction method, are changed. The elements of wall in this study are individually approached, and Fig. 1.
shows the assembly concept. The specific development scheme is divided into each part; stud, inner material
and surface material. Both wood and steel studs are considered. Inner materials are again divided into thermal
insulation and sound absorber, because they have different physical features and performance. Among these,
the thermal insulation is to be used with market products considering environmental performance, and the sound
absorber is to be developed using soil/foaming agent. Surface materials are further divided into interior wall and
façade wall. The façade wall material was developed considering strength, weather-proof performance and
water resistance, while interior wall materials were considered in terms of eco-friendly performance.

721
 Fig. 1. Concept of wall system assembling

As explained above, the façade wall element and sound absorber filling inside the wall were developed
according to the concept of wall system development. The façade wall material was made with high ductility
panel in consideration of strength and weather-proof performance. Then the panel was factory-assembled with
polyurethane thermal insulator and wood paneling. Table 1 shows the composite panel in question.
The sound absorber is soil sound absorber made of “hwangto”(yellow soil) and foaming agent. The foaming
agent used is of a vegetable-type, with openings to ensure cells fit together in a tight, thus soundproof way. Table
2 shows the sound absorber in detail.

Table 1. Composite panel with high ductility paneling, polyurethane and wood panels.
Item Detail
- Composite panel with poly urethane foam material
Concept of - Size : 805mm * 400mm (thickness : 67.5mm)
development - Method of construction : Fixed upon studs with screws
- Joint treatment : Filling with urethane foam

Plan and
section

722
 photo

Table 2. Sound absorber using hwangto(yellow soil)

Item Detail
- Use different kinds of foaming agents
Concept of
- Compared sound absorption by the internal foams
development
- Determining the size considering stud construction
Thickness : Within 5cm
Dimension
Size : 30cm(width) × 30cm(length)

Photo

Mold and curing Surface

Finally the wall system was designed, applying each element including those developed in labs or purchased in
markets. The table 3 expresses a wall system design proposal considering existing wall systems(Rudder et. al,
1985, ASTM, Quirt et. al, 1995); three exterior walls and one partition wall. W1 is the reference wall and W2/W3
were planned to compare the effects of air space and sound absorbers. The air space and sound absorbers
affect the thickness of walls and the area of the room. The reference wall, W1, consists of composite panel with
wood panel, air space, sound absorber and soil boards on steel stud. W2 was designed to see the effects of air
space and W3 to see analyze the absence of sound absorbers. W4 is the partition wall installed between rooms,
so it needs sound insulation performance above thermal insulation performance.

Table 3. Sound absorber using hwangto(yellow soil)

No. Composition Section Thickness Remark

Composite panel 70 + steel stud 50 +

wood panel 10 + vibration sheet 4 + Façade
W1 193.5 mm
hwangto absorber 50 + hwangto board wall
9.5

723
 Composite panel 70 + steel stud 50 +
Façade
W2 hwangto absorber 50 + hwangto board 179.5 mm
wall
9.5

Composite panel 70 + steel stud 50 + Façade

W3 129.5 mm
hwangto panel 9.5 wall

Hwnagto panel 9.5 + vibration sheet 4 +

Inner
W4 wood panel 10 + hwangto absorber 50 + 79 mm
wall
hwangto panel 9.5

EVALUATION OF THERMAL INSULATION AND SOUND INSULATION PERFORMANCE

The heat transmission coefficient was calculated to see if this wall system can satisfy the thermal insulation
performance requirement. The calculation for four wall systems was carried out with Eq.(1) and (2) using data in
Table 4,

K = 1 / RT (1)

RT = Ro + Σ(d/λ) + Ri (1)

Where,
K total heat transmission coefficient of wall
RT total heat resistance of wall
Ro outside surface heat transmission resistance
d thickness of element
λ heat conductivity of element
Ri inside surface heat transmission resistance

Table 4. The elements’ thermal characteristics calculations.

Heat conductivity(W/mK) or Remark
heat transmission
Element
resistance(㎡K/W) of
element
outside surface heat transmission 0.04 ㎡K/W
resistance
high ductility board 0.43 ㎡K/W plaster board
soundproof sheet 0.0336 W/mK soft rubber
plaster board 0.43 ㎡K/W
thermal insulator 0.034 W/mK glass wool 32K, 50mm
air space 0.086 ㎡K/W
soil porous absorber 1.32 W/mK lightweight porous concrete data
Hwangto(yellow soil) panel 0.114 W/mK
inside surface heat transmission resistance 0.13 ㎡K/W
2
W1 and W4 are satisfying the requirement of 0.36W/m K in heat transmission resistance, while W2, with no air
space, and W3, with no sound absorber, have not reached their applicable standards. Despite these liabilities, it
is not difficult to improve the heat transmission coefficient by increasing the thickness of the elements.

724
 Fig. 1. Comparison of heat transmission coefficient for the four wall systems.

The airborne sound insulation performance for the wall systems was evaluated by laboratory test. The test
was carried out in accordance with ISO 10140-2:2010 to see the difference by the variables. As a result, the
W1 has the highest value of Rw+C 53 dB, as anticipated, and the W4 has a slightly lower value of Rw+C 47
dB. The difference is caused by composite polyurethane panels and air space. The W2 doesn’t have air
space and wood panels with soundproof sheets, and the sound insulation performance was appeared to be
lower than the W1 by 3 dB. The W3 doesn’t have sound absorber and wood panels with soundproof sheet,
and is lower than W1 by 10 dB. These results express that the sound absorber in the wall is very important
and the airspace can provide some additional sound insulation. In the end, the W1 reached a “green building
certificate” value of grade 2, within an Rw+C of 53 to 58 dB. This could be applied to the traditional buildings if
another metric, such as environmental quality, can meet its applicable requirements.

Fig. 1. Comparison of airborne sound insulation performance for the wall systems.

CONCLUSIONS
As a method to improve Hanok’s weak residential performance, a new wall system was developed that
presents measurable benefits not only in residential performance but in physical performance. The
improvement targets were thermal and acoustic performance, and were tested to determine if they meet the

725
requirements already set up for the residential structures. The wall systems were assembled in various
conditions, and it was finally determined that composite polyurethane panels and hwangto(yellow soil) porous
materials make the best sound absorbers. For the composite paneling, high ductility panels were developed
to improve strength, airtightness and weatherproof performance for the exterior surface wall. In addition the
hwangto sound absorber was developed to increase sound insulation performance while possessing eco-
friendly characteristics. Two of those wall system designs satisfied the heat transmission requirement. The
airborne sound insulation performance was tested in a laboratory and was judged to be grade 2, ranging from
Rw+C 53 to 58 dB. Consequently the wall system designed in this study was determined to be applicable to
the Hanok as a substitute for existing traditional walls. It should be considered to have constructability in joint
areas with column and beams through a mock up test.

Acknowledgement
This research was supported by Basic Science Research Program through the National Research Foundation of
Korea(NRF) funded by the Ministry of Education, Science and Technology(2012006695, No. 2005-0049719) and a
grant(10High tech Urban B01) from High tech Urban Development Program funded by Ministry of Land, Transport and
Maritime Affairs of Korean government.

REFERENCES
ASTM E 497-99, Standard practice for installing sound-isolating lightweight partitions, ASTM international West
Conshohocken, PA, U.S.A.
Grimwood C., 1997, Complaints about Poor Sound Insulation between Dwellings in England and Wales, Applied
Acoustics 52(3/4): 211-223
Langdon F. J., Buller I. B. and Scholes W. E., 1981, Noise from Neighbours and the Sound Insulation of Party
Walls in Houses, J. Sound Vibr. 79(2): 205-228
Park G. S., Kim S. H. and Chae C. U. 2010. A Study on the Improving Scheme of Residential Performance
through the Assessment of Performance Han-Ok, Proc. Autumn Conf. of Architectural institute of Korea 30(1):193-
194.
Lee C. J. and Choi Il. 2010. Post Occupancy Evaluation of Hanok in Rural Area and Comparative Study on the
Perception of Hanok – Hanok Residents and Architectural Designers Comparison Based on-, J. of Architectural
institute of Korea 12(2):77-84.
Jang M. S. and Lee Y. S. 2011. Health Enhancing Architectural Features of Modern Hanok Perceived by
Apartment Residents, J. of Korean Institute of Ecological Architecture and Environment 11(4):3-10.
Park H. K. 2010. Analysis of Characteristics and User’s Evaluation for Lightweight Panel in Apartment, J. of the
Korea Institute of Building Construction 10(1):181-191.
Rudder F. F. Jr, 1985, Airborne sound transmission loss characteristics of woodframe construction. Gen. Tech.
Rep. FPL-43, Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory
ISO, Acoustics – Laboratory measurement of sound insulation of building elements – Part 2: Measurement of
airborne sound insulation, ISO 10140-2, Geneva: International Standards Organization; 2010.
Quirt J. D., Warnock A. C. C. and Birta J. A., 1995, Summary report for the consortium on gypsum board walls:
Sound transmission results, NRC Internal Report IRC-IR-693

726
 THE USE OF MAGNETIC PROPULSION SYSTEM IN PLACE OF INDUSTRIAL CONVEYORS

Hassaan Idrees
Electrical Engineering, Department of Electronics and Power Engineering,
Pakistan Navy Engineering College (PNEC), National University of Sciences and Technology (NUST),
Karachi-75350, Pakistan
hsidrees@gmail.com or h.idrees@ieee.org

ABSTRACT

In order to determine the feasibility of magnetically levitated and propelled linear systems for use in place of industrial
conveyor systems, a linear test track and vehicle was designed and constructed at National University of Sciences
and Technology, NUST, Karachi. The purpose of the facility was to demonstrate stable levitation, propulsion and
guidance of the maglev vehicle as it navigated the track. This paper reviews the analysis based on experimental work,
which was conducted in order to gauge the feasibility of using such a system in industry instead of employing
conveyor belt systems. The paper also includes a discussion of the hardware, software and instrumentation elements
of the controller and related systems.

Keywords–Magnetic levitation and propulsion; linear systems; stable levitation; linear test track

INTRODUCTION

Considering the fact that maglev trains are not currently practicable, especially for commercial applications in
extensive areas, ground-breaking projects like the E.M. train in Shanghai are already in use [1] and are leading the
way to revamp efficient transportation.

The motivation of this research to be used in the industry stems from this idea mentioned above. The experimental
project focused on magnetic levitation and propulsion systems, both areas which are not typically focused on, at least
not in small-scale workable experimentation set-ups. Many MAGLEV systems in such scenarios do not address
propulsion at all, or use mechanical systems to accelerate the train [2]. The goal of this project was to create a
platform for an effective, no-contact system for magnetic levitation and propulsion that has the capability to be
switched on and off for stopping and repeated movement.

MAGLEV systems use a number of methods to achieve the levitation and propulsion: by Electro-Magnetic
Suspension (EMS), electromagnets on the train produce repulsion to repel the train from the track [3][4], by Electro-
Dynamic Suspension (EDS), the electromagnets mounted on both train and track pushes the train away from the
track [5][6], in Hybrid Electromagnet Suspension (HEM), magnetic repulsion is obtained using permanent magnets [7],
and in direct diamagnetic levitation, permanent magnets are used on both sides, i.e. above the track and underside of
platform. Other methods use superconducting coils, exploiting the Meissner Effect [8], and liquid nitrogen on helium to
super-cool the magnets and achieve levitation superconducting magnets and coils. The basic purpose of the project
described in this paper was to develop a novel magnetically levitated and propelled system for position control
applications. Unlike passenger transportation systems, the linear motion system is comparatively smaller in size, with
a small carrier load and greater precision. The motion system is targeted for industrial automation machines by using
electromagnetic suspension in order to achieve the desired results.

METHODS AND IMPLEMENTATION

A. Working of Research

Firstly, the task was divided into two separate and solvable problems: levitation and propulsion. Using the flowchart
below, it was considerably easy to determine each step at a time. [9]

727
Fig.1: Flow-chart of the Maglev system

B. Experimental work on Levitation

An apparatus was set up in order to gauge the current which would vary with changing voltage in the electromagnet.
The hypothesis behind it was that it could help illustrate the effects of voltage with respect to current when the
electromagnet was activated for levitation. Here, the voltage was the controlled (independent) variable, while current
was the dependent variable. At a voltage of 23 V and current of 0.43 A, the electromagnet was attracted to the iron
sheet, hence, 0.42-0.43 A was the current range which used as benchmark. The table below shows the readings
obtained when the Hall Effect sensor A-1301 was subjected to changing voltage in the electromagnetic coil.

Table 1: Effect of changing voltage on current and force

Supplied Voltage Electromagnet current Force produced Proportionality
V (V) I (A) F (N) K
0 0.00 0.00 0.00
5 0.09 0.39 4.33
10 0.19 0.82 4.31
15 0.29 1.21 4.17
20 0.38 1.58 4.16
25 0.47 2.03 4.32

The charted results displayed a trend as shown below, where an increase in the current registered a proportional
increase in the attractive force produced by the electromagnet. This constant of proportionality, when averaged over
the range of values, was calculated as 4.258 signifying that this was the linear factor at which the ratio existed.

Fig.2: Graph for F(N)/I(A)

728
In order to construct a proper levitation system, a block diagram was developed as shown below.

Fig.3: Levitation block diagram

Mathematically, the normal force of levitation coil was determined by two methods: first, by using forces of gravity and
repulsion magnets, and second, by using Current-Force position equation [10]. Using the former method, the equation
for this force stands as follows:

Where
= normal force generated by levitation coil (variable)
= height of air gap (3.4 cm)
= mass of the cart (20 g)
= weight of moving platform (0.196 N)
= normal attraction force by the repulsion magnets

Using the Current-Force position equation, the equation for this force stands as follows:

Where
= normal force generated by levitation coil
= number of turns on levitation coil (200)
= cross-sectional area of the gap (2cm²)
= absolute permeability
= height of air-gap (2.5 cm)

C. Experimental work on Propulsion

Once, the levitation principle was accomplished, the propulsion system was developed using the block diagram
below.

Fig.4: Propulsion block diagram

This system circuit incorporated Hall Effect sensors A-1301, microcontroller, transistors C945, relays, diodes and pull-
up and pull-down resistors. The relays were connected to the electromagnets in order to switch the system for one
position to another. Here, we have three Hall Effect sensors instead of two, as in the levitation circuit previously.

729
Fig.5: Schematic of the propulsion circuit

After the propulsion was achieved, incorporation of both the levitation and propulsion systems was done in order to
fabricate an integrated workable structure.

Fig.6: Actual layout of propulsion circuit

Table 2: Key to the diagram above

Number Item name
1 Power Plug
2&3 Relays
4&5 Push Button
6 C828
7 Programmer Plug
8 Filtering capacitors
9 AT89S51 Microcontroller
10 Filtering capacitor
11&12 Transformers

730
D. System work-bench
The fabrication of the structure was done using the dimensions below.

Table 3: Dimensions of structure

Specifications Dimensions
Platform cart length 18 cm
Platform cart breadth 10.5 cm
Total track length 30 cm
Active track length 18.5 cm
Permanent magnet outer diameter 8.5 cm
Permanent magnet inner diameter 3.2 cm
Air gap between Permanent Magnets on track 1 cm
Air gap between electromagnets 3 cm

The diagram below was conceived as an initial structure on which the actual structure was made. [11]

Fig.7: Pictorial representation of the structure

The weights of the various components were done over a range and finally, an optimum was reached to balance both
levitation and propulsion using a guided mechanism.

The experimentally collected data showed Cart weight to be 20g, permanent magnet weight as 500g, and
electromagnet weight as 300g. The table below shows how the repulsion force correlated with increasing cart weight.

Table 4: Effect of increasing weight on repulsive loads

Experimental weight, (g) Total weight (exp. weight + cart weight), (g) Repulsion force, (N)
10 30 3.46
20 40 3.10
30 50 2.80
40 60 2.56
50 70 2.37
60 80 2.02

Fig.8: Experimental weights on cart

731
1. Track: The lower side of the track was made from wood so that a formidable support was provided. This material
was chosen after much deliberation and after both aluminum and acrylic were used and deemed unsuitable being
expensive and deformable, respectively. This was because aluminum though equally formidable was neither easily
available nor was it economical. An acrylic structure, on the other hand, was made but after the testing period the
acrylic proved weak and was unable to bear the weight of neither the electromagnets, nor the magnets.

Fig.9: Part of the track (left) and the lower side of the track (right)

Using the existing resources, the track built consisted of only two circular electromagnets extracted from clutch plates
of a scrap car. Even after very thorough scavenging, only four electromagnets from a local market were obtained, of
which two were burnt during the testing period. If available in larger quantities, a larger track would have been
constructed for better visualization and experimental methods.

2. Cart:
The cart was constructed of wood as well, because of the reasons mentioned above. For integration of propulsion
with levitation, two methods to achieve the desired solution were considered:

1. Energizing both cart (top-rail) and side-rail as active components: In this scenario, both the electromagnets
should repel each other but after giving a negligible push they attract each other.
2. Energizing cart (top-rail) as active and side-rail as passive: In this case, weak carbon magnets do repel but
they cannot produce the necessary lifting force. Moreover, strong magnets again attract to the electromagnet.

Eventually, an optimum solution was reached by using a passive cart and passive rails for a stable levitation [12]. This
is shown by the structure below:

Fig.10: Passive cart and passive rails

Effective and combined levitation and propulsion were achieved with the desired air gap and lateral distance, as
shown below.

732
Fig.11: The complete system with air gap

RESULTS AND DISCUSSION

The project, being a prototype, has a definite scope but is restricted by a weight limit of 85g, which was the maximum
weight which could be borne by the system. However, the positive aspect is that the end-product is appropriate to be
scaled so that it can be directly used in the industry.

Even though three MAGLEV designs have been developed to commercial readiness, MAGLEV systems in
continuous commercial operation are yet to be seen. Obstacles to commercial use of the MAGLEV systems include:

1. High expense, especially in guide-way construction,

2. Existence of conventional high-speed railways (French TGV),
3. Health concerns regarding exposure to electromagnetic fields,
4. Absence of commercially successful examples to reassure investors, and
5. Low possibility of selecting a guide-way design that will be compatible with future systems.

Parts of the video and animation were made public and can be viewed online. [13] [14]

AREAS OF CONSTRAINT AND IMPROVEMENT

Constraints

There were many issues which needed to be tackled during our project and they appeared in varying degrees of
difficulty and frequency.
1. Magnetic Retention: The Magnetic Cores available in the market have high Magnetic Retention. The hard
material which is available for the electromagnetic cores are not viable for our maglev purposes, whereas the
cores which should be used in our systems made from soft iron are not available in the market.

Fig.12: Magnetic retention of electromagnets

2. Neodymium Magnets: Propulsion can be achieved only with Neodymium magnets: Carbon magnets do not
produce enough horizontal forces to overcome static friction of the cart (top-rail).

733
Fig.13: Required Neodymium magnets

3. Strength and Shape of Magnets: Magnets of required strength and shape were unavailable in the market.
At least 180 (0.5”) square Neodymium magnets costing USD 16 each, which asked for approximately EUR
2,350.00 (including taxes), were found online. [15]

4. Circuitry: The design required four independent control circuits for each electromagnet which is needed to
balance the platform. This created extra circuitry and a cumbersome increase in weight causing the platform
to lose stability very quickly.

5. Mass of the platform: As the platform has a high mass compared to the weight of the electromagnets, this
created a non-uniform mass distribution (as compared to Ball levitation which involves only a small mass of
uniform shape)

6. Unstable platform: The coordination between the electromagnets for simultaneous stable levitation was
complicated since it posed a problem of balancing a very unstable platform which was apt to topple over
because of the uneven weights of the cart and especially of the electromagnets.

Risk areas

The risk analysis conducted for the experimental work posed the following questions:

1. Combating changing magnetic fields in the system,

2. Affixing power supply with microcontroller atop the cart,
3. Practical based on trial and error,
4. Endurance in electromagnets for the high current levels,
5. Structural accuracy of the required integrated system, and
6. Ruggedness of the sensors with respect to the system in an industrial environment.

Due to resource limitations, less pronounced magnetic propulsion was observed. This could be improved with
Neodymium (neodymium-iron-boron, Nd-Fe-B) magnets, because carbon magnets did not produce enough horizontal
forces, whereas the available Neodymium magnets were either too strong, too weak or not of required shape.

CONCLUSION

To improve on this design in the future efficiency should be optimized by adding a levitation system and using
superconducting electromagnets to prevent any energy loss due to friction or heat dissipation. What can also be done
is to improve the circuit using robust sensors and a PID controller to augment the structure and regulate energy
usage. The controller switches the electromagnet off once the requisite velocity is reached and switches the
electromagnet on only when it is required to maintain the velocity. MATLAB simulations during the trial phase can
prove a powerful tool for predicting solutions [17] [18]. Moreover, an oscilloscope, which is more accurate and
efficient, can be connected to the signal line for real time detection.

Most importantly, solving the economic factor can make this research more successful and productive. Unfortunately,
permanent magnets are very expensive, and building a linear propulsion system requires a lot of magnets. Even for a
36” track, over USD 50 are required on magnets alone. A levitation system can increase the costs further; for
fabrication and implementation of an actual MAGLEV train, the costs are proportionally higher. The design in this
project was not intended to minimize cost, however, for MAGLEV trains to become feasible, the technology would
need to become must more cost-effective. [19]

Conclusively, it appears that MAGLEV systems will not become popular without further funding and additional
advances in technology. The Japanese and German systems were developed at considerable government expense,
but still lacked a clear commercial advantage over conventional high-speed rail. The INDUCTRACK technology
showed a little more promise but required a long and expensive development period before it could be evaluated as a
candidate for commercial use. [20]

734
Recently, there is renewed hope for magnetic levitation transportation with the commitment of USD 1 billion for
MAGLEV research and development by the US government in their railway system [19]. MAGLEV trains prove to be a
promising technology in the future. The levitation and propulsion system using a linear motor built in this project
proved to be powerful. If a levitation system was to be built and superconducting electromagnets were used instead of
conventional electromagnets, the system would have been incredibly powerful and highly efficient.

ACKNOWLEDGEMENTS

The research reported herein was sponsored in part by grants of National University of Science and Technology and
the financial aid is gratefully acknowledged.

NOMENCLATURE

V, voltage or potential difference, V

I, current, A
F, force, N
A, area, cm²
H, height, cm
L, length, cm
W, weight, g

REFERENCES

[1] Shanghai Maglev Train, http://www.smtdc.com/en/ (2005)

[2] http://www.youtube.com/watch?v=6lmtbLu5nxw&feature=related (2009)
[3] Control strategies for Maglev electromagnetic suspension bogies, Paddison, J.E., Proceedings of the 35th
IEEE Conference on Decision and Control, 1996
[4] Robust controller design for Maglev transport vehicles with a guide-effective electromagnetic suspension
system, Morishita, M., Proceedings of the 35th IEEE Conference on Decision and Control
[5] Forces and moments for electrodynamic levitation systems--Large-scale test results and theory, Atherton, D.L.,
Mar 1978, IEEE Transactions on Magnetics
[6] HTSC Levitation Experiment With AC Current Modeling After EDS Maglev, Hungje Cho, June 2007, IEEE
Transactions on Applied Superconductivity
[7] A Nonlinear Method for Hybrid Electromagnetic Suspension, Junwei Cui, 2006, CES/IEEE 5th International
Power Electronics and Motion Control Conference
[8] In-Gann Chen, Journal of Applied Physics, Apr 1997, Magnetic levitation/suspension system by high-
temperature superconducting materials, Vol 81, Issue 8
[9] Y.-M. Che, S.-Y. Fan, W.-S. Lu, and T.-F. Wu, Electromagnetic Force Analysis of the Magnetically Levitated
Motor with Two Degrees of Freedom, Power Electronics Applied Research Laboratory, Department of Electrical
Engineering, National Chung Cheng University
[10] CASTELLI DEZZA, Francesco, DI GERLANDO, Antonino, FOGLIA, Gian Maria, Improved Magnetic Modelling
of EMS Maglev Levitators, Dipartimento di Elettrotecnica - Politecnico di Milano, Piazza Leonardo da Vinci 32,
20133 Milano, Italy
[11] Yong-Joo Kim, Pan-Seok Shin, Do-Hyun Kang and Yun-Hyun Cho, Design and Analysis of electromagnetic
system in a magnetically levitated vehicle, Korea Electro-technology Research Institute, Korea
[12] Koh Y.H. Eunice, Passive Magnetic Levitation, National University of Singapore, 21 Lower Kent Ridge Road,
Singapore 119077
[13] http://www.youtube.com/watch?v=9E6sBGCB6AA
[14] http://www.youtube.com/watch?v=zc5hDWz3csM
[15] http://www.rare-earth-magnets.com/c-7-cube-magnets.aspx
[16] Wang Ying, Passive Magnetic Levitation (Maglev) and Eddy Current, National University of Singapore,
Singapore 119077
[17] David W. Doll, Robert Kratz, Michael J. Newman, Allan B. Plunkett, Robert D. Blevins, Linear Synchronous
Motor Control for an Urban Maglev, General Atomics, 3550 General Atomics Court, San Diego, CA 92186, USA,
[18] Y.-M. Che, S.-Y. Fan*, W.-S. Lu*, and T.-F. Wu, Electromagnetic Force Analysis of the Magnetically Levitated
Motor with Two Degrees of Freedom, Power Electronics Applied Research Laboratory, Department of Electrical
Engineering, National Chung Cheng University, Ming-Hsiung, Chia-Yi, Taiwan, and *Linear Motor Research
Laboratory, Department of Electrical Engineering, Wufeng Institute of Technology, Ming-Hsiung, Chia-Yi, Taiwan
[19] Richard D. Thornton, Design Principles for Magnetic Levitation
[20] Arnie Heller, Maglev on the Development Track for Urban Transportation, Lawrence Livermore National
Laboratory

735
 FLOW AND HEAT TRANSFER CHARACTERISTICS OF AN EMPTY REFRIGERATED CONTAINER

E. Alptekin, M. A. Ezan, N. Kayansayan

Dokuz Eylül University, Faculty of Engineering, Mechanical Engineering Department
35397, Tinaztepe, Buca, Izmir

ABSTRACT
To preserve the quality, and the shelf life of perishable foods, it is essential to analyze the air flow distribution
inside a refrigerated truck for maintaining homogeneous temperature throughout. This study is an initiation of
such analysis, and air distribution patterns of ceiling-slot ventilated enclosure are numerically investigated for
three different geometries. Numerical analysis is carried out by using ANSYS-FLUENT software which is
based on control volume approach. In the analysis, the container cross-sectional area is kept constant.
However, by changing the container length as 6 m, 8 m, and 13.3 m, the aspect ratio of the container is
varied accordingly. The air velocity at the slot exit is varied in such a way that Reynolds number of the
5
discharge flow covers the range between 20 and 1.23x10 .

A three dimensional solution domain is considered in the analysis and is identified by a total number of
6
2.5x10 meshes. The Reynolds Stress Model (RSM) is applied in solving turbulent Navier-Stokes equations
at each computational node. The validity of the present computational method is checked by generating and
comparing the results with available work in literature. The flow distribution patterns of empty container
indicate that there exist circulation zones close to the surface at the opposite of air discharged slot. In fact,
two different vortices at two perpendicular planes coexist. The location of these circulation zones does not
change with the discharge flow rate, but the circulation intensity increases as the flow rate increases. It is
determined that variation of the container aspect ratio affects the flow behavior. For the case of container
length at 13.3 meters, the flow at the upper surface separates at a distance of 11.5 m from the air injection
plane.

1. INTRODUCTION
Cold chain is generally used to define a combination of temperature-controlled supply systems in food
industry. As illustrated in Fig. 1, from the cold storage (or warehouse) to the client, the food must be kept in a
predefined temperature range to preserve its quality and increase the shelf life. According to Flick et al.
(2012) about 60% of food products require refrigeration process in the cold chain. They also stated that
inadequate storage or transport conditions cause corruption of 25-30% of global food production. This huge
amount of lost can be regained by designing each component of the cold chain properly.

Figure 1 Cold chain

As indicated by Hoang et al. (2012), one hundred and twenty million tons of chilled foods are transported
each year in Europe. With the expansion of the urban areas, the temperature controlled of frozen foods in
refrigerated vehicles becomes crucial. Hoang et al. also mentioned that the ATP (International Agreement for
the Transport of Perishables) separates the refrigerated vehicles into two categories, which are ”Normally
Insulated” and “Heavily Insulated” vehicles. Heavily insulated trucks are adequate for frozen food
2
transportation with an overall heat transfer coefficients less than 0.4 W/m K, and normally insulated trucks
are considered to be proper for chilled foods (U ≤ 0.7 W/m K). In these trucks, the main issue is to control
2
3
the temperature distribution in a narrow range. To accomplish this in large volume (V > 50 m ) transportation,
the cooling system is usually driven by an engine separated from the vehicle’s main engine. The positioning
of the evaporator is a critical issue in acquiring homogeneous temperature distribution inside the container.
Neglecting the small influences of radiation and natural convection, the cooling process is mainly controlled

736
by forced convection, and the homogenous temperature distribution may be achieved by optimization of the
flow distribution inside the container.

Even though numerous experimental and numerical results have been published for thermal behavior of
warehouses, there are few studies conducted for refrigerated vehicles. As a pioneer, Moureh and his team
numerically and experimentally examined the flow ad thermal behavior of a refrigerated container under
several conditions. Moureh et al. (2002) used FLUENT CFD software to obtain velocity distribution in empty
and loaded containers. They compared the predicted velocity field with experimental LDA (Laser Doppler
Anemometer) measurements. Two different turbulence models are implemented, k-ε and RSM. As a result,
RSM turbulence model showed good agreement with the experimental data. After this pioneer study, Moureh
and his team conducted many other researches related with the refrigerated containers. Moureh and Flick
(2003) represented dimensionless temperature distribution inside container for one selected case at a single
Reynolds number and a container length. In addition, Moureh and Flick (2004) experimentally and
numerically studied the temperature and velocity distributions inside a loaded container for two different
spacing between loads, and the spacing is taken to be at 1 cm and 2 cm. Later, Moureh and Flick (2005)
tried to improve the performance of the container ventilation by locating the inlet section near lateral surface.
They compared the ventilation efficiency for lateral and central inlet orientations. Besides, Moureh et al.
(2009) tried to improve the cooling effect by changing the air duct configuration. Unlike the previous
numerical models, in this study, Moureh and his colleagues considered loads as porous medium to achieve
more realistic velocity field.

In this study, air is discharged from the evaporator located at mid-section of front surface having half width of
the container. Suction, on the other hand, takes places at all surfaces except the discharge plane of the
evaporator. First, the flow field is studied and the circulation zones are identified. To do this, flow Reynolds
number and container aspect ratio have varied, and the local and the average heat transfer characteristics of
such a flow are computed. The refrigerated truck is assumed to be moving for external flow simulations.
Thus, the container overall heat transfer coefficients and heat gains through the container surfaces are
realistically evaluated.

2. MATERIAL AND METHOD

2.1. Definition of the Problem

The geometry and the size of the refrigerated container are provided in Fig. 2. The evaporator is located at
the center of the front surface and the spacing between the ceiling and the evaporator is 4 cm. Air is supplied
to the computational domain from the inlet section as represented in Fig. 2, and collected by the openings at
the bottom, lateral and top surfaces of the evaporator. The container cross-sectional area is kept constant,
but the length is varied, so that the aspect ratio (L/H) is in the range between 2.4 and 5.32.

W = 2.5 m

Inlet Section
Outlet Section

H = 2.5 m

L = 6 m, 8 m and 13.3 m

Figure 2 Geometry of the container and evaporator

2.2. Governing Equations and Solution Method

To simplify the problem, instead of solving time dependent governing equations, time averaged Reynolds-
Averaged-Navier-Stokes (RANS) equations are solved. Considering steady and incompressible flow

737
conditions, the principle of conservation of mass, momentum and energy can be satisfied by the following
equations,
Mass:
∂
( ui ) = 0 (1)
∂xi
Momentum:
∂ ∂p ∂   ∂ui ∂u j  
∂x j
(
ui u j =
− )
+ µ  +
∂xi ∂x j   ∂x j ∂xi
 − ρ ui′u ′j 

(2)
  
Energy:
∂ ∂  c p µt  ∂T 
( ρ uiT ) =  k +  − ρ u ′jT ′ (3)
∂xi ∂x j  Prt  ∂x j 

where, u and u' indicate the average and fluctuating velocities, respectively. Similarly, T and T' are the mean
and instantaneous temperatures. Turbulence modeling by using RSM (Reynolds Stress Model) has been in
use by many authors. Norton et al. (2007) discussed the pros and cons of eddy viscosity and Reynolds
stress closure models in turbulence modeling for confined rooms with adverse pressure gradients and flow
separations. It is determined that RSM provided reasonable predictions for flow separations and Coanda
effects. Similarly, Schälin and Nielsen (2004), and Moureh and Flick (2005) indicated that for three
dimensional flow in confined spaces RSM method of calculations predict the experimental results with a
maximum of 10% deviation. As a result, Reynolds Stress Model (RSM) is considered in this study for
modeling the 3D turbulent flow.

Eqs. (1) to (3) are solved with the aid of ANSYS-FLUENT software. Software is based on the control volume
approach of Patankar (Patankar, 1980). QUICK scheme (Leonard, 1979) is applied for discretization of
convective terms and SIMPLE (Patankar and Spalding, 1972) algorithm is adapted for solution of momentum
and continuity equations. For evaluating the turbulence boundary layer near the wall, “log-law” function is
used (Launder and Spalding, 1974). Computational domain is divided into non-uniform quadrilateral grids. To
capture the temperature, velocity and pressure gradients close to the solid surface, fine grids are generated.
-5
Convergence criteria for all conservation equations are taken to be 10 .

2.3. Boundary Conditions and Mesh Independency

Because of container symmetry, half of the computational domain is taken into account, and at the symmetry
plane, the boundary condition is ∂φ ∂x x =W /2 =
0 . Uniform velocity (win) and temperature (Tin) profiles are
assumed at the inlet section of the evaporator. On the container walls no slip condition is sufficed by u = v =
w = 0. Due to large insulation thickness at the bottom and front (i.e. surface wıth evaporator) surfaces, these
surfaces are assumed to be adiabatic ( ∂T ∂n =0 ). The insulation material is selected to be
surface

polyurethane, with k = 0.02 W/mK, and the thickness of the insulation layers are 100 cm, 85 cm and 80 cm
for the ceiling, lateral and rear surfaces, respectively. On the top, lateral and rear surfaces, the combine
effect of solar radiation, convection and surface emission is considered and the following energy balance
equation is considered at those surfaces,

∂T
k = α I solar + σε (T 4 − Tsky
4
) + h (T − Tatm ) (4)
∂n surface

where α and Isolar are the surface absorptivity and the incident solar radiation, respectively. Meteorological
th
data of Izmir, Turkey for June 21 at 1 pm noon are considered for solar radiation intensity (Isolar), the sky
(Tsky) and atmospheric (Tatm) temperatures.

Average heat transfer coefficient (h) for each surface is evaluated by considering the results of a preliminary
analysis. To solve 3D flow over the truck and container combination, at a velocity of 90 km/h, a total of
sixteen million non-uniform grids are generated. The flow patterns around the truck are represented in Figure
3. At the top and back sides of the container, separation of flow can easily be located. As a result of this
preliminary analysis, average heat transfer coefficients for three surfaces are calculated as given in Table 1.

738
The predicted heat transfer coefficients are compared with the well-known correlation over a flat plate
(Incropera et al. 2009). Comparative results show that the numerical heat transfer coefficients are
approximately 30% higher than that calculated by theoretical methods.

Figure 3 Flow patterns around the refrigerated truck

Table 1 Mean heat transfer coefficients for external surfaces of container

2
Mean Heat Transfer Coefficient, h (W/m K)
Surface Theoretical
Predicted
Correlation Result
Top 64.48 50.73
Nu L = 0.037ReL4 5Pr 1 3
Lateral 66.18 56.4
Rear 42.05 − −

To get numerical results free of mesh size, the same geometry at the same flow conditions is studied for
three different mesh sizing which are respectively 2.5 million, 4.5 million and 10 million. As indicated in Fig.
4, no remarkable change of results is detected due to change in mesh size. Therefore, 2.5 million mesh is
selected for the current study.

3. RESULTS AND DISCUSSION

In this section, first the validity of the numerical method is discussed with regenerating the numerical data for
the same geometry and flow conditions as studied experimentally by Moureh and Flick (2005). After that, the
influence of selected parameters on the heat transfer inside the container is evaluated.

3.1. Validation of the Methodology

Recent experimental work of Moureh and Flick (2005) that provides flow patterns inside a refrigerated
vehicle is selected for comparison of present methodology. The container studied by Moureh has the same
cross-sectional dimensions as H = W = 2.5 m with the present study, and the container length is taken to be
13.3 m. Unlike the present study geometry, the air inlet section is at the top and the outlet is at the bottom of
the front plane. Moureh and Flick studied the airflow characteristics of such a container both experimentally
and numerically. In their analysis, LDA is used to capture the velocity distributions on the symmetry surface.
In addition, the inlet velocity is taken to be 11 m/s and the turbulence intensity of entering flow is defined to
be 10%. Moureh’s results and the present study are compared for the same flow geometry and flow
conditions.

In Fig. 4, airflow patterns on the symmetry plane of the container are compared with the experimental and
numerical results of the reference work. On the symmetry plane of the container, two separated circulation
zones are formed. Owing to the viscous effects, flow separates from the upper surface and generates a
secondary circulation at the rear region of the container. As illustrated Fig. 4, this study results predict the
experimental results of the reference work fairly accurately. In Fig. 5, decay of the jet axis velocity in
longitudinal direction is given and represents Moureh’s experimental results within acceptable engineering
accuracy.

739
Figure 3 Influence of number of grids on the jet velocity decay

(a) Experimental results (Moureh and Flick, 2005)

(b) Numerical results (Moureh and Flick, 2005)

(c) Current predictions

Figure 4 Comparison of airflow patterns with the reference work

740
Figure 5 Comparison of velocity variation on jet axis with the reference work

3.2. Parametric Results

In order to evaluate the inlet velocity effect on heat transfer through container surfaces, the flow Reynolds
5
number is varied in the range of 20 to 1.23x10 . In Reynolds number definition, characteristic length is taken
to be hydraulic diameter of the evaporator exit section. In addition, the container length is varied so that the
container aspect ratio (H/L) assumes the following values 2.4, 3.2 and 5.32. Hence, as shown in Table 2, a
total of eight parameters are varied in the analysis.

Table 2 Parameters that are used in the current study

Case Reynolds Number Inlet Temperature Length of the Container

#1 20
#2 200
#3 2000
4
8m
#4 1.23x10
4
-10°C
#5 6.23x10
5
#6 1.23x10

#7 5
13.3 m
1.23x10
#8 6m

3.2.1. Effect of Reynolds Number

Flow patterns and isotherms are illustrated in Fig. 6 for various Reynolds numbers. Temperature contours
are represented in non-dimensional form as, T * = (T − Tin ) (Tout − Tin ) . As the Re number increases from 20
5
to 1.23x10 intensity of circulation increases and dominates the rear side region of the container. In Fig. 6,
one can clearly see the Coanda effect for Re > 20. Coanda effect is defined as the tendency of the fluid jet
which is attracted to a nearby surface. After leaving the inlet section, because of Coanda effect, flow
becomes attach to the ceiling surface, however depending on the Reynolds number, flow separates from the
surface after travelling a certain distance. At low Re numbers (Re ≤ 2000), flow separates from the ceiling
and creates secondary circulation zone at the rear side of the enclosure.

741
 (a) Re = 20

(b) Re = 200

(c) Re = 2000

4
(d) Re = 1.23x10

4
(e) Re = 6.15x10

5
(f) Re = 1.23x10
Figure 6 Airflow patterns and isotherms on the symmetry surface

The existence of reverse flow for various Reynolds numbers are revealed with representing the velocity
distribution along the container length calculated at the jet axis (see Fig. 7). For Re = 20, almost 75% of the
enclosure is stagnant. However, for Re = 200 and 2000, flow separation from the ceiling is observed at
container lengths approximately z/L = 0.9 and 0.95, respectively. At higher Reynolds numbers, Re ≥ 12,300,
there is no separation of flow and a single circulation zone forms. Increasing Reynolds number slightly
changes the velocity variation especially at Reynolds numbers above 61,500.

742
Figure 7 Decay of velocity on jet-axis

Referring to Table 3, increasing the Reynolds number decreases the mean air temperature of the container.
In fact, average temperature decreases from 25.47°C to -9.68°C by increasing the Reynolds number as
indicated in the table. As shown in Fig. 6, container temperature distributions become more uniform at high
Reynolds numbers. As seen by Fig. 8, average values of inner surface heat transfer coefficients assume
high values as the Reynolds number increases. Due to decrease in surface thermal resistance, the surface
heat gains assume asymptotic values accordingly.

Table 3 Average air temperatures at various inlet conditions

4 4 5
Reynolds number 20 200 2000 1.23x10 6.15x10 1.23x10

Tmean 25.47°C 16.74°C 1.07°C -7.26°C -9.38°C -9.68°C

(a) Heat Transfer Coefficients (b) Heat Gains

Figure 8 Influence of flow Reynolds number on heat transfer

743
3.2.2. Effect of the container length
Fig. 10 shows the flow patterns and isotherms for three different lengths of the container. For the suction
design considered in this study, a stagnant region takes place just below the evaporator suction line. As
indicated in Fig. 10, increasing the container length increases the size of the stagnant region, this in turn,
deteriorates the uniformity of air temperature distribution of the container. Specifically for container length at
13.3 m, due to separation of flow from the ceiling surface at a length ratio z/L = 0.88 (see Fig. 11), the
temperature distribution even becomes worse. Table 4 presents the average air temperatures at three
different container lengths. One may conclude that the length effect on average temperature is insignificant.
Besides, as in Fig. 12, the heat gain per unit surface area varies insignificantly at all surfaces.

(a) L = 6 m

(b) L = 8 m

(c) L = 13,3 m
Figure 10 Airflow patterns and isotherms on the symmetry surface

Figure 11 Decay of velocity on jet axis

744
Table 4 Average air temperatures at various container lengths
Container Length 6m 8m 13.3 m

Tmean -9.72°C -9.68°C -9.47°C

Figure 12 Influence of container length on heat transfer

4. CONCLUSIONS
Flow and heat transfer characteristics of an empty container used for frigorific transportation is
numerically studied. The container has constant cross-sectional dimensions of 2.5 m x 2.5 m and the length
is varied in such a way that the container aspect ratio takes the values of 2.4, 3.2 and 5.32. Identifying the
turbulent stresses, Reynolds Stress Model is adopted. Similar studies of literature are also recalculated and
compared by this method of solution. Satisfactory results indicate that RSM method of analysis is
appropriate for studying the flow inside of refrigerated containers.

In the analysis, the flow Reynolds number is changed by varying the flow injection rate at the evaporator
outlet. For aspect ratios between 2.4 and 3.2, no separation of flow from the upper surface is noted.
However, in a region close to the rear surface of the container, two separate vorticities perpendicular to each
other occur at flow Reynolds numbers greater than 200. Occurrence of these vorticities indicates that a
stagnant region develops at the rear side of the container. For aspect ratio 5.32, the flow separates from the
upper surface of the container at a length ratio z/L = 0.88. Similar to smaller aspect ratios, two distinct
circulation zones occurs but their plane of occurrence is different from the cases with H/L = 2.4 and 3.2.

Non-dimensional temperature, T* should vary in the range between 0 and 1. If it is greater than the unity,
at a particular location, then that location would not be affected heat transfer wise by the air flow in the
container. In the analysis, as the Reynolds number increases, T* assumes less than 0.1 for the entire
container region. This also indicates that at high Reynolds numbers temperature uniformity throughout the
container is satisfactorily obtained. In addition, as the Reynolds number increases, the volumetric mean
temperature of air in the container drastically drops and approaches to the exit temperature at the
evaporator.

Current study investigates the essential features of the flow and describes the heat transfer
characteristics of an empty refrigerated container. Further studies should be conducted to evaluate the effect
of loading conditions on the flow and on the heat transfer characteristics in a frigorific container.

Acknowledgement
This research is supported by Republic of Turkey Ministry of Science, Industry and Technology under the
grant number: 01114.STZ.2011-2 and Dokuz Eylül University Scientific Research Foundation Grant BAP-
Project number: 2013.KB.FEN.9.

745
NOMENCLUTURE
cp specific heat, J/kgK
DH hydraulic diameter of the inlet section, m
2
h convective heat transfer coefficient, W/m K
H, L, W height, length and width of the container, m
2
Isolar incident solar radiation, W/m
k thermal conductivity, W/mK
Nu Nusselt number, hL/k
p pressure, Pa
Pr Prandtl number
Re Reynolds number, ρDHwin/µ
T temperature, °C
T* dimensionless temperature, T * = (T − Tin ) (Tout − Tin )
u, v, w velocity components on Cartesian geometry, m/s
2
U overall heat transfer coefficient, W/m K
3
V volume of the container, m
x, y, z Cartesian coordinates, m

Greek Letters
α surface absorptivity
ε surface emissivity
2
µ viscosity, Ns/m
ρ
3
density, kg/m

REFERENCES

ANSYS Inc, ANSYS-FLUENT 14.0 Theory Guide, 2009.

Flick D., H.M. Hoang, G. Alvarez, and O. Laguerre. 2012. Combined deterministic and stochastic
approaches for modeling the evolution of food products along the cold chain Part I: Methodology.
International Journal of Refrigeration 35:907-914.

Hoang M.H., O. Laguerre, J. Moureh, and D. Flick. 2012. Heat transfer modelling in a ventilated cavity
loaded with food product: Application to a refrigerated vehicle. Journal of Food Engineering 113:389–398.

Incropera F.P., D.P. DeWitt, T.L. Bergman and A.S. Lavine. 2007. Fundamentals of Heat and Mass Transfer.
John Wiley & Sons.

Launder B.E. and D.B. Spalding. 1974. Numerical computation of turbulent flows. Computer Methods in
Applied Mechanics and Engineering 3:269–289.

Leonard B.P. 1979. A stable and accurate convective modelling procedure based on quadratic upstream
interpolation. Computer Methods in Applied Mechanics and Engineering 19:59–98.

Moureh J., N. Menia, and D. Flick. 2002. Numerical and experimental study of airflow in a typical refrigerated
truck configuration loaded with pallets. Computers and Electronics in Agriculture 34:25–42.

Moureh J. and D. Flick. 2003. Wall air–jet characteristics and airflow patterns within a slot ventilated
enclosure. International Journal of Thermal Sciences 42:703-711.

Moureh J. and D. Flick. 2004. Airflow pattern and temperature distribution in a typical refrigerated truck
configuration loaded with pallets. International Journal of Refrigeration 27:464–474.

Moureh J. and D. Flick. 2005. Airflow characteristics within a slot-ventilated enclosure. International Journal
of Heat and Fluid Flow 26:12–24.

Moureh J., S. Tapsoba, E. Derens and D. Flick. 2009. Air velocity characteristics within vented pallets loaded
in a refrigerated vehicle with and without air ducts. International Journal of Refrigeration 32:220–234.

746
Norton T., D.W., Sun, J. Grant, R. Fallon and V. Dodd. 2007. Applications of computational fluid dynamics
(CFD) in the modelling and design of ventilation systems in the agricultural industry: A review. Bioresource
Technology 98:2386–2414.

Patankar S.V. 1982. Numerical Heat Transfer and Fluid Flow. New York: Taylor & Francis.

Patankar S.V. and D.V. Spalding. 1972. A calculation procedure for heat, mass and momentum transfer in
three-dimensional parabolic flows. International Journal for Heat and Mass Transfer 15:1787–1806.

Schälin A. and P.V. Nielsen. 2004. Impact of turbulence anisotropy near walls in room airflow. Indoor Air
14:159–168.

747
 CERTIFICATION AND MAINTENANCE OF INDUSTRIAL INSTALLATIONS IN THE
COMPANY TREFILEST ACCORDING TO STANDARD ISO 9001 /2000
1 2 3
DJEMAI-MOUNIRA ., HADJADJ-AOUL ELIAS, HAMAIDI BRAHIM
1
Genius Electromecanical Laboratory
2
Electromechanical Laboratory Of Genius Of Science Of The Engineer
3
Genius Electromecanical Laboratory
University Badji Mokhtar, BP 12 El Hadjar, Annaba
1
Email: MOUNHANACHI@yahoo.fr
2
Hadjadj.aoul.elias@yahoo.fr
3
Ham @yahoo.fr

ABSTRACT:
Open borders Algeria to the products and services imposed by the market economy, that the field of
competition between firms will no longer be confined to the region or country, but expect the world.
Sell a product to market requires a local Algerian company the same effort
its flow into China and Latin America.
The first major problem encountered in industrial production was the product meets customer needs,
therefore, the largest market share undoubtedly revert to that which will best meet customer
requirements. But the client now requires guarantees on the amplitude of the supplier to master its
production to ensure him a quality product that leads the company TREFILEST to set up a strict
quality system. This system allows the company to meet requirements, satisfy the customer and give
him confidence, it requires the company to be certified.
In our work, we studied the maintenance and certification of the company according to International
Standard ISO 9OO1: 2000 under a system of quality management. This work is to solve some
problems encountered during the manufacturing process to produce better and to satisfy the
customer.

KEY WORDS : Quality, standard, procedure, compliance, management, processes, ...

1. INTRODUCTION
All company and a set of services organized so that they work in a cooperative process to meet
customer satisfaction increasingly demanding, especially that which relates to the quality of the
product or service.
To establish or recognize the quality u service or business, it is necessary to have a system of
quality management
Certification of a company is the goal of any management whose main objective the satisfaction of
its customers, reducing production costs and improving profit margins.
The scope of our study concerns the maintenance department of the company production of long

748
 products (TREFILEST) has already begun the process of certification for all services since the
year 2OO2 obtained by SGS ISO 9001 2000.

2. IMPLEMENT MAINTENANCE
Maintenance and one of the cornerstones on which is built the history of industrial production. It
reveals
a lot of adaptability to mass production.

3. STAUS AND PROSPECTED

Embark resolutely on the path to effective maintenance involves firstly what type of maintenance
priority and secondly to assign precisely defined objectives.
Select a type of service over another can take on intuition and is based on a rational analysis,
whose prior is a self that will answer the following questions:
• What has been the evolution of maintenance or what has held in the company?
• What is the current situation?
• What are the broad guidelines of the maintenance policy in the future given the objectives
assigned to it?
This makes self-diagnosis officials loads of maintenance policy which must perform a system
scan productive in terms of cost, time quality e. Each piece of equipment constituting the
entire system can be multiplied by a coefficient expressing its potential influence on each of
these criteria.

4. THE QUALITY MANAGEMENT

The objective of quality is to satisfy stated and implied needs.
The evolution of quality concepts and their application in the growing industrial and administrative
sectors have led him to make quality management a recognized tool of management.
Processing and use of information quality in real-time: test results, identification of product defects
or failures of the organization, quality supplies, etc. .. , Provide corrective actions for each finding of
fault. So we gradually increased the simple concept of quality control of the product quality
assurance for products or services, and finally to quality management, which is part of the body u
dedicated to quality in mind the ISO 9000 series

Fig1. The eight principles of quality

management

749
 The quality procedures are traditionally grouped into a typology of five nival:

1. The quality control;

2. The control of the quality;
3. The quality assurance;
4. The quality management;
5. Total Quality (TMQ).
in the early '40s, companies were managed by a quantitative logic:

By cons in JAPAN, a qualitative logic was introduced by Dr. Deming, commissioned by the U.S. to
assist the Japanese in the recovery of their industry completely destroyed by the war:

750
5.ANALYSIS OF THE MAINTENANCE FUNCTION BEFORE AND AFTER CERTIFICATION
It will be to ensure that the organization and procedures are consistent with Article 6-3
"infrastructure" of ISO 9001 version 2000.
All charts of the company underwent a change TREFILEST approach to improvement, because they
were poorly trained in all services due to poor distribution of the different agents in different positions.
As the standard requires new posts and new names of some structures of the company.

6. ANALYSIS QUESTIONNAIRES MAINTENANCE FUCTION AFTER CERTIFICATION

For maintenance obtain good results for what is effective, it is based on what management.
Areas of maintenance management to define no standard rules for well maintained effectively. So we
propose a method of analysis decomposes maintenance management globally in twelve areas
identified from A to L:
A-The general organization
B-The working methods
C-Technical monitoring facilities
D-portfolio management work
E-management of spare parts
F-purchases parts and materials
G-organization of the maintenance workshop
H-Tools and measuring devices
I-Technical documentation
J-Staff and training
K-Outsourcing
L-control activity

751
A- General Organization:
It covers the general procedures for service organization, the rules by which the organization
is established and components of service policy.
B- Working methods:
We are electroplating the preparation work with, in particular, time estimates, methods of
interventions focusing on the methodologies used during troubleshooting ...
C- Technical monitoring facilities:
technical monitoring equipment includes all analysis petite to dose correctly according to the
objectives of availability and cost of remedial and corrective guidelines installations made
interventions actions, it is mainly to process information on specifications of facilities and
equipment modifications.
D- Portfolio management work:
This section covers the processing of applications for work and maintenance plans.
E- Management of spare parts:
Although the time is the elimination of stocks, minimum security still exist: how the parts are
they stored? ...
F- Procurement of parts and material:
It does not consider the organization buying the company, but whether the procedures used to
supply in good conditions, with the most appropriate suppliers.
G- The organization of the maintenance workshop:
Many spots are built in the workshop: it must provide well-equipped stations, conditions and
pleasant workspace.
H- Tooling and measuring equipment:
Maintenance materials need to be better tooling and should have many ways to handling and
it requires a serious organization and management.
İ- Technical documentation:
A good comprehensive documentation with easy access impeccable ranking and enjoying a
systematic updating the maintainer resists.
J- Personnel and training:
The most sensitive area even if the actual interning based management staff we went to the
mobilization of men.
K- Outsourcing:
The most difficult change that occurs at the moment is this shift toward outsourcing
maintenance.

L- The control of the activity:

What means are there to govern in the best maintenance?
Dashboard, information systems, activity reports?

752
 The analysis could also be three related areas: health and safety, energy conservation and
respects the environment

7.THE QUESTIONNAIRE ANALYSIS

We will present some analysis questionnaire maintenance function to some areas for continuous
improvement of the quality management system.
It covers twelve sections described above and has 105 questions, each question has 2-5 response
options:
• "yes" or "no"
• "rather yes" or "rather no", if not totally positive or totally negative
• "yes or no" if one of the options above are not suitable for use, simply check or circle the number in
the column closest to the assessment for the issue asked a subtotal is calculated for each of the
twelve categories.
From these questionnaires that have been made in the services business TREFILEST, we obtained
the following results:

753
Table 1.The questionnaire of A-General organization

754
Table 2.The questionnaire of B-working methods

755
Table 3.The questionnaire of C-Technical monitoring facilities

756
Table 4. . Analysis questionnaire maintenance function before certification

scores
Areas of analysis Maxi can Percent
obtained

235 250 94%

A- General A-Organization

135 250 54%

B- Working Methods

220 250 88%

C- technical monitoring equipment

300 300 100%

D- management work leaves door

110 200 55%

E- Stock Parts

120 200 60%

F- Purchasing and Supply Parts

G- Organization Workshop 130 200 65%

equipment
135 200 67%
H- Tools

135 200 67,5%

İ- Technical Documentation

210 400 52,5%

J- Personnel and Training

105 250 42%

K- Subcontracting

300 300 100%

L- activity control

Average and total score 2135 3000 70,41%

757
Table 4. . Analysis questionnaire maintenance functionafter certification

Analysis areas scores Maxi Percentage

obtained possible

A- General Organization 135 250 54%

B- Working Methods 135 250 54%

C- -technical monitoring equipment 120 250 54%

D- wallet management work 75 300 25%

E- Stock parts 110 200 55%

F- Purchase and supply of spare 105 200 52,5%

G- Organisation workshop material 130 200 56%

H- Tools 25 200 12,5%

İ- Technical Documentation 60 200 30%

J- Personnel and Training 150 400 37,5%

K- Subcontracting 120 250 48%

L- activity control 160 300 53,5%

Average and total scores 1325 3000 44

758
 Fig 2. Route profile maintenance

Table 5. . Summary Table

Areas of analysis Résult % Actions to be taken

- Development of job descriptions

- Update working procedures
A- General Organization 94%
- Establishment of a maintenance budget

- Creation of technical records for critical

equipment
C- technical monitoring
88% - Update on plan changes
equipment
- Management and operation of the historic
failures and maintenance costs

759
 - Implementation of preventive maintenance
policy, systematic and conditional
- Implementation of a CMMS
- Launch program daily visit: failure analysis,
prevention plan, operating range ...
D- management work
100% - Development program for preventive
leaves door
operations
- Preparation of a weekly schedule of
commencement of work

- Developing recommendation PDR

- Monitoring of consumption by PDR equipment
- Analysis of supply times for import purchase
E- Stock Parts 55% - Diversify suppliers PDR
- Establish procurement procedures

Areas of analysis Résult % Actions to be taken

- Mastery of supply times for import

purchase
- Diversify suppliers PDR
- Participation of the acquisition of the
F- Purchasing and Supply Parts 60%
PDR
- Establish procurement procedures

- Extension of repair areas

G – Organization Workshop equipment 65% (maintenance workshop)
- Improved working conditions

760
 - Codification of the collective and
individual equipment
- Purchase of test equipment (crack
H - Tools 67%
detection ...)
- Establishment of the list of equipment
to resupply

I- Technical Documentation
- Acquisition of technical
documentation for new equipment
67,5%
- Setting the provision of filing and
archiving

- Training of enforcement personnel

- Recycling of managerial staff
J- Personnel and Training 52,5% - Developing a career plan for staff

- Monitoring the work of subcontractors

K – Subcontracting 42%
and maintenance team

- Implementation of dashboards and

position indicator
- Analysis and improvement of
maintenance activities through the
L- activity control 100% results

8.RECOMMANDATIONS
- In the area of training, we obtained a score of 150 points total 200 points, so it is necessary to
develop staff training on the following functions:
 Methods and execution
 Maintenance Management (managing the costs of maintenance)
 Inventory Management
 Exploitation of software maintenance
-In the area of maintenance organization was found after certification a score of 130 points total 200
points, so you have:
 Establish a maintenance policy
 Creation of a documentary background
 Development of routings

761
 Updating historical records
 Improved preventive, conditional and systematic
 Improved healing work (use of work requests DT launch orders OT work)
 Development and installation of integrated computer system (CMMS)
 Follow the budget and the cost of maintenance.

Based on the route profile maintenance, there are some weak points (<44%) that can be corrected
with corrective action and that after certification that positively affects the maintenance department for
the purpose of conducting a product conforming to the short production that is to say, to ensure the
availability of the machines the entire period of production to obtain a product in accordance with the
customers requirements.

9. CONCLUSION
For maintenance and achieve good results for it to be effective, it must based on what good
management.
Starting with this idea, analysis of the maintenance function will be to ensure that the organization and
procedures are in compliance with Article "6-3" (infrastructure) ISO 9001 version 2000.
The method of analysis of the maintenance function we studied has three indisputable qualities:

 It is objective:
It does not extreme judgments as "good" or "bad" it leads to identify weak points and strong
points, achieve perfection, ie the twelve points on the circle profile maintenance is unthinkable, it is
turned to action since identifies the domain in which progress is possible.

 It allows dialogue:
The questionnaire is completed by the staff of the showdown response brings up differences
that are extremely interesting to analyze, it can struggle to find a common answer and reflect on the
useful points made in the smooth running of the service.

 It is reproducible
Repeat this maintenance is perfectly feasible and recommended.

9. REFERENCES
1. Isabelle Sulklaper,., J Sanchez, Educate and motivate quality, November 2002.
2. Jean Michel Morin, quality certification services, AFNOR 2001.
3. François Monchy, Maintenance methods and organizations in September 2001.
4. Seminar in ANNABA ISGA, quality and project certification, the November 13, 2000.
5. System of quality management, guidelines for performance improvement, ISO / DIS
9004:2000.
6. TREFILEST, concepts and principles, ISGA April 2001.
7. Company documentation TREFILEST, 2001.

762
 PLASMA-FUEL SYSTEMS UTILIZATION FOR ECOLOGICAL AND ENERGY EFFICIENCY OF
THERMAL POWER PLANTS
1 2
V.E. Messerle , A.B. Ustimenko
1
Combustion Problems Institute, Almaty, Kazakhstan,
Institute of Thermophysics of Russian Academy of Science, Novosibirsk, Russia
2
NTO Plasmotechnics, Research Institute of Experimental and Theoretical Physics at Physical Department of al-Farabi
Kazakh National University, Almaty, Kasakhstan
e-mail: ust@physics.kz

ABSTRACT

This paper presents the results of research and application of direct-flow, vortex and muffle plasma-fuel systems
(PFS) for coal-fired boilers of thermal power plants (TPP) at Ust-Kamenogorsk, Shakhtinsk, and Almaty (TPP-2
and TPP-3) (Kazakhstan). PFS are investigated for boilers with pulverizing systems with direct injection of dust
(Shakhtinsk TPP and Almaty TPP-2) and intermediate bunker (Ust-Kamenogorsk TPP and Almaty TPP-3). Also
this paper presents the results of numerical simulations of the plasma thermochemical enhancement of coal for
ignition and combustion in a furnace of power boiler. The calculations were performed for a low-rank bituminous
coal. 1D model Plasma-coal was used for plasma-fuel system computation. It describes a two-phase chemically
reactive flow in a plasma-chamber with an internal source of heat, i.e. an arc or plasma flame. The kinetic
scheme describes stages of coal devolatilisation, reactions of the gas phase volatile material and heterogeneous
reactions of carbon oxidation. Data to enable a 3D numerical simulation of coal combustion in a furnace
chamber were collected. The 3D numerical experiment was performed with the aid of Cinar ICE code applied to
the boilers of 420 t/h and 75 t/h steam productivity. A comparative analysis of the coal combustion process
enhanced through plasma-fuel systems and without plasma activation was carried out. As a result of the
numerical experiments the advantages of the plasma technology have been clearly demonstrated.

INTRODUCTION

The technology of plasma ignition of coal and its realizing PFS is electro- thermo-chemical preparation of fuel to
burning (ETCPF) [1-6]. In this technology pulverized coal is replaced traditionally used for the boiler start up and
pulverized coal flame stabilization fuel oil or natural gas. Part of the coal/air mixture is fed into the PFS where the
plasma-flame from plasmatron, having a locally high concentration of energy, induces gasification of the coal
and partial oxidation of the char carbon. As coal/air mixture is deficient in oxygen, the carbon being mainly
oxidized to carbon monoxide. As a result, a highly reactive fuel (HRF) composed of mixture of combustible
gases (at a temperature of about 1300 K) and partially oxidized char particles is obtained at the exit of the PFS.
On entry to the furnace, this HRF is easily ignited.
For instance, 75 ton steam productivity boiler (Fig. 1) of Ust-Kamenogorsk TPP has three main pulverized
coal turbulent burners and two kindling muffle burners. The last two were transformed to PFS. Kuznetsk
bituminous coal of 17.7 % ash content and 4878 kcal/kg calorific value was incinerated in the boiler. During the
PFS tests at this boiler the pulverized coal flow through each PFS was 1.5 t/h and the primary air - 2.6 t/h. The
pulverized coal flow through the main burners was 11.5 t/h. Plasmatron power was varied from 60 to 70 kW and
О
its heat efficiency was 85-86 %. HRF flame temperature at the PFS exit was in interval 1040-1240 С.
Plasmatron’s relative power consumptions were 0.5 – 0.7 % of the muffle burner heat power. NOx concentration
3
on the PFS exit was not more than 20 mg/Nm and synthesis gas (СО+Н2) yield exceeded 60 %. In 35 minutes
of the PFS start stationary heat regime of the muffle burner was achieved, plasmatrons were turned off and
heated muffles went on stabilizing the flame combustion. The flames from muffle burners were 3 m in length.
The boiler oil-free start-up lasted 3.25 h after which the boiler was linked up with the main steam pipeline of the
TPP.
75 ton steam productivity boiler of Shakhtinsk TPP has four burners (Fig. 2), two on the front and rear in one
layer. Bituminous coal of 30% ash content with the flow through the burner (or PFS) 3200 kg/h is incinerated in
the boiler. Primary air flow through the burner is 6400 kg/h, plasmatron power is 200 kW and PFS length is 2.3 m
(Fig. 3). Numerical modeling of the ETCPF in PFS is performed using a one-dimensional mathematical model

763
Plasma-Coal. The calculation results allowed defining the geometric dimensions of PFS, the required power of
plasmatron, temperature, velocity and composition of the products of ETCPF. These results can be used as
initial conditions for numerical simulation of HRF combustion in the boiler furnace using Cinar ICE code. 3D
modeling results showed that when operating PFS ignition of pulverized coal flame starts earlier, the combustion
front moves to the installation location of the PFS on the boiler, resulting in lower temperature of the exhaust
gases, the concentration of nitrogen oxides in them and unburned carbon, compared with the traditional mode of
coal incineration without plasma activation in PFS.

Fig. 1. PFS layout at a boiler of 75 ton steam productivity of Ust-Kamenogorsk TPP (cross-section view of the
furnace with PFS): 1 – plasmatron, 2 – chamber for plasma assisted incineration, 3 – muffle burner, 4 – flam of
high reactive two component fuel from PFS, 5 – pulverised coal flame, 6 – air-coal mixture from the main
burners, 7 – furnace, 8 – the main pf burners.

Fig. 2. Layout of the furnace of BKZ-75 power

boiler: 1 – burner throat; 2 – section of the Fig. 3. Sketch of the PFS for replacement of the traditional
swiveling chamber of the boiler. burners of the boiler BKZ-75.

Almaty TPP-3 boiler of 160 t/h steam productivity has four coal-fired blocks of two-layer slot burners.
Consumption of 45 % ash content and 3800 kcal/kg calorific value Ekibastuz bituminous coal was 4 t/h through
each burner. Two PFS were installed in the lower layer of the burners diagonally. Plasmatrons were running on
the power of 120-140 kW (350-450 A current, and 300-350 V voltage). Ignition of the flames in the furnace was

764
observed in 2-3 seconds after submitting of pulverized coal at a rate of up to 3 t/h through each PFS. Coal-dust
O
flame temperature at the exit of the PFS reaches 1200-1300 C, and is 5-6 m in length. Using these parameters,
the formation of the bright yellow core flame in the center of the furnace was observed. In 3.5 hours from the
kindling start parameters of the boiler reached operating values, and it was connected to the main steam
pipeline, after which air/coal mixture was filed to all the burners. According to the rule one start-up of the boiler
consumes 12 tons of fuel oil that by calorific value is equivalent to 30 tons of the coal. Instead, one start-up on
average consumed about 16.5 tons of coal that confirms ETCPF energy efficiency. The specific power
consumptions for plasmatrons were 1.2-1.4% of the heat capacity of pulverized coal burners.

a b
Fig. 4. BKZ-420 boiler furnace of 160 ton steam productivity (Almaty TPP-3) equipped with two PFS (top view)
(a) and sketch of direct flow PFS – burner assembly with plasmatrons: 1 – dust flue, 2 – chamber of plasma
ignition, 3 – plasmatron, 4 – plasma torch, 5 – partition plate, 6 – secondary air duct, 7 – furnace border, 8 –oil
fuel nozzle channel.

NUMERICAL SIMULATION AND THE COMPARISON WITH EXPERIMENTAL DATA

To understand physical mechanism of the ETCPF and combustion more completely, and to verify the software
code CINAR ICE, investigations of coal incineration in the equipped with PFS experimental furnace of 3 MW
thermal power have been fulfilled. Two computer codes were used for that. The first one was 1D code PLASMA-
COAL [1, 2, 7, 8]. It describes a two-phase chemically reactive flow in a plasma-chamber with an internal source
of heat (plasma flame). The second is 3D code CINAR ICE. It operates with real geometry of a furnace.
PLASMA-COAL code has been verified using experimental data on plasma gasification and ETCPF [1, 2, 7, 8]. It
was used for plasma-fuel system computation, and data to enable a 3D numerical simulation of coal combustion
in a furnace equipped with PFS were collected. CINAR ICE code was used for computation of coal co-
combustion with highly reactive two-component fuel in the furnace. The code has been verified for 3D simulation
of traditional furnace processes [9-14]. However, to use this code for computation of the equipped with PFS
furnaces verification was required.
Following the software code CINAR ICE verification numerical comparative simulation of ETCPF and coal
combustion in a furnace of the full-scale boiler was fulfilled.
CINAR ICE code (Cinar Integrated CFD Environment) [9-14] was designed to provide computational
solutions of industrial problems, especially those related to two-phase combustion. It solves the governing time-
averaged Eulerian equations for the gas phase mixture combined with Lagrangian tracking of the particles [15,
16] and simulates devolatilisation, volatiles combustion (fast diffusion combustion), char combustion and the
turbulence (k- model). “Fast chemistry kinetics” model of chemical reactions is used in the code. The kinetic
model is based on the conception of multi-fraction mixtures burning [9]. Radiative heat transfer is represented by

765
six-flow model of particles radiation and reemission [13, 14]. The Control Volume method is used for
discretization of the initial equations. SIMPLE [15] algorithm is used for calculation of pressure field.
To verify CINAR ICE code two regimes of experimental furnace were selected. The first one was
conventional coal combustion and the second regime was combustion plasma activated coal. Plasma activation
to coal combustion was provided with the plasma flame of 36 kW power. The parameters of two-component high
reactive fuel produced in PFS (Fig. 4) were calculated using PLASMA-COAL computer code. These parameters
were used as initial data for PFS equipped experimental furnace3D simulation.
The mathematical model of ETCPF describes the two-phase (coal particles and gas-oxidiser) chemical
reacting flow with an internal heat source (electric arc or flame). Generally coal particles and gas are admitted
into the PFS at ambient temperatures. Particle-to-particle, gas-to-particle transfer and gas-to-plasma heat-mass
transfers are considered. Heat and momentum transfer between the flow and the PFS wall have been
calculated. The average size of the coal particles was taken as a constant and equal to the particle mineral
skeleton size. In other words the particle was not resized at conversion. The particles were spherical and the
temperature gradients within the particle are negligible. Also, some chemical transformations of fuel were
considered. They are as follows: formation of primary volatile products from coal, conversion of evolved volatile
products in the gas phase and the char carbon gasification reactions. The coal composition was presented in the
model by its organic and mineral parts. The organic mass of coal was specified by the set of the functional
groups (CO, CO2, CH4, H2O, C6H6) and carbon. An entrained flow reactor was considered and the plug flow was
assumed. The resulting set of ordinary differential equations includes equations for species concentrations
(chemical kinetics equations) in conjunction with the equations for gas and particle velocities and temperatures,
3
respectively. The energy contribution from the plasmas had been found empirically and included into the
energy equation as a distribution of internal heat source. Also the model was distinguished by its detailed
10
description of the kinetics of the chemical reactions mentioned above . Kinetic scheme consists of 116
chemical reactions. The temperature dependence of rate constants is governed by the Arrhenius equation. In
Ref. 2, 3 and 5 the model was presented in detail.

Fig. 4. Layout drawing of plasma-fuel system: 1 – plasmatron, 2 – electromagnetic coil, 3 – air/coal mixture, 4 –
ETCPF chamber, 5 – air/coal mixture scroll, 6 – air/coal mixture, 7 – window for temperature measuring and gas
sampling, 8 – secondary air scroll, 9 – central tube of the burner, 10- windows for temperature measuring and
gas and coke residue sampling, 11 – furnace, 12 – secondary air, 13 – plasma forming air.

The swirl burner was mounted axial on the furnace top (Fig. 4). Ekibastuz bituminous coal of 45.2 % ash
content, 14.7 % devolatilization, 1.3 % humidity and 15960 kJ/kg heat value was incinerated in the experimental
furnace [7, 13]. Grinding of coal gave the average particle size of 60 microns. PLASMA-COAL computer code
has been used for calculation of ETCPF in the volume of PFS of 1.15 m length. The following initial parameters

766
were used for calculations: plasmatron power was 36 kW, initial temperature of pulverized coal (coal/air mixture)
О
was 300 С, coal and air consumption through PFS were 410 and 600 kg/h correspondingly. The air is
composed of nitrogen (79 vol. %) and oxygen (21 vol. %) and neither carbon dioxide nor noble gas were taken in
consideration.
1200 35

2
1000 30
1
Temperature, C

Velocity, m/s
2
25
O

1
800

20
600

15
400
10
0.00 0.15 0.30 0.45 0.60 0.75 0.90 1.05 1.15 0.00 0.15 0.30 0.45 0.60 0.75 0.90 1.05 1.15
PFS lenght, m PFS lenght, m

Fig. 5. Gas (1) and particles (2) temperatures along Fig. 6. Gas (1) and particles (2) velocities along the
the PFS. PFS.
As a result of calculations temperature distribution (Fig. 5), gas and particles velocities (Fig. 6), gas-phase
components concentrations (Fig. 7), gasification degree and carbon concentrations in coke residue (Fig. 8) were
found. Gas and coal particles temperatures (Fig. 5) increase along the PFS. The heat exchange between
plasma source and gas-phase is prevailing at the initial section (0< Х < 0.35 m). Coal particles are heated from
О
gas and their temperature increase up to 1121 С at the section 0.35≤Х<0.8 m due to carbon oxidation and
О
corresponding out of heat. It exceeds the gas temperature by 264 С. As a result of that inversion of temperature
О
curves is observed at the section. Gas temperature reaches maximum at 1015 С (Х=0.9 m) and goes down to
О
the outlet of PFS (T=1002 С). Gas temperature is 41 degrees higher than particles temperature, that is related
to heat-emission from the particles to the wall of PFS.
Equal at the PFS inlet velocities of gas and particles (Fig. 6) increase with the length achieving their maxima
of 33.8 and 33.2 m/s at the PFS exit accordingly. Meanwhile gas velocity is higher one of particles along the full
PFS. Note velocity of the flow at the exit of PFS is considerably higher than one for traditional pulverized coal
burner.
With coal particles heating it is observed devolatilization and carbon gasification (Figs. 7 and 8).
Concentration of combustible components (CO, H2, Н, CH4, C6H6) increases along the PFS and reaches its
maximum of 41.8 % at the PFS outlet. Oxidants concentration (CO 2, H2O, O2) does not exceed 7.3 % at the PFS
outlet. The nitrogen concentration (N2) decreases along the PFS from 79 % to 50.8 % at the outlet. The carbon
concentration in the coke residue decreases, and carbon gasification degree (Fig. 8) increases along the PFS
and achieves 69.5 % at the outlet. The two-component fuel with aforesaid characteristics is intensively ignited
mixing with the secondary air flow in the furnace volume. Calorific value of the coke residue was amounted to
7200 kJ/kg.
Integral characteristics of ETCPF at the PFS exit are gathered in Table 1. It the table T gas, Tsolid and VHRF are
temperatures of gas and particles and velocity of the HRF correspondingly. They were taken as initial data for
numerical simulation of ETCPF combustion in the experimental furnace using CINAR ICE software code.
Secondary air flow rate was 2322 kg/h. The furnace height was 7.5 m. Computing origin was the furnace top.
The x, y, z grid size was, respectively: 56 x 56 x 60.
The calculations results are shown in Figs. 9 and 10. Figure 9 visualizes the difference between
temperatures fields for the furnace two operational regimes. The traditional mode flame, with maximum
O O
temperature of 1580 С, generates the common flame body with the temperature 1300-1580 С. The PFS impact
appears as transformation of high reactive two-component flame shape, decrease of the flame length and
O
increase of temperature maximum up to 2015 С. That can be explained by earlier ignition and more complete
burning-out of ETCPF.

767
 100 70
N2
H2 60
CO
10 CO2
O2 50
1
Ci, vol, %

H2O
40

Y, %
1 C6H6
30
CH4 2
20
0.1
10
H
0.01 0
0.00 0.15 0.30 0.45 0.60 0.75 0.90 1.051.15 0.00 0.15 0.30 0.45 0.60 0.75 0.90 1.05 1.15
PFS lenght, m PFS lenght, m

Fig. 7. Gas phase components concentrations along Fig. 8. Gasification degree (1) and carbon
the PFS. concentration (2) in coke residue along the PFS.

Table 1. Characteristics of ETCPF at the PFS exit.

Content of gas phase, vol.% & kg/h Ash, Char carbon, Tgas, Tsolid, VHRF,
О О
H2 CO CH4 C6 H6 CO2 H2O N2 kg/h kg/h С С m/s
21.6 19.2 0.4 0.7 5.8 1.6 50.8
185.3 68.4 1002 961 33.8
14.0 174.2 2.2 18.0 82.2 9.1 462.0

Experimental and numerical temperatures dependences on the furnace height (Fig. 10) are similar. The
curves have typical maxima at the distance of 0.5 m from the top of the furnace. The maximal temperature level
for conventional mode of the furnace operation at the distance H<0.13 m is higher than one for plasma activated
fuel combustion. This difference is up to 164 degrees (0.025<H<0.13 м). It can be explained by more intensive
radiation from coal particles having higher concentration and total surface when the furnace operation is in
conventional mode in comparison with plasma activated regime of coal combustion. When PFS operates electro-
thermochemically prepared fuel (two component high reactive fuel) is gone from it to the furnace. The fuel
contains combustible gas and particles of coke residue which mass do not exceed 30 % of the initial coal mass.
That brings to triple decrease of radiative particles total surface. In section 0.13≤H<0.6 m temperature at plasma
activated regime is higher than one for conventional mode of the furnace operation. The maximum difference
reaches 260 degrees at height 0.4 m. The maximum combustion temperature of plasma activated fuel at the
experiment was 200 degrees higher one for conventional mode. Experimental temperature maximum was at the
furnace height of 0.67 m. Higher temperature of the flame of plasma activated fuel can be explained by more
complete burnout of the fuel.
Thus, the experimental data [8, 16] were used for verification of CINAR ICE computer code. Divergence of
experimental and calculated values of temperature of combustion products does not exceed 20 % along the full
height of the furnace. The divergence explains by imperfection of the "fast chemistry" scheme of fuel combustion
used by CINAR ICE program. The concentration of unburned carbon was measured at the exit of the
experimental furnace. Also divergence of the experimental and calculated values does not exceed 20 %. In
experiments unburned carbon was found as 3.1 and 1.8 % for traditional and plasma activated mode
correspondingly. Computed figures were 3.72 and 1.42 % accordingly for these two modes. Concentration of
nitrogen oxides at the furnace exit is significantly lower for ETCPF combustion (244.5 ppm) than for conventional
mode of coal combustion (537.9 ppm). Calculated values of NOx concentrations for these two modes were
122.1 and 228.1 ppm. However, essential quantitative difference of experimental and calculated values of NOx
concentrations shows that CINAR ICE kinetic scheme of NOx formation is in need of revision.

768
 1800

1600

1400

Temperature, C
O
1200
1
1000

800 2

600

400
0 1 2 3 4 5
Furnace height, m

Fig. 10. Furnace height distribution of the combustion

a b products maximal temperature: 1 – conventional
Fig. 9. Temperature fields along the furnace incineration of coal; 2 – incineration of coal in the furnace
height in the central section for conventional (a) and with PFS; ●, ○ – experimental data on coal incineration in
plasma supported (b) regimes. the furnace with PFS and without PFS correspondingly.

COMPUTATION OF PFS AND FULL-SCALE INDUSTRIAL BOILER'S FURNACE

The fulfilled verification of CINAR ICE code for plasma assisted coal combustion in the experimental furnace of 3
MW power confirmed legitimacy of the used codes complex (PLASMA-COAL and CINAR ICE) for simulation of
the furnaces equipped with PFS. Thus in this part the numerical study was performed for a power-generating
boiler with a steam productivity of 75 t/h. The boiler’s furnace (Fig. 2) is equipped with four swirl burners
arranged in one layers, by two burners, on the boiler front and backside. Low-rank bituminous coal of 35.1 %
ash content, 22 % devolatilization and 18550 kJ/kg heat value was incinerated in the furnace. Averaged size of
the coal particles was 75 micron. All the calculations were performed in accordance to the aforecited technique.
Three modes of the boiler operation were chosen for the numerical studies. The first one was traditional
regime, using four pulverized coal burners, the second one was regime with plasma activation of combustion,
using the replacement of two burners onto PFS (Fig. 3), and the third one was regime of the boiler operation
using four PFS instead of all burners.
PLASMA-COAL computer code has been used for calculation of ETCPF in the volume of PFS of 2.3 m
length. The following initial parameters were used for calculations: plasmatron power was 200 kW, initial
О
temperature of pulverized coal (coal/air mixture) was 90 С, coal and primary air consumptions through PFS
were 3200 and 6400 kg/h correspondingly.

769
 The results of numerical simulation by the PLASMA-COAL code are summarized in Table 2. Heat value of
the coke residue was 8580 kJ/kg. These data were taken as initial parameters for 3-D computation of the
furnace of the power-generating boiler equipped with PFS. This computation was performed using CINAR ICE
code to demonstrate advantages of plasma aided coal combustion technology.

Table 2. Characteristics of ETCPF at the PFS exit.

Content of gas phase, vol.% & kg/h Char carbon, VHRF,
O
Ash, kg/h THRF, C
H2 CO CH4 C6H6 CO2 H2O N2 O2 kg/h m/s
14.2 18.4 0.3 0.6 6.8 2.9 56.4 0.3
1123.2 435.0 997 189
88.5 1599.0 14.0 133.8 931.2 162.8 4911 31.0

Initial parameters for calculations of the furnace (Fig. 2) in different operational regimes were the following:
О
temperature of the secondary air was 290 С, coal productivity of the burner was 3200 kg/h and primary air flow
rate through the burner was 10260 kg/h. Secondary air flow rate to the boiler was 78160 kg/h. The grid is
defined by 85 x 69 x 116 grid lines in three directions (x, y and z).
The calculations results are shown in Figs. 11 and 12. Figure 11 visualizes the difference between
temperature fields in three regimes of coal incineration. In the traditional regime (Fig. 11 a), with maximum
О
temperature of 1852 С, four symmetrical flames are generated. In central space of the furnace they form overall
O
body of flame with the temperature about 1300 С. In Fig. 11 b two PFS are on top. The PFS impact appears as
O
increase of temperature maximum up to 2102 С and transformation of HRF flame shape, it becomes narrow
and longer. When the furnace operates with four PFS (Fig. 11 c) the flames length increases but maximal
О
temperature decreases to 1930 С.

a b c
Fig. 11. Temperature field within the combustion chamber at the level of the pulverized coal burners: a –
standard operational regime; b – plasma operational regime with two PFS; c – plasma operational regime with
four PFS.
Average characteristics of the boiler are compared in Figs. 12-15 for three modes of the boiler operation.
The temperature curves have a characteristic maximum in the zone of the burners arrangement at a height of 4
m (Fig. 12). In the traditional mode of combustion level of the average temperature in the furnace at a height of
up to 6 m higher than that for the boiler operating with PFS. The temperature difference reaches 75 degrees
(height between 2 and 3 m), due to more intense radiation from the coal particles having a higher concentration
and the total surface at the traditional combustion, compared to the operation mode with PFS. From the PFS
HRF enters the combustion chamber, consisting of fuel gas and coke residue particles, whose mass does not
exceed 30 % of the consumption of raw coal, which leads to a threefold reduction in the total surface of the
radiating particles. Further, the section of the furnace from 4.5 to 16.75 m, the temperature in the regime with

770
PFS higher than that for the traditional burning by 10 and 32 degrees in the case of 2 and 4 PFS respectively.
This is due to more complete fuel burnout (Fig. 13) by ETCPF confirmed by decreased oxygen concentration in
the furnace at the same location (Fig. 14). PFS improves the environmental characteristics of the combustion of
solid fuels. Compared with the traditional mode of coal incineration use of four PFS reduces the unburned
carbon at the outlet of the furnace (height of 16.75 m) 4 times, and nitrogen oxide emissions by more than 2.2-
fold (Fig. 15).

1300
8.8
1200
2 3 8.4
Temperature, С

1100 8.0 1
о

O2, mass %
1
1000 7.6
2

900 7.2

6.8
800
6.4
700 3
6.0
2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16

Furnace height, m Furnace height, m

Fig. 12. Furnace height distribution of mass average Fig. 13. Furnace height distribution of oxygen mean
temperature: 1 – standard operational regime; 2 - concentration: 1 – standard operational regime; 2 -
regime with two PFS; 3 - regime with four PFS. regime with two PFS; 3 - regime with four PFS.

COMPUTATION OF 420 T/H STEAM PRODUCTIVITY BOILER’S FURNACE EQUIPPED WITH PFS

The fulfilled verification of CINAR ICE code for plasma assisted coal combustion in the experimental furnace of 3
MW power confirmed legitimacy of the used codes complex (PLASMA-COAL and CINAR ICE) for simulation of
the furnaces equipped with PFS. Thus in this part the numerical study was performed for a power-generating
boiler with a steam productivity of 420 t/h. The boiler (Fig. 11) is equipped with 6 swirl burners arranged in two
layers, three burners each, on faced wall of the furnace. As it is seen from the figure three PFS are installed
instead of two burners of the lower layer and one of the upper layer. Low-rank Ekibastuz bituminous coal of
40 % ash content, 24 % devolatilization, 5 % humidity and 16700 kJ/kg heat value was incinerated in the
furnace. The coal grinding fineness is R90=15 %. All the calculations were performed in accordance to the
aforecited technique.
PLASMA-COAL computer code has been used for calculation of ETCPF in the volume of PFS of 3.687 m
length. The following initial parameters were used for calculations: plasmatron power was 200 kW, initial
О
temperature of pulverized coal (coal/air mixture) was 90 С, coal and air consumptions through PFS were 6000
and 8955 kg/h correspondingly.
The results of numerical simulation by the PLASMA-COAL code are summarized in Table 2. Heat value of
the coke residue was 6165 kJ/kg.These data obtained for the PFS exit were taken as initial parameters for 3-D
computation of the furnace of a power-generating boiler equipped with PFS. This computation was performed
using CINAR ICE code to demonstrate advantages of plasma aided coal combustion technology.
Initial parameters for calculations of the furnace (Fig. 16) in different operational regimes were the following:
О
temperature of the secondary air was 280 С, coal productivity of the burner was 12000 kg/h and primary air flow
rate through the burner was 17900 kg/h. Secondary air flow rate to the boiler was 446412 kg/h. Averaged size of
the coal particles was 60 micron. The furnace size is as follows: 27 m height, 7.7 m depth and 14.5 m width. The
x, y, z grid size was, respectively: 106 x 38 x 104.

771
 1.4 300
1
270
1.2
1
240
1.0

NOx, ppm
C, mass %

210
0.8
180
2 2
0.6 150

0.4 120 3

3 90
0.2
2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16
Furnace height, m Furnace height, m

Fig. 14. Furnace height distribution of carbon mean Fig. 15. Furnace height distribution of NOx mean
concentrations: 1 – standard operational regime; 2 - concentration: 1 – standard operational regime; 2 -
regime with two PFS; 3 - regime with four PFS. regime with two PFS; 3 - regime with four PFS.

a b
Fig.16. Scheme of industrial 420 t/h steam productivity boiler in Almaty TPP-2 (Kazakhstan) retrofitted with PFS
(a): 1 – standard pulverized coal swirl burner, 2 – PFS; and layout of the PFS for the boiler (b): 1 – channel of
the external flow of pf, 2 – secondary air duct, 3 - inlet of pf external flow, 4 – inlet of pf internal flow, 5 -
plasmatron, 6 – chamber for pf flow turning, 7 – chamber for plasma chemical preparation of fuel for combustion,
8 - chamber for mixing and thermochemical preparation of fuel, 9 - furnace.

The model predictions are presented in the following figures which show results for the plasma-activated
coal combustion in comparison with conventional coal combustion. Figs. 17 and 18 show temperature fields
along the furnace height in the mean cross-sectional plane for two regimes of the furnace operation, traditional
(Fig. 17) and plasma activated coal combustion (Fig. 18). The figures visually demonstrate the difference
between the temperature fields for the two modes of coal combustion. When the coal combustion is in
conventional mode, six symmetric pulverized coal flames are formed. Maximum temperature of these flames is

772
 O
1852 C. In Fig. 18 one can see influence of PFS on to the shape of the ETCPF flame and its maximal
temperature. In the presented plane PFS is upwardly. High temperature body of the flame are moved closer to
O
the PFS exit and upper in the furnace. Its maximal temperature is 1588 C.

Table 3. Characteristics of ETCPF at the PFS Exit.

Content of gas phase (vol.% & kg/h) Ash, Char carbon,
H2 CO CH4 C6H6 CO2 H2O N2 O2 kg/h kg/h
1.05 7.75 0.3 0.77 15.6 3.55 70.84 0.15
1518 261
7.272 751.4 16.75 207 2378 220.5 6870 16.49
Gas temperature (К) Solids temperature (К) Velocity of the flow (m/s)
1025 1025 48.2

o
Fig. 17. Temperature field ( C) along the Fig. 18. Temperature field along the furnace
furnace height in the mean cross-sectional plane of the height in the mean cross-sectional plane of the furnace
furnace when it works in conventional mode of coal when it works in plasma-assisted mode of coal
combustion. combustion using 3 PFS.

Averaged temperature curves (Fig. 19) have their maxima. The first one (H = 3 m) is generated by
overheating of the furnace back wall by the pulverized fuel flame. The second maximum is above level of the
burners of the upper layer due-to common forming of the flame body and observed moving as a result of natural
convection. Averaged temperatures in the furnace operated in conventional mode are higher one for the furnace
operated in plasma assisted regime using PFS. The difference achieves 350 degrees at the furnace exit. The
reason of this is more intensive radiation of coal particles which have higher concentration and total reacting
surface when the furnace operates in conventional mode in comparison with plasma assisted coal combustion.
When PFS operates two component fuel of combustible gas and particles of coke residue enters the furnace.
Mass of this fuel does not exceed 30 % of the initial coal mass. That decreases total surface of the radiative
particles.

773
 At the furnace exit when three PFS operate concentration of unburned carbon (Fig. 20) is 16 % less one
when the furnace works in traditional mode of coal combustion. Use of PFS improve ecological characteristics of
the process of solid fuel combustion. Fig. 21 demonstrates more than 33 % decrease of nitrogen oxides
concentration. Evidently decrease of unburned carbon and NOx concentrations at the furnace exit improves
ecology-economic indexes of TPP.

1300 6

5
1200
1
1 4
1100
3

C, %
T, С
о

1000 2
2
2
900
1

800 0
0 5 10 15 20 25 0 5 10 15 20 25
H, m H, m

Fig. 19. Mean concentration of temperature (T) along Fig. 20. Mean concentration of unburned carbon (C)
the furnace height (H): 1 – conventional incineration of along the furnace height: 1 – conventional incineration
coal, 2 – incineration of coal when 3 PFS operate. of coal, 2 – incineration of coal when 3 PFS operate.

240

1
210
NOx, ppm

180

150
2

120

0 5 10 15 20 25
H, m

Fig. 21. NOx mean concentration along the furnace height: 1 – conventional incineration of coal, 2 – incineration
of coal when 3 PFS operate.

CONCLUSIONS
Simulation and testing of PFS at existing coal-fired boilers of TPP confirmed the technical feasibility,
environmental and energy efficiency of fuel oil free boilers start-up and pulverized coal flame stabilization using
PFS.
PFS tests at the boilers BKZ-160 and BKZ-420 of Almaty Power System in the mode of the boilers start-up
from cold confirmed possibility of high ash Ekibastuz coal ignition.
Inculcation of PFS gives economical effect, which depends on coal/fuel oil price ratio (Table 4). Pay back
period varies from 12 to 18 months. Economical effect for TPP of Kazakhstan is 400 mln. of US dollars a year.

774
 Table 4. Economical comparison of the PFS technology with traditional one.
Conventional technology Plasma technology
1. Fuel Oil Rate for Russian TPP
5.1 mln. t/year (cost is more than $ 2 billion) 0
2. Fuel Oil Rate for Kazakhstan TPP
~1 mln. t/year (cost is about $ 400 mln.) 0
3. Investments for TPP
100% 3-5%
4. Operating costs
100% 28-30%
5. Electric power consumption for TPP auxiliary
3-5% 0.5-1.0%

REFERENCES
[1] Z.B. Sakipov, V.E. Messerle, Sh.Sh. Ibraev, “Electrothermochemical preparation of coal to burning” (in
Russian), Almaty: “Nauka”, 1993, 259 p. (in Russian)
[2] Messerle V.E., Ustimenko A.B. “Plasma ignition and combustion of solid fuel. (Scientific-and-
technological basics).” (in Russian). Saarbrucken, Germany: Palmarium Academic Publishing (ISBN: 978-3-
8473-9845-5), 2012, 404 p. Available: http://ljubljuknigi.ru/
[3] M.A. Gorokhovski, Z. Jankoski, F.C. Lockwood, E.I. Karpenko, V.E. Messerle, A.B. Ustimenko,
“Enhancement of Pulverized Coal Combustion by Plasma Technology”, Combustion Science and Technology,
2007, vol. 179, N 10, pp. 2065–2090.
[4] Karpenko E.I., Messerle V.E., Ustimenko A.B. “Plasma-aided solid fuel combustion.” Proceedings of the
Combustion Institute. Elsevier, 2007, vol. 31, part II, pp. 3353-3360.
[5] Karpenko E. I., Karpenko Yu. E., Messerle V. E., and Ustimenko A. B., “Using Plasma-Fuel Systems at
Eurasian Coal-Fired Thermal Power Stations”, Thermal Engineering, 2009, vol.56, N 6, pp.456-461.
[6] A.S. Askarova, E.I. Karpenko, Y.I. Lavrishcheva, V.E. Messerle, A.B. Ustimenko, “Plasma-Supported
Coal Combustion in Boiler Furnace”, IEEE Trans. Plasma Sci., Dec. 2007, 35 (6), pp.1607–1616.
[7] R.A. Kalinenko, A.P. Kuznetsov, A.A. Levitsky, V.E. Messerle, Yu.A. Mirokhin, L.S. Polak, Z.B. Sakipov,
A.B. Ustimenko, “Pulverized Coal Plasma Gasification”, Plasma Chemistry & Plasma Processing, 1993, vol. 13,
N 1, pp. 141-167.
[8] M. Gorokhovski, E.I. Karpenko, F.C. Lockwood, V.E. Messerle, B.G. Trusov, A.B. Ustimenko, “Plasma
Technologies for Solid Fuels: Experiment and Theory”, Journal of the Energy Institute, 2005, vol. 78, N 4, pp.
157-171.
[9] Z. Jankoski, F.C. Lockwood, V.E. Messerle, E.I. Karpenko, A.B.Ustimenko, “Modelling of Plasma Pre-
Treatment of Powdered Coal for Combustion”, Thermophysics and Aeromechanics, 2004, vol. 11, N 3, pp. 461-
474.
[10] F.C. Lockwood, A.P. Salooja, A.A. Syed, “A prediction method for coal-fired furnaces”, Combustion and
Flame, 1980, vol. 38, N 1, pp.1-15.
[11] F.C. Lockwood, T. Mahmud, M.A. Yehia, “Simulation of pulverised coal test furnace performance”, Fuel,
1998, vol. 77, N 12, pp. 1329.
[12] F.C. Lockwood and T. Mahmud, “The Prediction of Swirl Burner Pulverised Coal Flames”, Twenty-
Second Symposium (International) on Combustion, Pittsburgh, PA: The Combustion Institute, 1988, pp. 165.
[13] F.C. Lockwood, N.G. Shan, “Evaluation of an efficient radiation flux model for furnace prediction
procedures” Proc. Sixth Intern. Heat Transfer Conference, 1978, pp. 1405-1413.
[14] F.C. Lockwood, N.G. Shan, “A new radiation solution method for incorporation in general combustion
prediction procedures”, Proc. 18th Intern. Symp. On Combustion. Pittsburg, The Combustion Inst., 1981, pp.
1405-1413.
[15] B.E. Launder, D.B. Spalding, “The numerical computation of turbulent flows”, Comput. Meth in Appl.
Mech. And Eng., 1974, vol. 3, N 2, pp. 269-289.
[16] Messerle V.E., Karpenko E.I., Ustimenko A.B., Lavrichshev O.A. “Plasma preparation of coal to
combustion in power boilers” Fuel Processing Technology, March 2013. V 107. pp. 93–98.

775
 PLASMA TECHNOLOGIES OF SOLID AND GASEOUS FUELS PROCESSING

1 2
V.E. Messerle , A.B. Ustimenko
1
Combustion Problems Institute, Almaty, Kazakhstan,
Institute of Thermophysics of Russian Academy of Science, Novosibirsk, Russia
2
NTO Plasmotechnics, Research Institute of Experimental and Theoretical Physics at Physical Department of al-Farabi
Kazakh National University, Almaty, Kazakhstan
e-mail: ust@physics.kz

ABSTRACT

This paper presents review of developed plasma technologies of fuels and minerals utilization. Technical
realizability, energy and ecological-economical effectiveness of these technologies are shown. Also
mathematical models and realizing codes for numerical investigation of the processes of fuels and minerals
processing in plasma and electromagnetic reactors, plasma-fuel systems and power boilers are described.

INTRODUCTION

Technologies of fuel and minerals utilization, which were developed and mastered in commercial or
experimental-industrial scale, are discussed [1-10]. They are plasma technology of oil (gas)-free boilers start-
up and pulverized coal flame stabilization at coal fired thermal power plants; plasma technology of power
coals processing to carbonic sorbents; plasma gasification of coal; plasma technology of pulverized coal
ignition and the flame stabilization in rotary furnace for alumina aglomeration in aluminum industry and for
oil-free clinker firing at cement production; plasma technology of oil refining residuum utilization; plasma-
cyclone technology of bricks firing; energy-saving electromagnetic technology for mineral materials melting.
These technologies are realized with the aid of numerical modeling and plasma-fuel systems (PFS) which
were developed to improve efficiency of power coals combustion to decrease share of fuel oil and natural
gas in fuel balance of Thermal Power Plants (TPP) and to decrease harmful emissions.

PFS AND PLASMA TECHNOLOGY OF OIL (GAS)-FREE BOILERS START-UP AND PULVERIZED COAL
FLAME STABILIZATION AT COAL FIRED THERMAL POWER PLANTS

PFS are combination of pulverized fuel burners with DC arc plasma torches. Using PFS fuel oil is replaced
by pulverised coal, which is a subject to thermochemical pre-process to the combustion within the PFS. In
this technology part of the coal/air mixture is fed into the burner where the plasma-flame, having a locally
high concentration of energy, induces gasification of the coal and partial oxidation of the char carbon. The
resulting coal/air mixture is deficient in oxygen, the carbon being mainly oxidized to carbon monoxide. As a
result, a highly reactive mixture of combustible gases (at a temperature of 900-1200 K) and partially oxidized
char particles are obtained at the exit of the burner. On entry to the furnace, this combustible mixture is
easily ignited. This allows prompt ignition and much enhanced flame stability of the main portion of the coal
flame which is not directly treated by the plasma. Activation of coal combustion by this means eliminates the
need for supplementary fuel consumption (fuel oil or natural gas), traditionally used for the start-up of a coal-
fired furnace. Fig.1 illustrates the process of thermochemical preparation of coal to combustion. The figure
presents scheme of the developed direct flow PFS. The process is going on at relatively low specific power
consumption of 0.05-0.4 kWh/kg of coal. It was experimentally proven that PFS use increases efficiency of
coal ignition and combustion, eliminates fuel oil expenditure for start-up and flame stabilization, decreases
unburned carbon, NOx, SOx, V2O5 emission, and provides ample scope for the process automation.
PFS have been tested for boilers start-up by plasma ignition of pulverized coal and flame stabilization in
different countries on 30 pulverized coal boilers steam-productivity 75 to 670 t/h equipped with different types
of pulverized coal burners (direct flow and swirl burners). When PFS testing power coals of all ranks (brown,
bituminous, anthracite and their mixtures) were incinerated (Table 1). Volatile content of them was from 4 to
50%, ash varied from 15 to 56% and heat of combustion was from 1600 to 6000 kcal/kg. The advantages of
PFS technology were confirmed by 3D simulation of the boilers equipped with PFS.
For instance the schematic view of the boiler equipped with PFS and its primary dimensions are shown in
Fig. 2. The furnace is characterised by two symmetrical combustion chambers, each having 4 tangentially
directed main double burners in two layers. Combustion chambers are interconnected by a central section.

776
 3
Fuel consumption on the boiler was 121.3 t/h and total amount of air in the boiler was 553128 Nm /h at
temperature of the secondary air 350C.

Fig. 1. Direct flow PFS: 1- plasmatron; 2 – air/coal mixture input; 3 – secondary air input; 4 – side of a boiler
furnace; 5 – furnace.

Table 1. Technical characteristics of coals

Coal type W
w d
A V
daf
Qlw (kcal/kg)
Brown 25-35 15-20 35-50 3000-3800
Lignite 32-40 28-35 23-27 1900-2100
Shale oil 40-50 75-80 48-50 1600-2000
Bituminous 5-12 20-56 15-40 3500-5000
Anthracite 5-8 25-35 4-10 4300-6200
Coal mixture 10.4 48.5 38.2 3150
w d daf
Nomenclature: W = wet per working mass; A = ash per dry mass; V = volatile
per dry ash-free mass; Q = low heat of coal combustion per working mass.
w
l

Fig. 2. Scheme of the industrial furnace of BKZ 640-140 boiler.

Four PFS are mounted instead of four lower sections of the main double burners as it is shown in Fig. 2.
The plasmatrons operate during the boiler warm-up period and in the case of an unstable flame. When the
boiler performance is stabilised, the plasmatrons are switched off and the PFS continue to function as
conventional pf burners. In the case of flame instability, the plasmatrons are restarted.

777
 Knowledge of the specific power consumption of a plasmatron is required to estimate PFS efficiency. This
parameter is defined as the ratio of plasmatron electric power to pf consumption in the PFS. Figs. 3 and 4
present experimental results for NOX reduction and the decrease of unburned carbon during PFS operation
versus specific power consumption for the plasmatron. It is seen that the NOX concentration is halved, and the
amount of unburned carbon is reduced by a factor of 4. The NO X decrease is caused by the fact that the fuel
nitrogen, released from the coal inside the PFS in conditions of oxygen deficiency, forms molecular nitrogen in
the gas phase. Since the fuel nitrogen is evolved inside the PFS and converted to molecular nitrogen there,
mainly thermal nitrogen oxides are formed within the combustor volume. However, fuel nitrogen is the main
source of nitrogen oxide emission from conventionally-fired pf combustors. As to unburned carbon, its
decrease indicates a fuel reactivity increase which is explained by enlargement of the coal particles reactive
surface due to ‘‘heat explosion’’ and fragmentation as a result of their interacting with arc plasma.

Fig. 3. Influence of the specific power consumption for Fig. 4. Influence of the specific power consumption for
plasmatron on reduction of the nitrogen oxides (NOx) plasmatron on reduction of unburned carbon at
concentration at plasma-assisted pf combustion. plasma-assisted pf combustion.

PLASMA-STEAM GASIFICATION AND COMPLEX PROCESSING OF COAL TO PRODUCE SYNTHESIS

GAS AND VALUABLE COMPONENTS FROM ASH

Plasma gasification and complex processing of coal to produce synthesis gas and valuable components from
mineral mass of coal (MMC) were investigated using universal experimental setup. Plasma gasification and
complex processing of low-rank coals is oriented towards perspective of several industry branches
development (power engineering, chemical industry, metallurgy). Final products are synthesis gas (СО+Н2),
hydrogen, silicon, ferrosilicon, carbosilicon and alumina. From an environmental point of view, these
technologies are most promising. Their essence is heating of coal dust by the arc plasma, which is oxidant,
to complete gasification temperature at which the coal organic matter is transformed into an environmentally
friendly fuel - synthesis gas free of ash particles, nitrogen oxides and sulfur oxides. At complex processing of
coal simultaneously with gasification of organic matter in the same reaction volume coke carbon restores
MMC oxides and valuable components, such as industrial silicon, ferrosilicon, aluminum and carbosilicon
and microelements of rare metals: uranium, molybdenum, vanadium, and others, are formed. Fig. 5
demonstrates experimental installation for this process realization. It consists of heating the dust coal by arc
plasma, which is oxidant, up to temperature of complete gasification. Thus organic mass of coal is conversed
into environmentally friendly fuel, synthesis gas, which is free from ash particles, nitrogen oxides and sulphur
oxides. Simultaneously reduction of the oxides of mineral mass of coal is taken place and valuable
components like silicon, ferrosilicon, aluminum, carbosilicon and microelements of rare metals (uranium,
molybdenum, vanadium etc.) are produced.
The experimental installation is intended for work in the range of electric power 30-100 kW, mass
averaged temperature 1800-4000 K, coal dust consumption 3-12 kg/h and water steam flow 0.5-15 kg/h. The
steam/pulverised fuel mixture entering the arc zone is heated to high temperature by the arc rotating in
electromagnetic field to produce a two-phase plasma flow where the solid fuel gasification process occurs.
Petrocoke (CP), brown (PBC) and bituminous (KBC) coals of 3, 48.1 and 44 % ash content
correspondingly (Table 2) have been gasified in the plasma gasifier. The sieve analysis of the pulverized fuel
revealed that mean sizes of the CP, PBC and KBC dust particles were 105, 60 and 75 µm correspondingly.
Thermal efficiency of the reactor was determined as 76 % for the experiments. As a result of the solid fuels

778
gasification under steam plasma conditions concentrations of gas species, specific power consumptions
(SPC), carbon gasification degree (XC) and mass-averaged temperatures (TAV) in the reactor were revealed
(Table 3). High quality synthesis gas was produced in the experiments on solid fuels plasma gasification. It
has been found that the synthesis gas content of 85 to 99 vol.% can be produced, and gasification degree can
be reached as high as 94.2 %.

Fig. 5. Scheme of plasma unit for carbonaceous materials processing: 1 - reactor; 2 – chamber of gas and
slag separation; 3 – slag catcher; 4 – chamber of synthesis-gas removal; 5 – diaphragm; 6 – chamber of
hydration; 7 – pulverized coal feeder; 8 – the cooling system; 9, 10 – electric power supply system; 11, 12 –
gear for rod electrode delivery; 13 – steam generator; 14 – safety valve; 15 – stand for slag catcher.

At complex processing of a low rank brown coal (TBC) of 28 % ash content and 13180 kJ/kg caloricity
(Table 2) simultaneously with gasification of organic matter in the same reaction volume coke carbon restores
MMC oxides and valuable components are formed. Synthesis gas yield reached 95.2% and carbon
gasification degree was 93.2 % (Table 3). Table 4 presents degree of reduction (Θ) of the samples of solid
residue from different parts of the installation including special pool for melt near graphite orifice between the
reactor 1 and chamber of gas and slag separation 2 (Fig. 5). It is seen that restored material was found in the
slag in the form of ferrosilicon, silicon and ferric carbides. The highest degree of reduction of the oxides of
MMC 47 % was found in the slag from the wall of the reactor in the zone of maximal temperatures.
The installation (Fig. 5) was also used for realization some other following technologies. Technologies of
plasma conversion of carbonaceous materials (coal, petroleum coke, hydrocarbon gases) are characterized
by high levels of temperature and, therefore, a high degree of their thermochemical conversion to desired
products. Plasmochemical technology of cracking is to heat hydrocarbon gases in the combined electric arc
reactor to a temperature of pyrolysis (1900-2300 K) with the formation of fine carbon and hydrogen in unified
technological process. Plasmochemical hydrogenation of solid fuels, which is the pyrolysis of coal in a
hydrogen atmosphere, provides acetylene and other unsaturated hydrocarbons (ethylene C2H4, propylene,
C3H6, C2H6 ethane, etc.) from cheap low-rank coals by their treatment with hydrogen plasma.
Plasmochemical hydrogenation of coal is a new process of direct production of acetylene and alkenes in the
gas phase, in contrast to conventional hydrogenation (liquefaction) of coal. As a result of experiments on
low-rank coal hydrogenation in plasmochemical reactor at its power of 50 kW, coal consumption of 3 kg/h,
and propane-butane mixture flow of 150 l/h gas of the following composition is obtained, mass %: C2H6=50,
C2H2=30, С2Н4=10.

Table 2. Solid fuels chemical analysis, % dry mass basis.

Solid fuel C O H N S SiO2 Fe2O3 CaO MgO K2O Na2O Al2O3

KBC 48.86 6.56 3.05 0.8 0.73 23.09 2.15 0.34 0.31 0.16 0.15 13.8
CP 75.0 0.88 15.53 0.01 5.63 1.31 0.6 0.1 0.05 0.07 0.04 0.78

779
 PBC 33.60 8.52 6.50 0.88 2.40 28.52 1.73 0.41 0.46 - - 16.98
TBC 48.58 17.85 3.64 0.78 1.14 16.64 2.13 0.88 0.67 0.01 0.01 7.67

Table 3. Main Indexes of the Solid Fuels Plasma Gasification

Solid Consumption, kg/h SPC, TAV, CO H2 N2

N P, kW Volume % XC , %
fuel fuel steam kW h/kg K
1 KBC 4.0 1.9 25 4.8 3500 41.5 55.8 2.7 94.2
2 CP 2.5 3.0 60 9.6 3850 36.2 63.1 0.7 78.6
3 PBC 7.6 2.7 60 5.83 2600 34.1 51.1 14.8 92.3
4 TBC 7.1 4.5 60 5.17 3100 45.8 49.4 4.8 93.2

Table 4. Degree of reduction of MMC

Place of sampling T, K Θ, %
Slag from pool of melt 2600-2800 8.5-44.0
Slag from the wall of the reactor 2600-2900 16.5-47.3
Slag from slag catcher 2000-2200 6.7-8.3

Fig. 6. Images of the products of propane-butane plasma pyrolisis through transmission electron microscope
(TEM) – colossal carbon nanotube metal nanoparticle intercolated.

Experiments on plasma pyrolysis (cracking) of propane-butane gas mixture were performed in the
plasmochemical reactor of 100 kW nominal power. In the experiments, propane-butane mixture flow
amounted to 300 l/min, and the plasma reactor power was 60 kW. During the experiments hydrogen and
carbon (soot) were obtained. Physico-chemical analysis of the soot samples has shown that they are
different nanocarbon structures mainly in the form of "colossal" nanotubes (Fig. 6) with high electrical
conductivity and mechanical strength.

PLASMA TECHNOLOGY OF POWER COALS PROCESSING TO CARBONIC SORBENTS

One of the newest methods of active and inexpensive sorbents production is plasma technology of power
coals processing. Fig. 7 demonstrates a pilot three plasma steps installation for carbon sorbents production
from low rank coals. The installation’s productivity is 500 kg of carbonic sorbent per hour at plasmatron power
of 50 kW, coal consumption through the PFS of 100 kg/h and 200 kg/h through the chamber of heating. The
technology is in the following. Air/pulverized coal mixture is fed to PFS 1. Produced in PFS high reactive two
component fuel inflows to chamber of heating 4, where solid heat carrier (coal dust) is fed from coal dust
hopper 3. On the solid heat carrier achievement of specified temperature crashed coal is fed from crushed
coal hopper 5 to the flow. Coal particles dimension is within 5 mm. As a result of these particles pyrolysis at
the exit of chamber of the sorbent formation 6 desired product, sorbent, is produced.

780
 Cheap, active and extended inner surface sorbent was produced from bituminous coal. Experimentally
produced sorbent on this technology was tested. It confirmed principal realizability of the technology.

Fig. 7. Scheme of the carbon sorbent producing: 1 – PFS; 2 – plasmatron; 3 – coal dust hopper; 4 –
chamber of heating; 5 – crushed coal hopper; 6 – chamber of the sorbent formation.

PLASMA TECHNOLOGY OF PULVERIZED COAL IGNITION AND THE FLAME STABILIZATION IN

ROTARY FURNACE FOR ALUMINA AGLOMERATION IN ALUMINUM INDUSTRY AND FOR OIL-FREE
CLINKER FIRING AT CEMENT PRODUCTION

This technology is developed for the biggest in Eurasia Achinsky alumina plant, which is equipped with 11
rotary kilns of 5 m in diameter and 185 m in length. Dry alumina productivity of the kiln is 102 t/h. The kiln heat
power is 108 MW. Consumption of oil for the kiln is 17700 t/year. Traditionally to stabilize pulverized coal oil
nozzle continuously operates (Fig. 8). Fuel oil flow rate is 2.5 t/h and pulverized coal consumption is 5 t/h. To
substitute fuel oil for coal and to improve ecological and economical indexes of the plant plasma technology of
oil-free ignition and pulverized coal flame stabilization in inclined rotary kiln was developed. In accordance to
this technology continuously operated fuel oil nozzle is replaced with PFS. Fig. 9 illustrates this replacement.
The consumption of coal through the PFS is 10 t/h.
Without any essential modification this technology can be applied for oil-free clinker firing to produce
cement.

PLASMA TECHNOLOGY OF OIL REFINING RESIDUUM UTILIZATION

Existing petroleum refining plants dependently on quantity of petroleum produce 4 to 6 % of oil refining
residuum. For example processing 12 million ton of raw oil per year refinery “Neftochim” (Bourgas, Bulgaria)
gets 180000 ton per year of oil refining residuum.
Traditional technology of oil refining residuum utilization is its incineration in a special rotary gas furnace.
At that conversion degree of oil refining residuum does not exceed 30-40 %, and specific power consumption
for the process reaches 100-120 kW h/t of the utilizing oil refining residuum.

781
Fig. 8. Scheme of air/coal mixture ignition in a kiln (conventional installation): fuel oil flow rate is 2.5 t/h, coal
consumption is 5 t/h.

Fig. 9. Scheme of air/coal mixture plasma ignition in a kiln: coal consumption is 10 t/h.

To improve efficiency of the process of oil refining residuum processing the plasma technology was
developed and experimentally tested. It is based on preliminary plasma thermochemical preparation of coal to
burning. Instead of gas highly reactive mixture of combustible gases and partially oxidized char particles is
incinerated in the furnace ensure high temperature needed for oil refining residuum processing. Fig. 10
demonstrates the installation for oil refining residue incineration using PFS. Having oil refining residue
processing productivity of 1 ton per hour experimental specific power consumption amounted to 33 kW h/t.
This is 3 - 4 times less than existing one.
Having 2 t/h coal consumption through the PFS and plasmatron’s power of 200 kW 12 t/h of oil refining
residuum can be processed. For processing of such quantity of oil refining residuum in accordance to
3
traditional technology flow rate of natural gas amounts 1000 m /h.

Fig. 10. Scheme of the installation for oil refining residue incineration using PFS: 1 – rotary furnace
(conventional); 2 – PFS; 3 – plasmatron; 4, 5– power-supply source; 6 – pulverised coal hopper; 7 – coal-
dust feeder; 8 – fan; 9 – heated oil refining residue nozzle.

PLASMA-CYCLONE TECHNOLOGY OF BRICKS FIRING

Conventionally the bricks are fired in special circular or tunnel kilns using natural gas combustion products. As
it is shown in Fig. 11 high temperature gas can be produced by pulverized coal incineration in a plasma-
cyclone chamber. Produced in PFS in the process of plasma thermochemical preparation of coal to burning
two component high reactive fuel is fed tangentially to cyclone. During this fuel burning in the cyclone
produced liquid slag is thrown to the cyclone wall under centrifugal force and trickles down to the slag catcher.

782
High temperature gas (1000-1200 K) is directed to the hot exhaust gas reservoir. From the reservoir hot gas
is distributed to the kiln sections to fire the bricks. Calculated specific volume thermal intensity of plasma-
3
cyclone chamber is 3 – 4 MW/m .

Fig. 11. Scheme of bricks firing in the circular kiln with plasma-cyclone chamber: 1 – circular kiln for drying
and firing of the bricks; 2 - lined duct of hot exhaust gas supply; 3 – hot exhaust gas; 4 – cyclone furnace; 5 –
liquid slag; 6 – slag catcher, 7 – slag utilizer; 8 – plasmatron; 9 – bunker of the raw coal; 10 – the coal feeder;
11 – mill; 12 – dust separator; 13 – fan.

ENERGY-SAVING ELECTROMAGNETIC TECHNOLOGY FOR MINERAL MATERIALS MELTING

At present natural basalt, which is product of ancient volcanic processes, became one of the main raw
material for mineral fiber production. Industrial practice shows that basalt has great perspectives. Wide
assortment of high quality plate heat insulators and mats is produced from basalt wool by its mechanical and
chemical processing.
Almost all firms for basalt melting mainly use gas, coke and thermal-electric furnaces. But they are large
dimension, resource-demanding and ecology negative. This is because of the branch of building materials is
large-capacity production.
Compact industrial installation based on the efficient plasma electromagnetic reactor is offered (Fig. 12).
The base of the installation for heat insulators production is electromagnetic reactor. The reactor is made of
stainless steel in the form of water cooled sections. Four graphite electrodes are placed by the centre and the
corners of equilateral triangle in parallel with axis of the reactor chamber. The working chamber of the reactor
is enclosed into electromagnet with three poles. There are windings on the poles of electromagnet; they are
inserted in a certain order, sequentially with power electrodes, and create magnetic field. The melt, produced
in the reactor, is mixing as a result of interaction between the electric current, existing between the electrodes,
and the magnetic field of the three-phase electromagnet. It guaranties material heating rate acceleration, its
more uniform heating and homogenizing of the melt.
The melt from the reactor could be delivered either to blowing device, or centrifuges for producing super
thin fibers, and for manufacturing a number of new efficient heat insulating materials. It could also come to
draw plates where the continuous fibers are being drawn; further the roving is produced out of them. The melt
could also be directed to metallic moulds and molding forms, where decorative goods, jewelry, monuments,
and construction components for engineering industry are produced. External dimensions of the electrical
reactor, including electromagnet, are ~1.51.51.5m. At the reactor productivity of 200 kg/h specific power
consumption for basalt melt making came to ~0.9 kWh/kg. At that it is observed that the emission of gases
polluting atmosphere is minimal. The power source is the adapted to the process three phase controlled
thyristor converters. It is connected up to the power transformer having industrial frequency and voltage of 0.4
– 0.6 kV.
This three phase reactor with electromagnetic mixing of the melt could be used as waste vitrification
reactor or glass-melting furnace. At that, in comparison with gas glass-melting furnace specific power
consumption for the melt getting would decrease from 2.2 to 0.8-0.85 kWh/kg.

783
 Fig. 12. Four electrode plasma reactor for minerals melting and basalt flow from the reactor.

LONG LIFE PLASMATRON

The main part of the PFS is plasmatron. Its operational capability in industrial conditions depends on its
electrodes long life. Direct current air arc plasmatron with the cathodes life significantly exceeded 500 hours
has been developed. To ensure the electrodes long life the process of hydrocarbon gas dissociation in the
electric arc discharge is used. In this method atoms and ions of carbon from near-electrode plasma deposit on
the active surface of the electrodes and form electrode carbon condensate which operates as “actual”
electrode [11]. To realize aforesaid the construction of electro arc generator of air plasma has been developed
and tested. Fig.13 gives a photo of the plasmatron.

Fig. 13. Long life DC arc plasmatron.

Propane/butane mixture is supplied to the zone of the arc conjunction to the copper water-cooled
electrodes (cathode and anode). Linked with the arc in series, the magnetic coils guaranty stabilization of the
discharge on the electrodes. During the plasmatron operation a film of the cathode condensate is formed in
accordance with the processes of propane/butane molecules dissociation and carbon atoms ionization. Arisen
from ionization positive carbon ions deposit onto the cathode surface under the influence of near-cathode

784
decline in potential and form the film of the electrode condensate. It was determined the following parameters
of the films. The film consists of carbon of 96.74-98.47 mass %, hydrogen of l.24-2.26 mass % and cuprum of
0.30-1 mass %. Interplanar spacing is 0.333, 0.207 and 0.168 nm with the intensity of the X-ray picks of 100,
-8
1 and 5 % correspondingly. Specific electrical resistance of the electrode condensate is less than 10 Ω·m.
Thus the film of the cathode condensate is current installing polycrystalline graphitic material.
On the base of the microscopy and the Raman-spectroscopy investigation, it can be concluded that the
electrode condensate is composite carbonic stuff made of carbon nanoclusters which consists mainly of single
and multi-walled carbon nanotubes and other carbonic forms including some quantity of the copper atoms
intercalated to the carbonic matrix. The electrode condensate was examined using atomic power microscope,
scanning electron microscope and transmission electron microscope (Fig. 14). The basic mass of the carbon
sample (about 80 %) is represented as film and band graphite particles collected to aggregates of various size
and density. Width of the bands varies from 40 to 160 nm. Sometimes film and band nanoparticles are
collected into stratified packages which are gathered to well-ordered nanostructures.
Created long-life plasma torch was industrially tested for plasma ignition of a pulverised coal flame in
Gusinoozersk thremal power plant, Russia. The tests confirmed efficiency of the plasmatron in conditions of
TPP.

Fig. 14. TEM images of a sample of the electrode condensate.

THE PLASMA TECHNOLOGIES MODELING

To incorporate a new technology first of all it is necessary to demonstrate its advantages in comparison with
conventional technologies. Thereto mathematical modeling and numerical experiments are indispensable to
life. To investigate numerically the aforenamed technologies and develop equipment for their realization the
following developed and standard mathematical models and computer codes are used.

A. Software code TERRA for thermodynamic calculations of multi-component heterogeneous systems

Automatized software code TERRA for thermodynamic calculations of multi-component heterogeneous
systems is used to determine the conditions of thermal equilibrium of fuel and oxidant mixture at high
temperatures. This software was created for high-temperature process computations and in contrast to
traditional thermo chemical methods of equilibrium computation that use the Gibbs energy, equilibrium
constants and Guldberg and Vaage law of acting mass, TERRA is based on the principle of maximizing
entropy for isolated thermodynamic systems in equilibrium. Fundamental lows of thermodynamics, low of
conservation of mass, energy and electric charge are methodological base of the thermodynamic
computation. Code TERRA has database of thermo chemical properties for more than 3500 chemical agents
over a temperature range of 300 to 6000 K. The database includes thermodynamic properties of organic and
mineral components of hydrocarbon fuels. Code TERRA allows calculating products compound, specific
power consumption of plasma treated material in dependence on temperature, pressure or any other
thermodynamic parameter (Fig. 15).

B. Computer code PLASMA-COAL for one-dimentional computation of the processes in PFS or plasma
reactor
Computer code PLASMA-COAL is designed for computation of the processes of moving, heating and kinetics
of thermochemical conversion of coal-oxidant mixture in PFS or plasma reactor. In the base of this code is

785
one-dimensional model which describes two-phase (coal particles and gas-oxidizer) chemically reacting flow
with an internal plasma source. The set of the ordinary differential equations includes equations for
component concentrations (chemical kinetics equations) in conjunction with equations for gas and particle
velocities and temperatures, respectively. The plasma source is accounted as a member of the equation of
energy conservation. It is internal source of heat with empirically assigned distribution of heat-evolution along
the axis of chamber with plasma source. The model is also distinguished by its detailed description of the
kinetics of chemical reactions of 116 reactions. The kinetic scheme includes the stages of coal volatile matter
evolution, char carbon, sulphur and nitrogen gasification and conversation of evolved volatile products in the
gas phase. Code PLASMA-COAL allows calculating products compound, temperature, velocities of gas and
solids, gasification degree of char carbon, power of plasma source in dependence on geometrical parameters
of the device for plasma fuel processing (Fig. 16).
10 0
N2
8 SiO2
CO

C oncentrations, mass%
C oncentrations, volume %

H 6
10 H2
O
C Al2 O3
H 2O 4

1
SiO NO
CH 4 OH AlO 2
Al CaMgSi2 O 6
O2 Ca 2Al2SiO 7
H2 S MgSiO 3
AlOH F eAl2O 4 FeSiO3 CaAl4O 7
S
0. 1 0
60 0 1 100 16 00 2 100 26 00 3 10 0 36 00 6 00 110 0 1 600 2 100 26 00 310 0 3 600

T, K T, K

Fig. 15. Concentration of gas and condensed phases species dependence on temperature of coal plasma
processing.

70

60
Gasification degree, %

50

40

30

20

10

0
0.0 0.2 0.4 0.6 0 .8 1.0 1.2 1.4 1.6 1.8 2.0

X, m

Fig. 16. Char carbon gasification degree along a plasma-fuel system.

C. Computer codes for three-dimensional computation of the furnaces

Computer codes FLOREAN (Institut für Wärme- und Brennstofftechnik, Braunschweig, Germany), CINAR ICE
(Imperial College London, UK) and PFS-CFD the base of which is KIVA-F (Rouen University, France) are
appropriated for three-dimensional computation of the furnaces of the pulverized coal fired power boilers,
including the boilers equipped with PFS. These codes are based on numerical solution of the equations of
energy and mass transfer taking into account chemical reactions. All three mathematical models are
complicated system of non-linear three-dimensional partial differential equations consisted of equations of
continuity, gas equation, equation of two-phase medium motion, equations of heat transmission, chemical

786
kinetics and diffusion for components of reacting mixture with accounts of radiation and turbulent transfer by k-
ε model. All these codes consider simplified kinetic model of combustion. Intermediate reactions and forming
of intermediate components are not taken into account. Only coal volatile release, their oxidation to carbon
oxides and burn out of carbon are considered. Components of the velocity, fields of temperatures, pressures,
concentrations of combustion products, including NOx forming can be found using the codes (Fig. 17).

Fig. 17. Temperature fields on the level of the burners and along a furnace in plasma operational regime.

To simulate gas phase of two phase reactive flow codes CINAR ICE and KIVA-F use Euler approach. But
to trace individual particles they use Lagrange. Code FLOREAN uses Euler approach to describe both gas
phase and particles. It is supposed that velocities of particles and gas phase are equal.
All the aforesaid codes were validated comparing with experiments in laboratory and full-scale conditions.
The used codes ensure competitiveness of new technologies due to decrease of expenditure onto research
and experimental development and also essentially reduce the duration the last and accelerate the process of
the technologies incorporation.

CONCLUSIONS

The described results of long-term studies of plasma-chemical technologies of pyrolysis, hydrogenation

thermochemical preparation for combustion, gasification and complex processing of solid fuels and cracking
of hydrocarbon gases and the application of these technologies to produce desired products (hydrogen,
carbon, hydrocarbon gases, synthesis gas, valuable components of the mineral mass of coal, basalt melt)
meets modern ecological and economic requirements of the power industry, metallurgy and chemical
industry.
Table 5 summarizes the results of investigations of carbonaceous raw plasma conversion.
At complex processing of coal conversion of the mineral mass requires high temperatures (2200-3100
K), which leads to higher specific energy consumption up to 2-4 kW h/kg. Thus the high degree of carbon
conversion (90-100 %) is achieved. Plasma-steam gasification provides a transition to the gas phase mainly
organic mass of coal that does not require such high temperatures, as in the complex processing, and
enables process at relatively low specific power consumption (0.5-1.5 kW h/kg) and achieving a high degree
of conversion (90-100%). Plasma-chemical hydrogenation of coal requires high temperatures (2800-3200 K),
which leads to high specific power consumption for the process (6.5-8 kW h/kg), which allows achieving high
conversion (70-100 %) for direct (one-step) receipt of acetylene and alkenes in the gas phase. To ensure a
high degree of conversion (98-100%) of hydrocarbon gas in the combined plasmochemical reactor there is
no need in such a high temperature. That allows caring the process at relatively low specific power
consumption (2.2-3.8 kW h/kg). Note that all the processes of plasmochemical processing of fuels (Table 5)

787
are characterized by extremely low concentrations of nitrogen oxides and sulfur emissions not exceeding
3
20 mg/Nm , which is much lower than when conventional use of fuels.

Table 5. The optimal range of process parameters for plasma conversion of carbonaceous raw.
3
Fuel / Т, К Specific power Fuel conversion Concentration, mg/Nm
plasma consumption, degree, %
forming gas kW∙h/kg of fuel NOx SOx
1. Plasmochemical preparation of coal for combustion (air)
1.5–2.5 800–1200 0.05–0.40 15–30 1–10 1–2
2. Complex processing of coal (water steam)
1.3–2.75 2200–3100 2–4 90–100 1–2 1
3. Plasma gasification of coal (water steam)
2.0–2.5 1600–2000 0.5–1.5 90–100 10–20 1–10
4. Plasmochemical hydrogenation of coal (hydrogen)
10 2800–3200 6.5–8 70–100 0 0
5. Plasmochemical cracking of propane-butane mixture
3
18 m /ч 1500–2500 2.2–3.8 98–100 0 0

The considered computer codes have independent scientific value. They are powerful tool for investigation
of modern technologies and processes of fuel and minerals processing.

REFERENCES

[1] E.I. Karpenko, V.G. Lukyachshenko, V.E. Messerle, A.B. Ustimenko, A.V. Yakovenko, “New
Technologies of Fuel Utilization and Mineral Raw Processing,” Combustion and Plasma Chemistry,
2004, vol. 2, N 2 pp.117-145.
[2] E.I. Karpenko, V.E. Messerle, A.B. Ustimenko, “Plasma-Aided Solid Fuel Combustion,” Proceedings of
the Combustion Institute, 2007, vol. 31, Part II, pp.3353-3360.
[3] M.A. Gorokhovski, Z. Jankoski, F.C. Lockwood, E.I. Karpenko, V.E. Messerle, A.B. Ustimenko,
“Enhancement of Pulverized Coal Combustion by Plasma Technology,” Combustion Science and
Technology, 2007, vol.179, N.10, pp.2065–2090.
[4] V.E. Messerle, A.B. Ustimenko, “Plasma-Supported Coal Combustion Modelling and Full-Scale Trials,”
Advanced Combustion and Aerothermal Technologies, N.Syred and A.Khalatov (eds.), Springer, 2007,
pp. 115-129.
[5] V.E. Messerle, A.B. Ustimenko, “Solid Fuel Plasma Gasification,” Advanced Combustion and
Aerothermal Technologies, N.Syred and A.Khalatov (eds.), Springer, 2007, pp.141-156.
[6] R.A. Kalinenko, A.A. Levitski, V.E. Messerle, L.S. Polak and A.B. Ustimenko, “Pulverized Coal Plasma
Gasification,” Plasma Chemistry and Plasma Processing, 1993, vol. 13. N. 1, pp. 141-167
[7] M. Gorokhovski, E.I. Karpenko, F.C. Lockwood, V.E. Messerle, B.G. Trusov and A.B. Ustimenko,
“Plasma Technologies for Solid Fuels: Experiment and Theory,” Journal of the Energy Institute, 2005, 78
(4), pp. 157-171.
[8] A.S. Askarova, E.I. Karpenko, Y.I. Lavrishcheva, V.E. Messerle, A.B. Ustimenko, “Plasma-Supported
Coal Combustion in Boiler Furnace,” IEEE Trans. Plasma Sci., Dec. 2007, 35 (6), pp.1607–1616.
[9] V. Galvita, V.E. Messerle, A.B. Ustimenko, “Hydrogen production by coal plasma gasification for fuel cell
technology,” International Journal of Hydrogen Energy, 2007, 32 (16), pp. 3899-3906.
[10] Messerle V.E., Ustimenko A.B. “Plasma ignition and combustion of solid fuel. (Scientific-and-
technological basics).” (in Russian). Saarbrucken, Germany: Palmarium Academic Publishing (ISBN:
978-3-8473-9845-5), 2012, 404 p. Available: http://ljubljuknigi.ru/
[11] V.I.Golish, E.I.Karpenko, V.G.Lukyashchenko, V.E.Messerle, A.B.Ustimenko, “Nanocarbon Coating of
Electrodes for Plasma Torch Life Prolongation”, Nanotec2009.it. Nanotechnology, Competitiveness and
innovation for industrial growth / Book of abstracts, March 31 – April 3, 2009, Italy, Rome, National
Research Council, pp. 141-142.

788
 EXERGY ANALYSIS OF AN ETHYLENE GYLCOL RECOVERY SYTEM USING BATCH
DISTILLATION
1* 2
Suha Orçun MERT , Erkan ERSÖZ
1
Yuzuncu yıl University, Chemical Engineering Department, Van, Turkey
2
Ege University, Chemical Engineering Department, Bornova, İzmir, Turkey
email: orcunmert@yyu.edu.tr, erkan.ersoz@ege.edu.tr

ABSTRACT
Exergy analysis of an ethylene glycol recovery system is conducted in this study in details considering a broad
system modeling. In the process ethylene glycol is tried to be recovered from the waste of a chemical plant that
includes water, glycols and some anhydrides with experimental studies. The purification is sustained by a
distillation system that operates in a batch manner which is heated by an electric heating system. Heating system,
reboiler and condenser groups are investigated for understanding the characteristics of the system and revealing
the efficiencies exergetically. The calculations are done by developing a computer program by using Visual Basic.
Overall system exergetic efficiencies with respect to the time are calculated depending on the real data and
gathered from the experiments and the exergy destruction point for the system is pointed out for sustaining a
better efficiency and cost. As result of the investigation the overall efficiency for the distillation system is found as
56.4% and the overall exergy destruction is 12 kW.

INTRODUCTION
Distillation is one of the most important methods for separating mixtures of chemical process industries and can be
used for both production and reuse applications. However it is an energy intensive process. Most of some earlier
studies are concentrated on optimum reflux ratio and column pressure to increase column efficiency. But nowadays,
reduction of energy consumption in distillation processes becomes more important subject because of limited energy
sources (Dhole, V.R., and Linnhoff, B., 1993). Energy consumption can be reduced with minimizing energy
dissipation or lost work and energy efficient distillation processes can be designed. For this reason exergy term and
by extension exergy analysis comes into prominence (Demirel, 2004).

Exergy analysis has an increasing trend in the engineering and multi-disciplined working fields for seeking better
efficiency and sustainable use of energy. Exergy can be defined as the maximum useful work that can be obtained
from a system at a given state in a given environment (Mert, 2012). By the way distillation columns can be analyzed
thermodynamically to compute the net energy gained and the amount of losses during the production in a process.
So its inefficiency on energy usage can be described with the help of this analysis. Thus exergy analysis considers a
system within its environment and gives us the true characteristic and capacity theoretically (Dincer and Rosen,
2007).

From this point of view it is obvious that exergy deal with the quality of the energy differing from the energy that only
tries to measure the quantity of the energy. Exergy analysis is used as a tool for identifying the types, locations and
magnitudes of the energy loss point in a system. So for better system efficiencies and sustainable productions it is
crucial to identify and quantify these losses (Rosen and Dincer, 2003).

In this study recovery of ethylene glycol from a waste stream solution which taken from polymer chemical industry, by
distillation is investigated exergetically. Waste solution contains mainly Ethylene Glycol, Propylene Glycol, water and
some other chemicals as insignificant level. Experimental studies carried out both laboratory scale and pilot scale
distillation setup. During the laboratory scale batch distillation process, waste solution is distillated in setup and
temperature of the boiler and condensate volume are recorded for defined time range. After experimental data is
taken, process carried out to pilot size. Pilot size distillation process is consists of glass rasching ring filled glass
column, insulated boiler and condenser as seen in Fig. 1. Heat requirement of the process provided by hot oil system
heated with electrical heater. Heater has 18 resistance cells to heat oil and each of them requires 3*2kW energy (Fig.
2).

EXERGY ANALYSIS
Exergy analysis depends on the second law of thermodynamics with considering the entropy. So differing from the
energy, exergy can be destroyed and non-conserved. For the calculation of the systems exergy both the
surroundings and the system must be defined. The surrounding is defined as the reference environment and all the
exergy calculations are done depending on these conditions. The temperature, pressure and composition of the
reference environment are assumed to be constant and specified. The reference environment may be the
atmosphere, earth crust or the ocean. The reference environment may also be thought as the thermal reservoir.

789
There are various components of the exergy exists. Depending on the system definition and characteristic some of
these components should have been may excluded from the investigation.

Fig.1. Pilot size distillation column

Fig.2. Electrical heater and its control panel.

Most common known components are physical, kinetic, chemical, potential, nuclear, magnetic, electrical, and surface
tension forms of exergy. In this study only physical and chemical exergies of the streams are taken into account.

Ex  Ex phy  Exch (1)

The physical exergy is the result of the transformation of the systems conditions to the reference states defined by P0
and T0 by physical processes only involving thermal interaction with the environment.

Ex phy  U  U o   P0 V  V0   T0 S  S0  (2)
Ex phy  H  H 0   T0 S  S0  (3)

where 0 refers for the reference environment.

790
The chemical exergy of a stream of any system is the substances calorific value. In a brief way it is defined as the
energy results from the difference in the substance composition with respect to the common components of this
substance in the environment. Chemical exergy is evaluated by bringing the pure component in chemical equilibrium
with the environment. So the departure of the chemical composition of the material from the reference condition
increases the chemical exergy of it.

 P
Ex ch  RT0 ln  
 P0  (4)

The chemical exergy of the substances is calculated with respect to its reference environment,

Ex ch  G   v j Ex 0,ch , j
j (5)

where G denotes the change in the Gibbs energy of formation, v denotes the stoichiometric coefficient and Ex0,ch,j
denotes the standard chemical exergy of the substance.
Standard molar chemical exergies of some substances can be found in the literature (Kotas, 1985). By using the
standard chemical exergy values from the table, the chemical exergies may be defined as follows;

N N
Ex ch   xi Ex ch,i  RT0  xi ln xi
i 1 i 1 (6)

where x is the mole fraction of the substance.

Second-law based exergy efficiency of a process may be defined as exergy recovered (exergy output) to the exergy
input. Exergy output can be defined as the desired exergy output or useful exergy output. (Mert, 2012)

Ex out

Ex in (7)

SYSTEM DESCRIPTION AND EXERGY ANALYSIS

The investigated system is composed of two stages. First the distillation operation is studied in the laboratory for the
investigation of the true system characteristics of separation operation and then with the gathered knowledge of
experimental study and the theoretical investigation a pilot operation is applied using a batch distillation system.

The experimental studies are carried out on an electrically heated flask which is connected to a glass column used for
the operation of distillation that is filled with rasching rings of sized 1 cm. The system is operated in a batch manner
for replicating the pilot system and the temperature profiles and components fractions in the processes while the time
passes and separations continues are observed and necessary calculations are done for sustaining the system
characteristics.

The scheme of the system investigated for exergy analysis is given in Fig. 3.

4
Condenser

5
Column

2
3
Electricity - 1
Reboiler

791
Fig. 3. The scheme of the experimental setup

The mixture that is tried to be recovered is composed of glycols and water wit trace amount of anhydrides. This
mixture is a waste solution from a polymer chemical plant. Because of the values of the glycols that is used various
fields for heat transfer fluid applications to raw material for various polymer production processes.

During the separation process the volatile components are evaporated rapidly in the early stages of the batch
distillation operation than the less volatile components fractions in the distillate is increases. According to this principle
the water is removed from the system by this evaporation and the remaining mixture became ricer from glycols point
of view. But because of the various polymer residuals and other chemicals this mixture is continued to the distillation
operation and most of the glycol mixture is taken from the top of the columns as distillate and remaining residual
became a dense chemical waste.

Table 1. The weight fractions of the components and boiling points

Component Weight Percent Boiling Point [°C]
Ethylene Glycol 18.00 197.3
Propylene glycol 61.00 188.2
Water 20.00 100
Others 1.00 -

300
Conderser
250
Reboiler
Temperature [°C]

200

150

100

50

0
0 20 40
60 80 100 120 140
Time [m]
Fig. 4. The temperature profile for the system

During the distillation operation the temperatures are changing both for condenser and reboiler since the fractions of
the components in the tank and distillate is changing (Fig. 4). As the more volatile component’s fraction decreases in
the tank the temperatures are increase for both condenser and reboiler.

792
 100
Water
90
Ethyelene Glycol
80
Propylene Gylcol
70
Weight Fraction [%]

60
50
40
30
20
10
0
0 20 40 60 80 100 120 140
Time [m]
Fig. 5. The components present in the distillate during the distillation process
The components mass fraction in the distillate during the distillation operation is given in the Fig. 5 with respect to the
time. As it is seen the fraction of the water (more volatile component) decreases while the other components fractions
are increasing as the temperature increases and time passes.

Stream outlet

Stream inlet System

Stream outlet

Stream inlet
Fig. 6. Flows in the system

A sample system is shown in Fig. 6 for exergy calculations. In this study every streams exergy is calculated with
respect to the time passes in distillation operation and exergetic efficiencies and exergy destruction is calculated
according to the exergy balance.

(8)

RESULTS AND DISCUSSION

The results for the investigation of the distillation operation in a batch manner is tabulated and shown in graphs with
respect to the time passes in the operation. For every step of the time increment the evaporated amount of mixture is
calculated and depending on the electricity input the exergy balance is applied. Table 2 shows the values for
exergies, and efficiencies with respect to the time spent in distillation operation.

Table 2. Exergy values in the separation operation

Time [m] Energy Input [kW] Tank Exergy [kW] Distillate Exergy [kW] Efficiency [%]
0.00 19.32 27.32 0.00 99.87
6.00 16.98 27.29 0.00 99.20
37.00 6.54 27.07 0.00 99.78
43.00 15.11 27.01 0.06 100.01
47.00 13.20 26.95 0.14 99.92
55.00 11.80 26.79 0.29 99.09
64.00 15.95 26.26 0.77 99.38

793
 74.00 16.03 25.32 1.32 96.17
76.00 21.19 24.97 1.68 98.20
82.00 17.67 22.84 2.43 97.12
86.00 28.67 21.06 3.08 87.96
90.00 32.69 20.52 3.25 89.50
103.00 33.76 15.12 5.36 96.63
116.00 29.67 9.25 8.09 52.23
128.00 39.48 5.58 9.97 21.41

Fig. 7 shows the variation of the exergies of tank and distillate with respect to the time as well as the exergy
efficiency. It is seen that as the time passes the exergy in the tank decreases as expected since some portion of the
mixture is evaporates exergy is transferred via mass transfer but besides this depending on the exergy destruction a
large portion of the exergy is also cannot be used.

When the exergy efficiency is investigated it is seen that most of the exergy destruction is occurred in the late steps of
the separation operation where the temperatures are at the highest values. It is also reasonable since the heat
transfer operations at high temperature differences and depending on the evaporation operation the exergy
destruction rate increases accordingly.

30
100,0

25
"Tank Exergy" 80,0
"Distilate Exergy"
20

Efficiency [%]
Exergy Efficiency
Exergy [kW]

60,0
15

40,0
10

5 20,0

0 -
0 20 40 60 80 100 120 140
Time [m]
Fig. 7. The variation of the exergies and efficiencies in the separation operation

In Table 3 the overall results for the operation is calculated and tabulated. This values is the representation of the total
energy input to the system by electricity and exergy by mass. The efficiency is around 60% values which is
considerable high when a thermal system is investigated. This high values are represents that of the benefits of
insulation and usage of electricity instead of fuel oil, coal or natural gas that includes a combustion process.

Table 3. Overall results

Total Exergy Destruction [kW] 12.03
Total Exergy Efficiency [%] 56.39

CONCLUSION
In this study the exergy analysis of a batch distillation operation of a mixture of water and glycol compounds are
investigated. The investigation is based on the time intervals and the exergy values of distillate and the residuals in
the tank are calculated as well as the efficiencies and overall exergy destruction.

It is seen that the efficiency values are higher in the early steps of the operation where the temperatures are low with
respect to the late stages of operation where the temperatures are reaches up to the 560 K.

The overall efficiency is found as 56.4% and the overall exergy destruction is 12 kW. The destructed exergy mostly
arises at the reboiler section of the system where the heat transfer and evaporation operation takes place. Following
to the reboiler the other mechanism of the exergy destruction is the condenser where the distillate is condenses.

Following to the exergy analysis and exergoeconomic analysis and an optimization study will increase the depth of
investigation about the system.

794
REFERENCES
Dhole, V.R., and Linnhoff, B. 1993. Distillation Column Targets, Computer and Chemical Engineering, 17(5-6): 549 –
560.

Demirel, Y. 2004. Thermodynamic Analysis of Separation Systems, Separation Science and Technology, 39(16):
3897 – 3942.

Mert, S.O., 2012. Exergetic Based Multi-Objective Optimization of Selected Fuel Cell Systems, PhD thesis, Ege
University.

Dinçer, İ. and Rosen, M., 2007. Exergy: energy, environment, and sustainable development, 1:11-14

Rosen, M., Dincer, I., 2003. Exergy-Cost-Energy-Mass Analysis of Thermal Systems and Processes, Energy
Conversion And Management 44; 1633-1651

Kotas, T. J., 1985. The Exergy Method of Thermal Analysis, Anchor Brendon Ltd., Tiptree, Essex,Great Britain

795
 INVESTIGATION OF EXERGY RATIOS OF A SOLAR POND AT VARIOUS REFERENCE
TEMPERATURES
1 2
Mehmet Karakilcik and Ibrahim Dincer
1
Department of Physics, Faculty of Sciences and Letters, University of Cukurova, Adana 01330, Turkey
E-mail: kkilcik@cu.edu.tr ,
2
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology (UOIT) 2000 Simcoe Street
North, Oshawa, Ontario L1H 7K4, Canada,
E-mail: Ibrahim.Dincer@uoit.ca

ABSTRACT
In this paper, we present an experimental and theoretical investigation for exergetic assessment of a
rectangular solar pond (with dimensions of 2 m × 2 m × 1.5 m) at various reference temperatures. The
experimental solar pond was designed and built at Cukurova University in Adana, Turkey. The system was filled
with salty water to form three zones (e.g., Upper Convective Zone (UCZ), Non-Convective Zone (NCZ) and
Low Convective Zone (LCZ)) accordingly. A data acquisition device was used to measure the temperatures
hourly at various locations in the pond. An exergy model is developed to study the exergetic performance of the
pond for its three zones in terms of exergy efficiencies. Lowest and highest reference environment
temperatures for various days in the months are considered. The results show that the efficiency ratios of the
LCZ are found to be higher than the corresponding the NCZ and also the UCZ.

Keywords: Rectangular solar pond, solar energy, energy and exergy efficiency, exergy ratios

INTRODUCTION
The sun is the largest source of solar energy which is renewable, clean, abundantly and easily available
everywhere in the world and also can easily be converted to thermal energy by using the renewable energy
systems. These systems generally absorb the solar radiation falling on them to collect and store as thermal
energy. The thermal energy is used for cooling, water and space heating, and power production using air,
water, steam or thermal fluids to transfer the heat to the location where it is used. For these reasons the
renewable energy technologies are improved day by day. It is widely known that these have a significant role
in reducing the energy consumption and plays an important role in energy production and saving policies. All
of these are important factors to increase the use of the solar energy to reduce non-renewable energy
consumption, carbon footprints and improving solar energy technologies. One of these is solar pond which
is a large-scale energy collector with integral heat storage for supplying thermal energy and can be used for
various applications (e.g., process heating, water desalination, refrigeration, drying and power generation).
Solar ponds can suitably be constructed anywhere, regardless of the distance to the nearest power
outlet, as long as there is access to direct sunlight on the site. It is economical to construct them at places
where there is low cost salt and bittern, a good supply of sea water or water for filling and flushing, high solar
radiation, low cost land, etc. The maintenance of solar ponds is generally simple and of low cost, and does
not emit greenhouse gas emissions. The solar pond works on a very simple principle. The solar radiation
heats the water, and the heated water from within the pond rises and reaches the top, but loses some heat
to the atmosphere. Heat is then properly stored. The solar ponds restrict this tendency of dissolving salt in
the bottom layer of the pond and make it too heavy to rise (Saxena et al., 2013).
The solar ponds are generally consisted of three main zones as we briefly define as follows. The
upper convective zone (UCZ) is composed of low salinity brine. This zone requires continuous flushing with
fresh water or low salinity water to compensate for evaporation and flush away the rising salt to the top
surface by the natural process of salt diffusion through the non-convective zone (NCZ). Below it, the NCZ
contains a salinity gradient such that the water close to the surface is always less salty than the water below
it. The NCZ constitutes a thermally insulating layer since natural convection currents are suppressed. The
lower convective zone (LCZ) is also called as the storage zone. This zone is where the heat is collected and
stored. It is very hot and salty water, and its temperature can vary between 50–90°C (Akbarzadeh et al.,
3
2005) whose density distribution is changed between 1150-1190 kg/m (Karakilcik, 1998).Usage of solar
energy to collect and store energy in salty water zones of solar pond becomes the heart of this study for
analysis and assessment.
There have been experimental and theoretical investigations on exergetic performance analyses of
integrated and non-integrated solar ponds through exergy efficiencies (Karakilcik et al., 2008; 2012). These
studies are in fact the key motivation for the present work. In this work, a detailed investigation on exergetic

796
efficiencies is carried out through exergy analysis of the zones at higher and lower reference temperatures.
Accordingly the exergy destructions and losses as well as exergy efficiency relations are determined for each
of the system components and for the whole system. In this regard, exergetic efficiency ratios at lowest and
highest reference temperatures are studied.

EXPERIMENTAL APPARATUS AND PROCEDURE

Generally, solar ponds are divided into three zones as follows:
a) The upper convective zone (UCZ) is the first zone with the fresh water layer at the top of the pond. This
zone is fed with fresh water in order to maintain its density as close as to the density of fresh water in the
upper part to protect the cleanness of the pond and meet the lost water due to evaporation.
b) The second zone is non-convective zone (NCZ). It consists of several different salty water layers. The
layers are heavier than fresh water and will not rise or mix by natural convection. So, the body of the
salty water inner zone creates a salt gradient zone. The zone permits solar radiation to pass through, but
reduces the heat loss from the hot low convective zone (LCZ) to the cold upper convective zone (UCZ).
It creates a larger temperature gradient within the pond which becomes hotter with depth and also the
salt content increases with its depth, hence causing a salinity or density gradient. This gradient acts as a
transparent insulator permitting sunlight to reach the bottom zone but also entrapping it there. Heat
transfer through this zone is by conduction only.
c) The third zone, non-convective zone (LCZ) is called heat storage zone (HSZ) which is composed of salty
water with highest density. Considerable part of the solar energy is absorbed and stored by this bottom
region. The LCZ has highest temperature and, hence the strongest thermal interaction occurs between this
zone and the insulated bottom and side walls surrounding it.
2
In the experimental work, a rectangular solar pond with a surface area of 4 m and a depth of 1.5 m was built
as shown in Fig. 1, and the pond was then filled with salty water to form the inner zones of UCZ, NCZ and LCZ.
A data acquisition system was used to measure the hourly temperature variations in a day in these inner zones.
The bottom section and the side-walls of the pond were insulated with a glass-wool of 50 mm thickness. The
temperature measurements were taken using 16 temperature sensors, as some of which were placed into the
inner zones to measure the temperature distributions of various layers. A computerized data acquisition system
was used for data recording, monitoring and processing. The temperatures of inner zones of the pond were
measured hourly during a day by sensors with a range of −65 to +155 °C, and with a measurement accuracy of
±0.1 °C for the temperature range of 0°C to 120°C. The solar energy data measured by a pyranometer and
hourly average air and daily average weather temperatures taken from a local meteorological station were used
as input parameters for analysis and assessment. Further information on experimental system and
measurement details, as well as some thermophysical properties of materials and fluids are available
elsewhere (Karakilcik et al., 2006a, b).

Figure 1. Schematic representation of the rectangular solar pond.

Fig. 1 illustrates the three different zones of the solar pond as described above. The LCZ is high density salty
water. Solar energy is absorbed and stored by this zone. The NCZ consists of six layers of salty water at
various densities, and the last zone (UCZ) is formed by fresh water layer. The incoming solar exergy is
calculated by (Petela, 2003):

797
  4T 1  Tre  
4

Ξ so = E net 1 − re +    A UCZ
3  
Ts
 3Ts
where, Ξ so is the exergy of solar radiation reaching UCZ surface; Enet is the net incident solar radiation reaching UCZ
surface; AUCZ is the net surface area of UCZ; Tre is the reference environment temperature; and Ts is the sun’s
surface temperature taken as 6000 K. Also, AUCZ is the net upper surface area of UCZ, as the effective area that
[ ( ) ]
receives incident solar radiation and is defined as A UCZ = L w L L − δ + (I − 1) ∆x tanθ (here θ the angle of the
reflected incidence, ∆x is the thickness of each layer in UCZ and taken as 0.005 m in the calculations, and Lw and LL
are the width and length of the pond) (Karakilcik et al. 2008).

EXERGY ANALYSIS
The energetic performance considers no effect of the environment temperature within the scope of first-law
of thermodynamics. As know, the real performance is affected by many aspects of the surroundings. In this
regard, the higher the emissivity of the absorbing zones inside of the pond, the higher the efficiency. The
higher efficiency is for the largest absorbing surface of the pond if the zones do not losses heat from upper
surface and side walls or if the environment temperature is equal to reference temperature, the efficiency
can be high, whereas the practical value of heat absorbed at reference temperature is zero. In contrast to the
work, heat is driven by the temperature difference, which makes the quality aspect of heat a critical element.
Conversion of radiation energy into heat causes exergy losses during the processes.
An understanding of the relations between exergy and environment may reveal the underlying
fundamental patterns and forces affecting changes in the environment, and help researchers to deal better with
environmental damage. Exergy analysis permits many of the shortcomings of energy analysis to be overcome.
Exergy analysis acknowledges that, although energy cannot be created or destroyed, it can be degraded in
quality, eventually reaching a state in which it is in complete equilibrium with the dead state (Dincer et al.,
2012). It appears to be a potential tool for design, analysis, evaluation, and performance improvement of solar
pond systems. The effectiveness of conversion of the incident radiation into heat can be evaluated by the
exergy conversion efficiency. Again interpreting appropriately the terms in the following exergy conservation
equation can assess for a balanced system in Fig. 2. It is completed by exergy loss due to irreversibility of
components of the inner zone. Here, Fig. 2 shows each of the zones and the respective exergy flows.
Ex
erg
eti

∆Ξlost to air
cS
ola

∆Ξsolar
rE
ner
gy

∆ΞUCZ
Ξlost, UCZ
UCZ Ξlost, NCZ

NCZ
∆ΞNCZ
∆ΞUCZ,recovered
Ξlost, NCZ

LCZ
∆ΞNCZ,recovered ∆ΞLCZ

Figure 2. Exergy flows in the zones of the solar pond.

Exergy efficiency for UCZ

As shown in Fig. 2, the exergy flows in UCZ can be illustrated through exergy balance equations. We first write
the exergy balance equation for UCZ as

Ξ i,UCZ = Ξ O,UCZ (1)

798
where Ξ i,UCZ is the total exergy input to UCZ, including the exergy of solar radiation and exergy gained by UCZ,
and Ξ O,UCZ is the total exergy losses, including exergy recovered and destruction. So, we can write the exergy
efficiency equation of UCZ as

Ξ r,UCZ Ξ wa,UCZ + Ξ sw, UCZ + Ξ dest.,UCZ

ψUCZ = = 1− (2)
Ξ i,UCZ Ξ so + Ξ g,UCZ

where Ξ r,UCZ is the recovered exergy of UCZ, Ξ so is the exergy of solar radiation reaching UCZ can be
expressed (as modified from Petela, 2003), Ξ g,UCZ is the exergy gained from NCZ, Ξ wa, UCZ is the exergy loss
from UCZ to the ambient air and Ξ sw,UCZ is the exergy loss through side walls of UCZ, Ξ dest,UCZ is the exergy
destruction in UCZ.
we can now define the exergy efficiency for UCZ as follows:
   T  
 m UCZ C p,UCZ (TL-H,UCZ − TL-H, re ) − TL-H, re  ln L-H,UCZ   
  TL-H, re   
 
 

 TL-H,UCZ  
 UCZ p,sw,UCZ (TL-H,UCZ − TL-H,sw,UCZ ) − TL-H,re  ln TL-H,sw,UCZ   + Ξ dest.,UCZ 
+ m C    
ψ UCZ =1−     
    
4
 E A 1 − 4TL-H,re + 1  TL-H,re   + m C   T
 (TL-H, NCZ − TL-H,UCZ ) − TL-H,re  ln L-H, NCZ   
 net UCZ  
3  Ts  
 UCZ p, UCZ
  T  
L-H,UCZ   
3Ts  
   
(3)
where mUCZ = ρUCZ VUCZ is the mass of salty water in UCZ, and is given in Table 1; ρUCZ is the averaged density
3
distribution as seen in Fig 3, VUCZ is the average volume of the salty water in UCZ as VNCZ = 0.4 m . C p,UCZ is
the specific heat of salty water in UCZ. TL-H,re , TL-H,sw,UCZ , TL-H,UCZ and TL-H,NCZ denote lowest and highest
temperatures of reference environment, UCZ, side wall of the UCZ and NCZ, respectively. Ts is the sun’s
surface temperature taken as 6000 K, C sw,UCZ is the specific heat surrounding wall of UCZ, Enet is the net
incident solar radiation reaching UCZ surface. Also, AUCZ is the net upper surface area of UCZ that receives the
solar incident.
The exergy efficiency ratio of the UCZ at the lowest and highest temperatures can be written as
ψ
χ UCZ = ψ UCZ (4)
T, UCZ

Exergy efficiency ratios for NCZ

Based on the exergy flows in NCZ as shown in Fig. 2, and the balance equation can be written for NCZ as
Ξ i, NCZ = Ξ O, NCZ
(5)

where Ξ i, NCZ is the total exergy input to NCZ, including the recovered exergy of UCZ and the exergy gained by
NCZ, and Ξ O,NCZ is the total exergy output to NCZ, including exergy losses and destruction. So, we can write the
exergy efficiency equation of NCZ as
Ξ r,NCZ Ξ + Ξ sw,NCZ + Ξ dest.,NCZ
ψNCZ = = 1 − up,NCZ (6)
Ξ i,LCZ Ξ r,UCZ + Ξ g,NCZ

799
where Ξ r,UCZ is the exergy recovered of UCZ which is given by Eq. 2, Ξ g,NCZ is the exergy gained by NCZ,
Ξ up,NCZ is the exergy losses from NCZ, Ξ sw, NCZ is the exergy loss through side walls and Ξ dest,NCZ is the
exergy destruction in NCZ.
We can now define the exergy efficiency for NCZ as the ratio of the exergy recovered from NCZ to the
total exergy input to NCZ as follows:

   T  
 mNCZ C p,NCZ (TL-H, NCZ − TL-H, UCZ ) − TL-H. re  ln L-H, NCZ   
   T   
 L -H, UCZ 
  
 TL-H, NCZ  
 sw,NCZ Cp,sw,NCZ (TL-H,NCZ − TL-H, sw,NCZ ) − TL-H, re  ln TL-H, sw,NCZ   + Ξ dest.,NCZ 
+ m    
= 1−  
  
ψNCZ (7)
  TL-H, LCZ 
Ξ r,UCZ

( )
+ mNCZ C p,NCZ  TL-H, LCZ − TL-H, NCZ − TL-H,re  ln
 T


  L-H,NCZ 

where m NCZ = ρ NCZ VNCZ is the averaged mass of salty water in LCZ, and is given in Table 1. ρNCZ is the
3
averaged density and VNCZ is the volume of salty water in NCZ as VLCZ = 2.4 m . Cp,NCZ is the specific heat of
salty water in NCZ. TL-H,re , TL-H,NCZ , TL-H,sw,NCZ and TL-H,LCZ denote lowest and highest temperatures of
reference environment, NCZ, side wall of the NCZ and LCZ, respectively. Cp,sw,NCZ is the specific heat
surrounding wall of NCZ.
The exergy efficiency ratio of NCZ can be written as

ψ
χ NCZ = ψ NCZ (8)
T, NCZ

Exergy efficiency ratios for LCZ

The exergy flows in LCZ are clearly shown in Fig. 2 and the stored exergy equation of LCZ can be written as

(
ΔΞ st = Ξ i,LCZ − Ξ O,LCZ = Ξ r,NCZ − Ξ up,LCZ + Ξ sw,LCZ + Ξ dw,LCZ + Ξ dest.,LCZ ) (9)

where Ξ i,LCZ is the input exergy from NCZ to LCZ and is called recovered exergy ( Ξ r,NCZ ) as given in Eq. 6,
Ξ o,LCZ is the exergy output from NCZ; Ξ up,LCZ is the exergy losses from LCZ to NCZ, Ξ sw,LCZ and Ξ dw,LCZ are
the exergy losses through side and down wall and have a similar effect due to the fact that both outside walls
have the same insulating materials and are surrounded by the ambient air, Ξ dest.,LCZ is the exergy destruction
in LCZ.
Using Eq. 9, we can now define the exergy efficiency for LCZ as the ratio of the exergy stored in LCZ to
the total exergy input to LCZ which is essentially the exergy recovered from NCZ:

ψLCZ =
ΔΞ st
= 1−
(Ξ up,LCZ + Ξ sw,LCZ + Ξ dw,LCZ + Ξ dest.,LCZ )
(10)
Ξ r,NCZ Ξ r,NCZ

800
    T  
 mLCZ C p,LCZ (TL-H, LCZ − TL-H, NCZ ) − TL-H, re  ln L-H, LCZ   
   TL-H,NCZ   

   

+ m C  (TL-H, LCZ − TL-H, sw,LCZ ) − TL-H, re  ln L-H, LCZ   
T
 LCZ p,sw,LCZ   T  
 L -H, sw,LCZ  
 
    
 + mLCZ C p,dw,LCZ  (TL-H, LCZ − TL-H, dw,LCZ ) − TL-H, re  ln TL-H, dw,LCZ   + Ξ dest.,LCZ 
 T 
 L -H, LCZ

ψLCZ = 1−
    
(Ξ )
(11)
r,NCZ

where mLCZ = ρLCZ VLCZ is the averaged mass of salty water in LCZ, and is given in Table 1; ρLCZ is the
3
averaged density and VLCZ is the volume of salty water in NCZ as VLCZ = 3.2 m . Cp,LCZ is the specific heat of
salty water in LCZ. TL-H,re , TL-H,LCZ , TL-H,sw,LCZ and TL-H,dw,LCZ denote the lowest and highest temperatures of
reference, LCZ, side and down wall of the LCZ, respectively. C p,sw,LCZ and Cp,dw,LCZ is the specific heat side
and down wall of LCZ, Ts is the reference environment temperature.

The exergy efficiency ratio of the NCZ can be written as

ψ
χ LCZ = ψ LCZ (12)
T, LCZ

Table 1. Average mass of the salty water of the inner zones of the solar pond.
Month mUCZ (kg) mNCZ (kg) mLCZ (kg)
Jan. 611.52 2400.58 3828.58
Feb. 619.20 2371.65 3819.30
Mar. 615.68 2362.26 3806.59
Apr. 613.70 2350.62 3798.80
May 618.80 2353.17 3784.88
Jun. 623.40 2357.26 3778.07
Jul. 612.24 2342.46 3769.62
Aug. 614.23 2335.09 3766.30
Sep. 619.10 2341.85 3776.87
Oct. 618.60 2354.64 3794.32
Nov. 613.65 2354.15 3803.03
Dec. 613.80 2345.68 3806.03

RESULTS AND DISCUSSION

We now present the results of the present model calculations for exergy efficiencies and their ratios for each
zone at lowest and highest reference environment temperatures in the experimental rectangular solar pond on
various days in the year for the inner zones and also compare these results to show how exergy is crucial for
determining true performance at various temperatures.
Fig. 3 shows the averaged experimental density variations of salty water versus the height of the
pond from the bottom in twelve months at different dates. One can conclude from the figure that the salt
density distribution in the inner zones of the solar pond becomes nearly stable which is of great significance.
It is now well known that any deterioration in salt gradient region decreases the pond’s ability to store heat
energy. The primary reason for differences during different months is likely the high temperatures in summer.
This change is mainly attributable to the thermo-physical property of the salty water, heat losses from the pond
to the air, and the absorption and reflection of incident solar radiation on the surface. The reason for the
fluctuations of the density in the zones is caused by the evaporation from the water surface, changes in the
temperature of the zones in time, salt diffusion, insufficient protection and the deterioration of salt gradient
zone. Such changes can be reduced by continuously added fresh water to the top of the pond and using a well-
functioning salt gradient protection system.
There were significant differences between the density distributions measured in January and
August, due to the temperature changes and evaporation of salty water from the pond. As expected,

801
increasing temperature decreases the density more in the summer months. As further seen in Figure 3, the
density distribution of LCZ and NCZ are decreased due to an increase in the salt diffusion as a result of the
increase in the temperature of the zones, especially in August. At same time, the density profile of UCZ
fluctuates because of the salty diffusion from bottom zone to upper zone and fresh water added to the upper
surface to compensate evaporation losses. The decreases in the ratio of the density distribution of the LCZ
and NCZ are found to be approximately 1.6% in January and 2.78% in August. The maximum decrease in
the density distribution of UCZ is found to be 1.94% in June due to the salty diffusion from bottom zone and
evaporation losses. Although the zone’s salty water is changed with fresh water (with a density of 1000
3
kg/m ) in July, increasing the ratio of the zone density is found, to be 0.44% in August, 1.24% in September
and 1.16% in October. Generally, these decreases of the LCZ and NCZ can be restored by continuously
injecting high density salty water and by using a well-working salt protection system. On the other hand,
increases in UCZ can be eliminated by adding of fresh water to the top of the pond or changing with fresh
water.
1250
LCZ NCZ UCZ
1200
Density (kg/m3)

1150

1100

1050

1000
Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.
Month
Figure 3. The averaged density variations of the inner zones of the solar pond.

The zone temperatures of the pond were measured throughout the day and averaged to find the daily
average temperature values at the respective points. It is clear that the zone temperatures vary with a day of
month in the year, depending on the reference environment temperature and incoming exergetic solar energy
on the surface (the useful part of solar energy). The temperatures of the zones generally increase with incident
solar energy per unit area of surface. There are heat losses from each zone and this is the largest in the
storage zone which affects the storage performance directly and drastically. In order to improve the
performance and increase the efficiency, we should minimize the losses appropriately.
The temperature distributions are varied with time and effect on the performance of the pond. Thus,
lowest and highest averaged temperature distributions of the zones were measured during day time for each
month in the year. Regarding the experimental temperature distributions in Fig. 4, for LCZ, NCZ and UCZ, the
temperatures of LCZ, NCZ and UCZ are observed to be a minimum of 13.06 °C, 11.27 °C and 10.10 °C on the
1 day (in January) and a maximum of 51.92 °C, 43.76 °C and 37.84 °C on the 243 day of the year (in
st rd

August), respectively.
70
LCZ (lowest temp.)
60 NCZ (lowest temp.)
UCZ (lowest temp.)
Temperature (°C)

50

40

30

20

10

0
1 46 60 91 121 193 243 271 289 333 359
Days of the year

802
Figure 4. Lowest temperature distributions of the solar pond.

Fig. 5 shows highest temperature distributions during day time for each month in the year of the zones.
The temperatures of the LCZ, NCZ and UCZ are observed to be a minimum of 17.80°C, 15.58 °C and 14.37 °C
on the 31 day (in January) and a maximum 61.45 °C, 49.93 °C and 42.75 °C on the 231st day of the year (in
st

August), respectively. As seen in the figure the temperature profiles of the zones are increasing and decreasing
seasonally as the same of the surrounding air of the solar pond. So, the temperature profiles of the inner zones
st
are seen to be a maximum on the 231 day of the year.

70
LCZ (highest temp.)
60 NCZ (highest temp.)
UCZ (highest temp.)
Temperature (°C)

50

40

30

20

10

0
31 58 90 120 148 200 231 253 276 327 335
Days of the year
Figure 5. Highest temperature distributions of the solar pond.

Figs. 6 and 7 exhibit in a bar chart the exergetic efficiency distributions at lowest and highest
temperatures taking place in the inner zones of the solar pond and their variations based on the
experimental data measured during the different day of the year. Fig. 6 shows the exergetic efficiency
distributions at lowest temperature for each zone of the solar pond. The highest exergetic efficiencies of the
LCZ, NCZ and UCZ are calculated, using the experimental data, to be a maximum 22.55%, 55.85% and
th st
92.56% at the 204 day of the year and a minimum 4.28%, 10.47% and 15.85% at the 1 day of the year,
respectively.

35
LCZ NCZ UCZ
30
Exergetic Eff. (%)

25

20

15

10

0
1 46 60 91 121 193 243 271 289 333 359
Day of the year
Figure 6. The exergetic efficiency distributions at lowest temperature of the solar pond.

Furthermore, Fig. 7 exhibits that the exergetic efficiency distributions at high temperature for each
zone of the solar pond. Highest exergetic efficiencies of the LCZ, NCZ and UCZ are calculated, using the
st
experimental data, to be a maximum 24%, 61.15% and 104.30% at the 211 day of the year and a minimum
st
5.57%, 13.12% and 20.97% at the 31 day of the year, respectively.
As seen in the Figs. 6 and 7, the differences in the exergetic efficiencies among the zones are smaller
in cooler days than the respective warmer days of the year. As expected that the efficiencies for the LCZ are
higher than the corresponding NCZ and UCZ.

803
 35
LCZ NCZ UCZ
30
Exergetic Eff. (%)

25

20

15

10

0
31 58 90 120 148 200 231 253 276 327 335
Day of the year
Figure 7. The exergetic efficiency distributions at highest temperature of the solar pond.

We now will explain the exergetic efficiency ratios of the each zone at lowest and highest reference
environment temperatures taking place in the inner zones of the solar pond and their variations based on the
experimental data measured during the different months of the year.
Figs. 8, 9 and 10 show the minimum and maximum exergetic efficiency ratios at lowest and highest
reference environment temperature of the LCZ, NCZ and UCZ. As seen in Fig. 8, the minimum and maximum
exergetic efficiency ratios of LCZ are calculated by using Eq. 12, using the experimental data and found to be a
minimum 5.27% at lowest and 5.20% at highest reference environment temperatures in January, and to be a
maximum 14.10% at lowest and 14.51% at highest reference temperatures in August, respectively. There are
no significant variations between the ratios at lowest and highest references environment temperatures of
LCZ. The exergy efficiency ratios are almost same in cooler months from December to March, and very
small in warmer months from July to November.
20
Lowest Temp.
18
Highest Temp.
Exergetic Eff. Ratio (%)

16
14
12
10
8
6
4
2
0
Jan. Feb. Mar. Apr. May Jul. Aug. Sep. Oct. Nov. Dec.
Month
Figure 8. The exergetic efficiency ratios of LCZ.

Fig. 9 shows the minimum and maximum exergetic efficiency ratios of NCZ. The ratios are calculated
by using Eq. 8, using the experimental data and found to be minimum 1.40% at lowest and 2.97% at highest
reference environment temperatures in January, and to be maximum 16.07% and 15.33% at lowest and
reference environment temperatures in August, respectively. As one can see the differences of the exergetic
efficiency ratios are small in warmer day than cooler days of the year. Although the ratios in winter season at
highest reference environment temperatures are higher than at lowest reference environment temperatures,
after in summer season are lower than at lowest reference environment temperature. There are small
variations between the exergetic efficiency ratios of NCZ but those are significantly effect on the performance
of the solar pond.

804
 20
18 Lowest Temp.
Highest Temp.
16
Exergetic Eff. Ratio (%)

14
12
10
8
6
4
2
0
Jan. Feb. Mar. Apr. May Jul. Aug. Sep. Oct. Nov. Dec.
Month
Figure 9. The exergetic efficiency ratios of NCZ.

Fig. 10 shows the minimum and maximum exergetic efficiency ratios of UCZ. The ratios are calculated
by using Eq. 4, using the experimental data and found to be minimum 3.23% at lowest and 3.98% at highest
reference environment temperatures in January, and to be maximum 15.52% at lowest and 14.87% at highest
reference environment temperatures in August, respectively. There are significant variations between the
exergetic efficiency ratios at lowest and highest references environment temperatures of UCZ. As one can
see the differences of the exergetic efficiency ratios between winter and summer seasons are changed with
at lowest and highest reference environment temperatures. The ratios at the highest reference environment
temperatures of UCZ as NCZ are higher in winter season than warmer season. But, the ratios at the lowest
reference environment temperatures of UCZ are higher in summer season than winter season of the year.
The variation of the ratios at lowest and highest reference environment temperatures between winter and
summer are especially created by ambient air temperature, humidity and salt diffusion from bottom layer to
upper layer. The upper layer (UCZ) is homogeneous and convective, where the density of saline is close to
fresh water. In the middle zone layers (NCZ) saline density increases in depth, thereby natural convection is
stopped. In this layer, mass or thermal energy is transported only by molecular diffusion that is a very low
process.
18
Lowest Temp.
16
Highest Temp.
Exergetic Eff. Ratio (%)

14
12
10
8
6
4
2
0
Jan. Feb. Mar. Apr. May Jul. Aug. Sep. Oct. Nov. Dec.
Month
Figure 10. The exergetic efficiency ratios of UCZ.

As seen in the Figs. 8-10, the fluctuations in the zones between of the ratios are significantly effect on the
performance of the solar pond. As a result, the inner zones of the pond store more exergy in August than in
January due to the considerable temperature differences between the zones. The exergy destruction and
losses significantly affect the performance of the pond and should be minimized to increase the system
efficiency. Furthermore, the advantages of exergy analysis of such systems for design, analysis and
performance improvement purposes are that it helps achieve the goal of more efficient energy resource
utilization, enabling locations, types and true magnitudes of wastes and losses to be determined, and

805
revealing whether or not and how much it is possible to design more efficient energy systems by reducing
the inefficiencies in the process/system.

CONCLUSIONS
In this paper, we have studied the performance analysis of a solar pond through exergy efficiency analysis. The
exergy efficiencies calculated for each zone at lowest and highest temperatures using the experimental data
are compared, and the corresponding minimum and maximum exergetic efficiencies and exergetic efficiency
ratios are obtained. As expected, the exergy efficiencies and ratios appear to be smaller at lowest temperatures
than the ones at highest temperatures. In this regard, the ratios are different at lowest and highest reference
environment temperatures in the same month for each zone of the pond due to the exergy destructions in the
zones and losses to the surroundings. Furthermore the exergy analysis shows that it is important to determine
the true magnitudes of the exergy efficiencies by using ratios for performance improvement studies.

Acknowledgement
The authors acknowledge the support provided by University of Cukurova (Grant Nos: FEF 2004 BAP4 and
FEF 2010 BAP5).

NOMENCLATURE
2
A surface area, m
Cp specific heat, Jkg−1k−1
E total solar energy reaching to the pond, J
HSZ heat storage zone
I number of the layers
LCZ lower convective zone
m mass
NCZ non-convective zone
T temperature, °C
UCZ upper convective zone
3
V volume, m

Greek Letters
Ξ exergy efficiency
δ thickness where the long wave solar energy is absorbed
∆x thickness of horizontal layers, m
θ angle of the reflected incidence, °
ρ
-3
density, kgm
ψ exergy efficiency, J
χ exergetic efficiency ratio

Subscripts
dest destruction
dw down wall
g gained
i input
L length
L-H lowest and highest
net net irradiation
o output
r recovered
re reference environment
s sun
so solar
st stored
sw side wall
T total
up upper
w width
wa water to air

806
REFERENCES
Akbarzadeh, A., Andrews, J. 2005. Integration of solar ponds in salinity mitigation schemes to produce low
grade heat for industrial process heating, desalination and power. ISES–2005 Solar World Congress, August
6–12, 2005, Orlando, Florida, USA.

Dincer, I. and Rosen, M. A. 2012. Exergy: energy, environment and sustainable development and energy
nd
analyses. 2 ed. Elsevier, 58.

Karakilcik, M. 1998. Determination of the performance of an insulated prototype solar pond, Ph.D. Thesis,
University of Cukurova, Adana, Turkey (in Turkish).

Karakilcik, M., Dincer, I. and Rosen, M.A. 2006a. Performance investigation of a solar pond, Appl. Thermal
Engrg. 26:727–735.

Karakilcik, M., Kıymaç, K. and Dincer, I. 2006b. Experimental and theoretical distributions in a solar pond,
Int. J. Heat Mass Transfer, 49:825–835.

Karakilcik, M. and Dincer, I. 2008. Exergetic performance analysis of a solar pond, Int. Journal of Thermal
Sciences, 47:93–102.

Karakilcik, M., Bozkurt, I. and Dincer, I. 2012. Dynamic Exergetic Performance Assessment of an Integrated
Solar Pond. Int. J. of Exergy 12:70–86.

Petela, R. 2003. Exergy of undiluted thermal radiation. Solar Energy 74:469–488.

Saxena, A. and Goel, V. 2013. A technical note on fabrication and thermal performance studies of a solar
pond model, Journal of Renewable Energy, 2013:1–5.

807
ENERGY AND EXERGY ANALYSIS OF AN R134A AUTOMOTIVE HEAT
PUMP SYSTEM FOR VARIOUS HEAT SOURCES IN COMPARISON
WITH BASELINE HEATING SYSTEM

1,* 2 3 1 4
Murat Hosoz , Mehmet Direk , Kadri S. Yigit , Mustafa Canakci , Ali Turkcan ,
4
Ertan Alptekin ,

1
Department of Automotive Engineering, Kocaeli University, Kocaeli, 41380, Turkey
2
Yalova Community School, Yalova University, Yalova, 77100, Turkey
3
Department of Mechanical Engineering, Kocaeli University, Kocaeli, 41380, Turkey
4
Department of Mechanical Education, Kocaeli University, Kocaeli, 41380, Turkey

* Corresponding author. Tel: +90 262 3032279, Fax: +90 262 3032203,
E-mail:mhosoz@kocaeli.edu.tr

ABSTRACT

Performance of an automotive heat pump (AHP) system using R134a and driven
by a diesel engine has been evaluated in this study. For this purpose, an
experimental AHP system, capable of providing a conditioned air stream by
utilizing the heat absorbed from the ambient air, engine coolant and exhaust gas,
was developed. The experimental system was equipped with instruments for
measuring engine torque and speed, refrigerant and coolant mass flow rates,
refrigerant and air temperatures as well as refrigerant pressures. The system was
tested by varying the engine speed, engine load and air temperatures at the inlets
of the indoor and outdoor coils. Using experimental data, an energy analysis of the
system was performed, and its performance parameters for each heat source
were evaluated for transient and steady-state regimes. Then, the performance of
the AHP system for each source was compared with that of the system using
other heat sources and with that of the baseline heating system. The investigated
performance parameters include air temperature at the outlet of the indoor coil,
heating capacity, coefficient of performance, and exergy destruction in the
components of the AHP system. The exergy destruction rate in the AHP with
engine coolant is higher than those in the AHP with ambient air and with exhaust
gas mainly because of the greater refrigerant mass flow rate and heating capacity.

Keywords: automotive heat pump, R134a, exergy

808
INTRODUCTION
Under cold weather conditions, comfort heating of the passenger compartments of
vehicles with an internal combustion engine is usually performed by utilizing the
engines waste heat. However, this coolant-based heating system cannot provide an
appropriate thermal comfort in the compartment until the coolant temperature rises to a
certain value. This problem is more critical for the vehicles employing high-efficiency
diesel engines due to the lack of sufficient waste heat within an acceptable duration of
operation after the engine is started (Wienbolt and Augenstein, 2003). With the
intention of obtaining thermal comfort rapidly, some vehicles utilize heaters using fuel or
electricity. However, these systems contain several disadvantages such as high initial
and operating costs, low efficiency and leading to air pollution as well as global
warming. On the other hand, the problem of insufficient heating can be solved by
adding some low cost components to the present air conditioning system of the vehicle
to operate it as a heat pump. The automotive heat pump (AHP) system can heat the
passenger compartment individually, or it can support the present heating system of the
vehicle (Meyer et al., 2004).

There have been very few research studies on AHP systems due to its competitive
nature. Domitrovic et al. (1997) simulated the steady state cooling and heating
operation of an automotive air conditioning (AAC) and AHP system using R12 and
R134a, and determined the change of the cooling and heating capacities, COP and
power consumption with ambient temperature at a fixed compressor speed. They found
that R134a and R12 yield comparable results, while the heating capacity of the AHP
system was insufficient in both refrigerant cases. Antonijevic and Heckt (2004)
developed and evaluated the performance an R134a AHP system, which was
employed as a supplementary heating system. They carried out the tests at very low
ambient temperatures and compared the performance of the AHP system with that of
other supplemental heating systems.

Rongstam and Mingrino (2003) evaluated the performance of an R134a AHP system
using engine coolant as a heat source, and compared it with the performance of a
coolant-based heating system at –10ºC ambient temperature. They found that, during a
drive cycle operation, the average cabin temperature provided by the heat pump
system at the 10th minute of the tests was 8ºC higher than those provided by the
baseline heating system. Scherer et al. (2003) reported an on-vehicle performance
comparison of an R152a and R134a AHP system using engine coolant as a heat
source. They presented the air temperatures at several locations inside the passenger
compartment as a function of time, and found that both refrigerants yield almost
identical performances and heating capacities. Hosoz and Direk (2006) evaluated the
performance of an air-to-air R134a AHP system, and compared its performance with
the performance of an air conditioning system. They observed that the AHP system
using ambient air as a heat source could not meet the heating requirement of the
compartment when ambient temperature was extremely low. Direk and Hosoz (2008)
carried out an energy and exergy analysis of an R134a AHP system using ambient air
as a heat source. They observed that the heat exchangers of the system are
responsible for most of the exergy destruction. Tomoichiro et al. (2005) studied on the

809
experimental performance of an AHP system using CO2 as a refrigerant. They found
that the performance of the heat pump system using CO2 is equal to or exceeding that
of the system using R134a. Kim et al. (2007) investigated the heating performance of
an AHP system with CO2. They performed transient and steady state tests under
various operating conditions, finding that the use of their system improved heating
capacity compared to the baseline heating system. Cho et al. (2009) reported on the
heating performance characteristics of a coolant source heat pump using the waste
heat of electric devices for an electric bus and suggested a heat pump for an electric
bus to be a heating system of over 23.0 kW for cold weather conditions.

In this study, the performance parameters of an experimental AHP system for the
cases of using the heat from ambient air, engine coolant and exhaust gas were
evaluated, and compared with each other along with the performance of the baseline
heating system. The investigated performance parameters are heating performance,
compressor power, coefficient of performance, exergy destruction in the components
and total exergy destruction in the AHP system.

DESCRIPTION OF THE EXPERIMENTAL SETUP

As schematically shown in Fig. 1, the experimental AHP system is usually made from
original components of a compact size AAC system. It employs a seven-cylinder fixed-
capacity swash-plate compressor, a parallel-flow micro-channel outdoor coil, a
laminated type indoor coil, two thermostatic expansion valves, a reversing valve to
operate the system in reverse direction in the heat pump operations, a brazed plate
heat exchanger between the engine coolant and the refrigerant to serve as an
evaporator and another plate heat exchanger to extract heat from the exhaust gas. All
lines in the refrigeration circuit of the system were made from copper tubing, and
insulated by elastomeric material. The indoor and outdoor coils were inserted into
separate air ducts of 1.0 m length. In order to provide the required air streams in the air
ducts, a centrifugal fan and an axial fan were placed at the entrances of the indoor and
outdoor air ducts, respectively. These ducts also contain electric heaters located
upstream of the indoor and outdoor coils. The indoor and outdoor coil electric heaters
can be controlled between 0–2 kW and 0–6 kW, respectively, to provide the required air
temperatures at the inlets of the related coils. The refrigeration circuit was charged with
1600 g of R134a.

Table 1 Characteristics of the instrumentation

Measured variable Instrument Range Uncertainty
Temperature Type K thermocouple –50–500 ºC ± % 0.3
Pressure Pressure transmitter 0–25 bar ± % 0.2
-1
Air flow rate Anemometer 0.1–15 m s ±%3
-1
Refrigerant mass flow rate Coriolis flow meter 0–350 kg h ± % 0.1
Compressor speed Digital tachometer 10–100000 rpm ±%2
Torque Hydraulic 5–750 Nm ±%2
dynamometer

810
 Fig. 1. Schematic illustration of the experimental AHP system using various heat sources

In order to gather data for the performance evaluation of the experimental AHP system,
some mechanical measurements were conducted on the system. The employed
instruments and their locations are depicted in Fig. 1. The characteristics of the
instrumentation can be seen in Table 1. The refrigerant mass flow rate was measured
by a Coriolis mass flow meter. In order to ensure that the refrigerant is in a liquid phase
before entering the mass flow meter, a liquid receiver with a volume of 1.1 litres was
placed between the condenser and filter/drier. Additionally, in order to observe if the
refrigerant stream contains moisture, a sight glass was located just upstream of the
mass flow meter. The temperatures of the refrigerant and air at the inlet and outlet of
each component were measured by K-type thermocouples. The refrigerant pressures
at the inlet and outlet of the compressor were monitored by both Bourdon gauges and
pressure transmitters.

The AHP system was driven by a Fiat Doblo JTD diesel engine with a cylinder volume
of 1.9 litres and a maximum power of 77 kW at 4000 rpm. The engine torque and speed
were measured by means of a hydraulic dynamometer (Baturalp-Taylan BT-190 FR)
with a maximum measuring power of 100 kW, a maximum torque of 750 Nm and a
maximum speed of 6000 rpm. The fuel consumption was measured using an electronic
scale and a chronometer. The values of the measured variables were usually acquired
through a data acquisition system and recorded on a computer. The data acquisition
system has 16 bit - 200 KHz frequency with a 56 channel thermocouple input module
and an 8 channel transducer interface module.

811
Fig. 1 also illustrates the refrigerant flow paths in the experimental heat pump system
for the cases of using ambient air or exhaust gas as a heat source. In order to perform
the heat pump operation, the reversing valve is energized. Then, the reversing valve
directs the high temperature superheated vapour refrigerant discharged from the
compressor to the indoor coil (condenser). The refrigerant passing through the indoor
coil rejects heat to the indoor air stream, thus providing a warm air stream for the
potential passenger compartment. After rejecting heat to the indoor air stream, the
refrigerant condenses and leaves the indoor coil as subcooled liquid. In heat pump
operations, thermostatic expansion valve #1 (TXV1) is bypassed. Then, the refrigerant
flows through valves V1 and V2 and reaches the receiver tank, which keeps the
unrequired refrigerant in it when the thermostatic expansion valve decreases the
refrigerant flow rate at low cooling loads. After passing through the filter, sight glass and
Coriolis flow meter, the refrigerant reaches TXV2 located at the inlet of the outdoor coil.
Since valves V5 and V8 are closed, the refrigerant passes through TXV2, which
reduces the pressure, and thereby the temperature, of the liquid refrigerant. TXV2 also
controls the refrigerant mass flow rate circulating through the circuit so that a constant
superheat at the outlet of the outdoor coil is maintained under all operating conditions.
Next, it enters the outdoor coil, where it absorbs heat taken from the outdoor air stream,
and leaves the outdoor coil as low pressure superheated vapour. After leaving the
outdoor coil, the refrigerant passes through valve V6, and enters the reversing valve.

The refrigerant flow paths in the experimental AHP system using exhaust gas as a heat
source are the same as those shown in Fig. 1. In this case, the outdoor air stream is
heated by the exhaust gas in a heat exchanger located upstream of the outdoor coil,
and then the evaporating refrigerant absorbs its heat in the outdoor coil. Note that in
this operation, the bulb of TXV2 is located at the outlet of the outdoor coil serving as an
evaporator to sense the superheat of the leaving refrigerant and regulate the refrigerant
mass flow rate circulating through the circuit properly. Similar to the operations in the
previous cases, in the AHP system using engine coolant, the liquid refrigerant enters
TXV2 located at the inlet of the outdoor coil. Since valve V4 is closed, it flows through
valve V8 and enters the heat exchanger serving as an evaporator. After absorbing heat
from the engine coolant, the refrigerant evaporates and leaves the heat exchanger as
superheated vapour. Then, through valve V7, the refrigerant enters the reversing valve,
and the operation goes on similar to the previous cases.

THERMODYNAMIC ANALYSIS
The performance parameters of the experimental AHP system can be evaluated by
applying the first law of thermodynamics to the system. Using this law for the indoor coil
(condenser), the heating capacity of the experimental AHP system can be evaluated
from
Q cond m
 r (hcond,in  hcond,out ) (1)

Assuming that the compressor is adiabatic, the power absorbed by the refrigerant
during the compression process can be expressed as

812
Wcomp m r (hcomp,out  hcomp,in ) (2)

Coefficient of performance of the system can be defined as

Q cond (3)
COP 
Wcomp

The general form of exergy rate balance equation given below can be utilized (Ozgener
and Hepbasli, 2007).

 T0  
 1  T Q Wcv   m in in   m out out  E xd
 j
(4)
 j 

The specific flow exergy in this equation can be evaluated from (Ozgener and Hepbasli,
2007)

  (h  h0 )  T0 (s  s0 ) (5)

where subscript “0” stands for reference (dead) state.

In the adiabatic compressor, the rate of exergy destruction can be evaluated from

 r  comp,in  comp, out 

E xd , comp  m (6)

The rate of exergy destruction in the condenser can be determined from

E xd ,cond  m r ( cond,in  cond,out )  m a,cond ( B  C ) (7)

The total flow exergy of air to be used in Eq. (7) at the locations B and C can be
calculated from (Ozgener and Hepbasli, 2007)

 a  (C p ,a  C p ,v )T0 (T / T0 ) - 1 - ln(T / T0 )  (1  6078 )( RaT0 ln( P / P0 )

(8)
(1  1.6078 ) ln(1  1.60780 ) /(1  1.6078 )
 RaT0  
 1.6078 ln( / 0 ) 

Neglecting the heat transfer, the rate of exergy destruction in the expansion valve can
be obtained from

E xd ,valve  m
 r ( valve, in  valve, out ) (9)

813
 Then, using Eq. (5) and considering that h5  h4 , we obtain

E xd ,valve  m rT0 svalve, in  svalve, out  (10)

The rate of exergy destruction in the evaporator can be evaluated from

E xd ,evap  m r ( evap,in  evap,out )  m a,evap ( E  F ) (11)

Finally, the total rate of exergy destruction in the refrigeration circuit of the system can
be determined by summing up the individual destructions, i.e.

E xd ,total  E xd , cond  E xd , evap  E xd , comp  E xd , valve (12)

RESULTS AND DISCUSSIONS

Fig. 2 shows the variations in the heating capacity versus compressor speed at the fifth
minute of the tests when the temperatures of the air streams entering the evaporator
and condenser are maintained at 5 ºC. The coolant based AHP system has the highest
heating capacity at all compressor speeds. The heating capacity at the end of the fifth
minute increases on rising the compressor speed for a fixed dynamometer load of 60
Nm. In all tests performed at 60 Nm, the coolant based AHP system has the highest
fifth minute heating capacity, which is followed by the baseline heating system, AHP
with exhaust gas and AHP with ambient air in descending order.

Fig. 2. The change of the heating capacity as a function of the compressor speed at the end of the five minute
operation period

The effects of the compressor speed on some of the performance parameters of the
AHP system at steady state conditions are shown in Figs. 3–6. Fig. 3 shows the
variations of the heating capacity versus compressor speed when the temperatures of

814
the air streams entering the evaporator and condenser are maintained at 5 ºC and at
the fixed dynamometer load of 60 Nm. It is seen that the heating capacity usually gets
higher on increasing the compressor speed. In the AHP system with the engine
coolant, the refrigerant absorbs a greater amount of heat from the coolant heat
exchanger, thereby providing considerably higher heating capacity compared with the
AHP system with ambient air or exhaust gas. On the other hand, when the engine load
is kept at 60 Nm, the baseline heating system yields higher steady-state heating
capacities than all AHP systems at all speeds. In all tests, The AHP system using
exhaust gas usually results in higher heating capacities than the AHP system using
ambient air as a heat source.

Fig. 3. The change of the heating capacity as a function of the compressor speed

Fig. 4 indicates that the COP for heating decreases by increasing the compressor
speed. Because the compressor power increases faster than the heating capacity
does, the COP for heating decreases with increasing compressor speed. The AHP
system using engine coolant provides the highest COP compared with the AHP system
using ambient air and exhaust gas.

Fig. 4. The change of the coefficient of performance for heating as a function of the compressor speed

815
The effect of the compressor speed on the distribution of the exergy destructions in the
components of the AHP system can be shown in Fig. 5 for the engine speed of 850 rpm
at 5 ºC of air inlet temperatures. It is seen that the indoor coil (condenser) usually
causes the highest exergy destruction followed by the outdoor coil (evaporator),
compressor and thermostatic expansion valve in descending order.

Fig. 5. The comparison of the exergy destructions at n=850 rpm and T=5 Nm

Fig. 6 indicates that the AHP with engine coolant has the highest exergy destruction at
1550 rpm engine speed and 60 Nm dynamometer load settings. As the heat transfer in
the outdoor and indoor coils takes place across a higher temperature difference, the
exergy destructions in these components are higher than other components.

Fig. 6. The comparison of the exergy destructions at n=1550 rpm and T=60 Nm

Fig. 7 shows the effect of the compressor speed on the rate of total exergy destruction
in the AHP system. It can be seen that the rate of total exergy destruction in the AHP
system increases on rising the compressor speed. This is mainly due to the increased
refrigerant mass flow rate and condensing pressure together with decreased

816
evaporating pressure with increasing compressor speed. Furthermore, as the
compressor speed increases, the heat transfer in the condenser (indoor coil) and
evaporator (outdoor coil) takes place across a higher temperature difference, and the
exergy destructions in these components get higher. The exergy destruction rate in the
AHP with coolant is higher than those in the AHP with ambient air and with exhaust gas
mainly because of the greater refrigerant mass flow rate and heating capacity.

Fig. 7. The change of the total exergy destruction as a function of the compressor speed

CONCLUSIONS
The performance of an experimental AHP system for the cases of using ambient air,
exhaust gas and engine coolant as a heat source has been evaluated and compared
with each other as well as with the performance of the baseline heating system.

AHP systems using different heat sources provided higher heating capacities than the
baseline heating system at the end of five minute operation period when the engine
was operated at the idling conditions (n=850 rpm and T= 5 Nm). However, only the
AHP system using engine coolant as a heat source yielded a better fifth-minute
performance than the baseline heating system at higher speeds when the engine load
was kept at 60 Nm. On the other hand, when the engine load was 60 Nm, the baseline
heating system provided higher steady-state heating capacities than all AHP systems
at all speeds. In all tests, the AHP system using exhaust gas usually resulted in higher
heating capacities than the AHP system using ambient air as a heat source. It was
determined that the AHP system with engine coolant yielded the highest COP values,
while the AHP system with ambient air yielded the lowest ones.

The exergy destructions in the components of the AHP system and the total exergy
destruction in the AHP system increased on rising compressor speed and temperatures
of the air streams entering the evaporator and condenser. Among the AHP system
components, the condenser usually caused the highest exergy destruction, followed by

817
evaporator, compressor and expansion valve. These results reveal that the condenser
has a high potential for further improvement.

Acknowledgement

The authors would like to thank The Scientific and Technological Research Council of
Turkey (TUBITAK) for supporting this study through a Research Project (Grant No.
108M132).

NOMENCLATURE Tj instantaneous temperature, K

AAC automotive air conditioning W power, W

AHP automotive heat pump Wcv power produced in the control
COP coefficient of performance volume, W
c p ,a specific heat of air, kJ/kg K
Greek symbols
c p ,v specific heat of water vapour,  humidity ratio
kJ/kg K  specific flow exergy
E x d the rate of exergy destruction, W
h enthalpy, kJ/kg Subscripts
mass flow rate, g/s 0 reference (dead) state
m
a air
n engine speed, rpm
comp compressor
p pressure, Pa
cond condenser
Q heating capacity, W
cv control volume
Q j
time rate of heat transfer, W evap evaporator
R ideal gas constant, kJ/kg K in inlet
s entropy, kJ/kg K ind indoor
temperature, °C out outlet
T
environmental temperature outd outdoor
T0
r refrigerant
representing the dead state, K uni unit

REFERENCES

Antonijevic, D. and R. Heckt. 2004. Heat pump supplemental heating system for motor
vehicles. International Journal of Automobile Engineering Part D 218:1111-1115.

Cho, C.W., H.S. Lee, J.P. Won and M.Y. Lee. 2012. Measurement and evaluation of
heating performance of heat pump system using wasted heat of electric devices for an
electric bus. International Journal of Energies 5:658-669.

818
Direk, M. and M. Hosoz. 2008. Energy and exergy analysis of an automobile heat pump
system. International Journal of Exergy 5:556-566.

Domitrovic, R.E., V.C. Mei and F.C. Chen. 1997. Simulation of an Automotive Heat
Pump. ASHRAE Transactions 103:291-296.

Hosoz, M. and M. Direk. 2006 Performance evaluation of an integrated AAC and HP

system. Energy Conversion and Management 47:545-559.

Kim, S.C., M.S. Kim, I.C.Hwang and T. Lim. 2007. Heating performance enhancement
of a CO2 heat pump system recovering stack exhaust thermal energy in fuel cell
vehicles. International Journal of Refrigeration 30:1215-1226.

Meyer, J., G. Yangand E. Papoulis. 2004. R134a heat pump for improved passenger
comfort. SAE paper 2004-01-1379.

Ozgener, O. and A. Hepbasli. 2007. Modeling and performance evaluation of ground

source (geothermal) heat pump systems. Energy and Buildings 39:66-75.

Rongstam, J and F.A. Mingrino. 2003. Coolant – based automotive heat pump
system. C599/067/2003 VTMS 6 Conference.

Scherer, L.P., M. Ghodbane, J.A. Baker and P.S. Kadle. 2003. On-Vehicle performance
comparison of an R-152a and R-134a HP system. SAE Word Congress, Detroit,
Michigan, USA.

Tamura, T., Y. Yakumaru and F. Nishiwaki. 2005. Experimental study on automotive

cooling and heating air conditioning system using CO2 as a refrigerant. International
Journal of Refrigeration 28:1302-1307.

Wienbolt, H.W. and C.D. Augenstein. 2003. Visco heater for low consumption vehicles
SAE Word Congress, Detroit, Michigan, USA. 2003-01-0738.

819
 EXERGY AND ENERGY EVALUATION OF DISTRICT HEATING SYSTEM
a* b c a
K. Comaklı , B. Yuksel , U.Cakır , O. Comaklı
a
Atatürk Uni. Mechanical Engineering Dept. 25240 Erzurum Turkey
b
Balikesir Uni. Mechanical Engineering Dept. 10165 Balikesir Turkey
c
Bayburt Uni. Mechanical Engineering Dept. 69000 Bayburt,Turkey

e-mail: kcomakli@atauni.edu.tr

Abstract
This paper deals with the analysis of energy and exergy utilization in the district heating system (DHS) of university
campus. The mass, energy, entropy and exergy balance relations are derived and applied to DHS. The performance
characteristics of this DHS are evaluated in terms of energetic and exergetic aspects. The energy and exergy efficiency
values for the DHS are found to be 74.99 and 8.46%, respectively. The analysis exergy loss which occurs during the
transport of thermal energy to consumers indicates that this loss is large and primarily dependent on the temperature of
the hot water. The total exergy losses increase 0.75% with increasing the supply water temperature about 10 0C, which
is the case for the return water temperature. This analysis shows that this loss during heat transfer in district heating
network can be reduce, by reducing the consumption of electrical energy during the transport of hot water to the
consumer and by reducing heat loss in pipelines.

Keywords: exergy analysis, district heating

1. Introduction

The global problems, e.g. environmental pollution, population growth, global warming and increasing energy demand
have become obvious within the last decades due to the increasing industrial use of natural resources during the past
centuries [ 1]. Energy consumption is one of the important problems of modern life. Most of the energy we use up comes
from fossil fuels. We have to save energy since the limited resources are available and those fuels cause significant
environmental pollution. The energy saving is maintained by reducing the energy consumption in buildings. Energy
savings is more vital for Turkey which imports most of the energy used up. A significant part of Turkey consumes great
amount of energy for heating.

The demand for space heating contributes greatly to total energy demand in many industrialized countries. Energy
consumption for space heating in Turkey is very high (%34). Low quality fuel consumption together with the increasing
energy demands for space heating have caused very high air pollution and poor air quality on occasion during heating
period in Turkey. Buildings with low insulation thickness, increasing building construction and lack of passive solar
applications in buildings are causes of increasing fuel consumption. For heating, coal, fuel oil, natural gas, LPG, and
biomass is used.

Faster urbanization caused by industry revaluation made to emerge the idea finding a remedy to human needs from a
centre in addition to the social services like water supply, sewer system, public transportation and DH system. DH, the
utilization of centrally produced heat for space heating and domestic hot water generation, has the potential to contribute
to the eco-efficient use of energy resources in the parts of the world where space heating is needed. A district energy
system traditionally refers to centrally producing heat and/or cooling for multiple customers in a concentrated area, such
as a city center, a university campus, or a hospital complex. As district heating and cooling systems spend great
quantities of fuel at national economies level to provide heating, cooling and domestic hot water in buildings. These
systems attracted great attention in the literature [2–8].

DH is important not only because of the use of energy sources more efficiently but also being received with energy need
more regularly, sufficiently and cheaper than other ways. In DH system, heat generated in a plant is transported to the
consumers in an extensive area. Thus, it is provided to heat a region consisting of many buildings and hot water from a
centre. The DH system, transport from the heat plant by using primary pipe network, via substation, to secondary pipe
network where heat is finally distributed to consumer.

Energy and exergy analyses are performed recently to show where energy efficiencies occur within processes, donated
as energy and exergy losses. Energy and exergy losses can directly be translated to an increase in primarily fuel
consumption. Energy analysis is based on the first law of thermodynamics, which is concerned with the conversation of
energy. Exergy analysis is based on second law of thermodynamics. Many researchers propose that the thermodynamic
performance of a process is best evaluated with exergy analysis [9].

From the thermodynamics point of view, exergy is defined as the maximum amount of work, which can be produced by
a system, or a flow of matter or energy as it comes to equilibrium with a reference environment. Unlike energy, exergy is
not subject to a conservation law (except for ideal, or reversible, processes). Rather, exergy is consumed or destroyed,

820
due to irreversibilities in any real process. The exergy consumption during a process is proportional to the entropy
created due to irreversibilities associated with the process. Exergy analysis is a method that uses the conservation of
mass and conservation of energy principles together with the second law of thermodynamics for the analysis, design and
improvement of energy and other systems. The exergy method is a useful tool for furthering the goal of more efficient
energy-resource use, for it enables the locations, types, and true magnitudes of wastes and losses to be determined [10].

In this study, energy and exergy analysis in the district heating system of Atatürk university were analyzed. In the
University, heating and domestic hot water in building of supplied heat is produced by DH. This heating system operates
24 h a day, every heating season (winter) of year. Atatürk University is in the Erzurum which is one of the coldest cities
in Turkey. In one season, approximately 10000 tons fuel-oil is consumed in the heating system. Because the pipe
network is about 11988 m, which is assumed to be quite long, there have been energy and exergy losses considerably in
the heating network system revision and re-submittal.

2. System description and analysis

The system consists of heating plant, pipeline network, heat stations, and heat consumer (Fig1). The essential element
of such a system is heating plant. One other important part of the system is pipeline which enables the transport of
energy and which is the source of the heat losses. The other important part of the system is the heat stations, where the
heat is transferred from a high to a low temperature medium, resulting in decreased heat quality. In the analysis, the
pipeline network was the insulated pipes with diameter from 65 to 250 mm. The total length of the pipeline network is
11988m. The average pressure in the primary network (Ts=1750C, Tr=1100C), was 15 bar, while in the secondary
network (Tc,s=850C, Tc,r=650C), the pressure depended on the atmosphere pressure.

Flue
gases I2 I1 Ef Ehw
Exf Exhw Eev
Wf Exev

Ep.loss I5
Exp.loss

EFLUE
EXFLUE Wp1 I3 I4 Wp2

Heating Plant Buildings

Pipes
Channel
Fig.1. Illustration district heating system
Energy and exergy balance for any control volume at steady- state, with negligible kinetic and potential energy variations,
can be expressed respectively by the equations: [11]

 Q   m h  W   m h
j
j
i
i i
j
j
o
o o (1)

Ej
Q
x, j   m e  W   m e
i
i i
j
j
o
o o  I (2)

Where I is the irreversibility rate (exergy destruction rate) of the control volume under analysis and is the exergy
associated to a heat flow available at a temperature Tj, given by

821
  
E Q   T0 
x  Q j 1  
(3)
 Tj 
The specific flow exergy (e) is given by

e  (h  h0 )  T0 (s  s 0 )  e ch (4)

where the subscripts (0) stand for the restricted dead-state (ambient pressure and temperature) and superscript (ch)
stands for the standard chemical component. In fig. 2 mass flows and energy flows are identified (numbered streams).
These are: boiler (heating plant), district heating network, heating heat exchanger and hot water heat exchanger. Energy
and exergy balance for this system volume can be expressed respectively by the equations:[12]

 W W
 W    
E f f p1 p2  E flue  Ep.loss  Ehw  E ev (5)

 W  W
 W E          
E x,f f p1 p2 x,flue  Ex,p.loss  Ex,hw  Ex,ev  Itot Itot  I1  I2  I3  I4  I5 (7)

13
Hot water
heat exchanger

6
12
3 4 5 11
7
Heating water
1 heat exchanger
2
Boiler 10
9 8
Fig.2 Flow sheet diagram of system

For system, exergy efficiency , also known as “rational efficiency” has been determined according to the definitions given
in equation [13]:

exergy of product
 (8)
exergy of f uel

The expressions for the rational efficiency of boiler, heating heat exchanger, hot water heat exchanger and district
heating system are given respectively by:

 4 (e 4  e 9 )
m
 boiler  (9)
Ex,f

 10 (e11  e10 )
m
 hhex  (10)
m 5 (e 5  e 8 )

 12 (e13  e12 )
m
 whex  (11)
m 6 (e 6  e 7 )

Boilers are composed of four boilers. For analysis purpose they were considered as a one equivalent boiler. Similar to, it
were considered as a one equivalent heat exchanger.

822
3. Result

The thermodynamic properties of water and steam were obtained from steam tables. The properties of gases (air and
combustion products) were calculated employing the methodology proposed by ref.çengel [14], with polynomial
regressions for specific heat. The ideal gas model was applied for air and combustion products. The chemical exergies of
different substances (ech) were determined according to Kotas [13]. Temperature, pressure, mass flow, enthalpy, and
exergy for each flow are shown in Table 1, according to the in Fig. 2. It used to one year value of system in analysis.

Table 2 shows the evaluation result of irreversibility’s in sub-systems. As shown in table 2, the largest and second largest
losses are irreversible loss in the boiler heat exchanger (I 2) and combustion (I1), respectively. Fig 3 and Fig 4 show the
energy and exergy grasman diagrams.

Table 1. Flow identification and properties

Flo T m h e
w (K) (kg/h) (kJ/kg) (kJ/kg)
Combustion air 1 298 30481 0.0 0.0
Fuel (fuel-oil) 2 298 1548.9 42650 45458
Flue gases 3 438 32030 150.6 112.58
Supply boiling 4 443 216370 741.2 117.06
water
Supply boiling 5 440 196906 728.1 112.81
water
Supply boiling 6 440 19464 728.1 112.81
water
Return boiling 7 378 19464 457.8 39.83
water
Return boiling 8 378 196906 457.8 39.83
water
Return boiling 9 378 216370 457.8 39.83
water
Heating return 10 343 597600 294 13.04
water
Heating supply 11 363 597600 378 26.17
water
Feed water 12 286 31421 54.3 0.248
Hot water 13 328 31421 230.5 5.93

Table 2. Irreversibility’s of sub-system

Symbol Irreversibility
Subsystem %
(kW)
Combustion I1 6553 40.02
Boiler heat exchanger I2 7364 44.98
Hot water heat exchanger I3 374 2.28
Heating heat exchanger I4 1489 9.10
Radiators I5 593 3.62
Total Itot 16373 100

823
Table 3. Efficiencies of system component
Fist law Exergy effi.
efficiency (  ) ()
Boiler 92.82 23.73
Hot water heat 97.90 12.57
Heating
exchangerheat 98.10 54.60
exchanger
Overall system 74.99 8.46

Table 3 shows the evaluation result of efficiency in system component. As shown in table 3, the largest and second
largest exergy efficiency is exergy efficiency in the heating heat exchanger and boiler, respectively. Energy losses in
pipelines which are used for the distribution of heat from the heating plant to the exchanger in heating systems are very
important. Heat losses in pipelines were calculated using daily average outer temperature values. Yearly heat loss in
network was shown in table 2. Heat losses due to the network Qloss is about 8.62% of the energy provided by heating
plant, Qi.

Thickness of thermal insulation materials is the most effective factor which causes reduction of heat losses in pipes.
Heat losses decrease by increasing thickness of insulation. Especially, an increase up to a thickness of 20 cm results in
an important decrease in heat losses. It would be a decrease of about 25% in heat losses if an insulation thickness of 20
cm was used instead of 8 cm in pipes.

Ep.loss
8.69 % Wp2
0.28 %

Ef
98.31 %

Wf Eev
1.06 % 74.99 %

Wp1 Eflue Ehw

0.35 % 8.12 % 8.21 %

Fig.3. Energy flow (percentage) diagram of DH system.

Exp.loss
2.91 %

I1 I2 Wp2
32.97 % 37.05 % 0.27 %
I5
2.98 %
Exf
98.41 %
Exev
8.46 %
Wf I4
0.99 % 7.49 %
I3
1.89 %

Exhw
Wp1 0.21 %
0.33 %

Fig.4. Exergy flow (percentage) diagram of DH system.

824
 0 0
The supply and return water temperatures are 140-180 C and 105-135 C, respectively, which therefore have been
taken in to account in the analysis in this paper. Exergy losses in hot water distribution pipes, in heat exchanger and from
the circulation of hot water, change with the temperature of hot water supplied and returning. In fig. 5, variation of exergy
losses in hot water distribution pipes due to the change of supplied and returning water temperature may be seen.
Exergy losses increase with an increase of Ts while it decreases with an increase of Tr. A stronger dependence on the
supply temperature rather than the return temperature is the case, when looking at fig.5.

Tr (0C)

Fig.5. Exergy loss due to heat loss

Electrical energy is supplied to the system to drive the circulating pumps. With a constant demand for the
amount of thermal energy to be supplied, the mass flow of the water varies and, therefore, so does the
amount of supplied electrical energy. The exergy losses formed during heat transfer in exchangers are so high. With
0
an increase of 10 C in Ts causes an increase of about 1.6% in exergy losses. Besides, exergy losses increase
withincreasing Tr. Finally, exergy losses in exchanger are about 50% of the ones provided for consumer.

The heat supplied to consumers also delivers a certain amount of exergy, which can be calculated using above Eq. If
the total exergy loss is reduced to that of the amount of fuel exergy, the curve shown in Fig. 6 is obtained. The maximum
values reach a factor of almost 16%, which means that in this case, almost 16 % of the fuel exergy is lost during the
transport and distribution of heat.

Tr (0C)

Fig. 6. Total exergy loss reduced to total fuel exergy

4. Conclusion

In this paper, the energy and exergy analysis occurring in the district heating system have been investigated. The energy
and exergy efficiency values for the DHS are found to be 74.99 and 8.46%, respectively. The analysis exergy loss which
occurs during the transport of thermal energy to consumers indicates that this loss is large and primarily dependent on
the temperature of the hot water. The total exergy losses increase 0.75% with increasing the supply water temperature
about 10 0C, which is the case for the return water temperature. This analysis shows that this loss during heat transfer in
district heating network can be reduce, by reducing the consumption of electrical energy during the transport of hot water
to the consumer and by reducing heat loss in pipelines. These heat losses should be kept at a minimum, which is

825
possible by lowering the supply temperature from the plant and by increasing of thermal insulation thickness in pipes.
However, lowering the supply temperature could result in unacceptably low temperature levels at the customers.
Furthermore, if the supply temperature is reduced, the water flow in the system increases, resulting in higher pumping
costs. The authors hope that these findings will help consumers used district heating systems.

References

[1]. B. Ludwig., On the sustainability of future energy systems. Energy Convers. Magmt, 38 (1997), 1765-1776.
[2]. D.K. Baker, S.A. Sherif, Heat transfer optimization of a district heating system using search methods. Int. J.
Energy Research, 21(1997), 233-252.
[3]. T. Yılmaz, District heating, M.Sc. thesis (in Turkish) İstanbul technical university, Turkey, 1988
[4]. I.Adamo, G. Cammarata, A. Fichera, L. Marletta, Improvement of a district network through thermoeconomic
approach. Renewable Energy, 10(1997), 2-3, 213-216.
[5]. A. Benonysson, B.Bohn, H.F. Ravn, Operational optimization in a district heating system, Energy. Convers.
Magmt, (1995) 36, 5, 297-314.
[6]. M. Bojic, N. Trifunovic, S.I. Gustafsson, Mixed 0-1 sequential lineer programming optimization of heat distribution
in a district-heating system,. Energy and Building, (2000) 32, 309-317.
[7]. B. Bohn, On transient heat losses from buried district heating pipes. Int. J. Energy Research, (2000) 24, 1311-
1334.
[8]. L. Gustavsson, A . Karlsson, Heating detached houses in urban areas, Energy 28 (2003) 851-875
[9]. M.A. Rosen, W.H. Leong, M.N. Le, Modelling and analysis of building systems that integrate cogenaration and
district heating and cooling. www.esim.ca/2001/documents/
[10]. I. Dincer, The role of exergy in energy policy making. Energy Policy, 30 (2002) 137-149.
[11]. A. Bejan, Fundamentals of exergy analysis, entropy generation minimization, and the generation of flow
architecture. Int. J. Energy Research, (2002) 26, 545-565
[12]. K. Çomaklı, Energy and exergy analysis of district heating plant of Atatürk university, Ph.D. thesis (in Turkish),
Ataturk university, Erzurum, Turkey,2003
[13]. T.J. Kotas, The exergy method of thermal plant analysis, 2nd edition Krieger Publishing Company USA 1995
[14]. Çengel YA., Boles MA., Thermodynamics: An engineering approach. 5. ed. McGraw-Hill.

826
 EXERGETIC PERFORMANCE ASSESSMENT OF A BINARY GEOTHERMAL POWER PLANT
a, * b c
Yildiz Kalinci , Arif Hepbasli and Ibrahim Dincer
a
Department of Technical Programs, Izmir Vocational School, Dokuz Eylul University, Education Campus Buca, Izmir,
Turkey
b
Department of Energy Systems Engineering, Faculty of Engineering, Yasar University, 35100 Bornova, Izmir, Turkey
c
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology (UOIT), 2000 Simcoe Street
North, Oshawa, Canada L1H 7K4
yildiz.kalinci@deu.edu.tr, arif.hepbasli@yasar.edu.tr, Ibrahim.Dincer@uoit.ca
* Corresponding author. E-mail address: yildiz.kalinci@deu.edu.tr

ABSTRACT
Electricity generation is achieved by means of the medium-temperature geothermal water in the range of 90–
140ºC in binary plants with the Organic Rankine Cycle (ORC). In this study, we present a case study using the
literature data as the respective energy and exergy analyses of a binary geothermal power plant are performed.
Also, exergy destruction rates throughout the plant are quantified and illustrated for comparison purposes. In the
plant considered, the brine injection and reinjection temperatures are 140 ºC and 80 ºC, with a mass flow rate of
64.87 kg/s, respectively. The energy and exergy efficiencies are calculated as 5.34% and 30.84%, respectively,
based on the heat and exergy input rates to the system at the net power. Furthermore, we examine the effects of
some parameters on energy and exergy efficiencies and net power output (e.g., brine injection temperature, brine
mass flow rates, turbine inlet temperature and inlet pressure).

Keywords: Energy, exergy, efficiency, geothermal power plant, ORC.

INTRODUCTION
"Geothermal" comes from the Greek words geo (earth) and therme (heat). So, geothermal means earth heat. Our
earth's interior - like the sun - provides heat energy from nature. This heat - geothermal energy - yields warmth and
power that we can use without polluting the environment (Geothermal Energy Facts, 2012).
The utilization of geothermal source either for power generation, cooling or heating applications depends mainly
on the source temperature. High temperature geothermal resources above 150 ºC are generally used for power
generation. Moderate-temperature (between 90 and 150 ºC) and low temperature (below 90 ºC) geothermal
resources are best suited for direct applications, such as space and process heating, cooling, aquaculture, and
fish farming (Ghafghazi et al., 2010; Kilkis, 2011; Ratlamwala et al., 2012).
Basic use of geothermal source at higher temperature is that of power production. Today, world installed electrical
generation capacity is more than 10 GW (Ratlamwala et al., 2012). High temperature geothermal resources, such
as dry steam and hot water as well as medium temperature geothermal resources such as water of moderate
temperature can be profitably used to generate electricity using three types of geothermal power plants (GPPs):
dry steam, flash and binary power plants. Dry steam geothermal power plants use very hot steam (> 235 ºC) and
limited amounts of water from the geothermal resources. Flash steam power plants (single or double) use hot
water (> 180 ºC), while binary cycle system uses water at moderate temperatures (100-180 ºC) coming from
geothermal resources (Coskun et al., 2011). An application ranges of varies types of geothermal power plants can
be seen in Fig.1.
Note that binary plants convert medium temperature resources into electricity more efficiently than other
technologies. In binary plants a heat exchanger transfers heat from the produced hot geofluid in a primary loop to
a low boiling-point working fluid in a secondary loop, such as propane, isobutene, pentane, isopentane, etc. This
thermodynamic cycle is known as ORC because initially organic compounds are used as the working fluid. The
working fluid in the secondary loop is evaporated in the vaporizer by the geothermal heat provided in the primary
loop. The vapour expands as it passes through the organic vapour turbine which is coupled to the generator. The
exhaust vapour is condensed in a water-cooled condenser or air cooler and is recycled to the vaporizer by the
feed pump. The cooled geofluid can be discharged or reinjected into the reservoir without flashing, which
minimizes scaling problems. It makes decentralized geothermal production feasible and economically attractive in
many remote or less developed regions of the world, where financial incentives promote low CO 2 emission energy
production technologies (Guzović et al., 2010).
There are some studies about geothermal binary power plants in the literature as follows: DiPippo (2007)
discussed the Carnot and the triangular efficiencies for geothermal binary power plants. He said that a more useful
model was the triangular (or trilateral) cycle because binary plants operating on geothermal hot water used a non-

827
isothermal heat source. The triangular cycle imposed a lower upper bound on the thermal efficiency and served as
a more meaningful ideal cycle against which to measure the performance of real binary cycles.

Fig.1. Application ranges of varies types of geothermal power plants (Clauser, 2006; Guzović et al., 2012).
Kose (2007) examined the potential and utilization of the existing geothermal energy resources in Kutahya–Simav
region. The temperature of the geothermal fluid in the Simav–Eynal field was too high for the district heating
system. Therefore, the possibility of electrical energy generation by a binary cycle was researched and the
preliminary feasibility studies were conducted in the field. For the environmental reasons, the working fluid used in
this binary power plant was chosen as HCFC-124. It was concluded that the Kutahya–Simav geothermal power
plant had the potential to produce an installed capacity of 2.9 MW e and a minimum of 17,020 MWh/year electrical
energy could be produced from this plant.
Kanoglu and Bolatturk (2008) studied exergy analysis of a binary geothermal power plant using actual plant data
to assess the plant performance and pinpoint sites of primary exergy destruction. Exergy destruction throughout
the plant was quantified and illustrated using an exergy diagram, and compared to the energy diagram. The sites
with greater exergy destructions included brine reinjection, heat exchanger and condenser losses. The energy and
exergy efficiencies of the plant were 4.5% and 21.7%, respectively, based on the energy and exergy of
geothermal water at the heat exchanger inlet. The energy and exergy efficiencies were 10.2% and 33.5%,
respectively, based on the heat input and exergy input to the binary Rankine cycle.
Yari (2010) presented a comparative study of the different geothermal power plant concepts, based on the exergy
analysis for high-temperature geothermal resources. The considered cycles for this study were a binary
geothermal power plant using a simple ORC, a binary geothermal power plant using an ORC with an internal heat
exchanger (IHE), a binary cycle with a regenerative ORC, a binary cycle with a regenerative ORC with an IHE, a
single-flash geothermal power plant, a double-flash geothermal power plant and a combined flash-binary power
plant. The maximum first-law efficiency was found to be related to the ORC with an IHE with R123 as the working
fluid and was calculated to be 7.65%. In contrast, the first-law efficiency based on the energy input into the ORC
revealed that the binary cycle with the regenerative ORC with an IHE and R123 as the working fluid has the
highest efficiency (15.35%). Also, the maximum first-law efficiency was shown to be given by the flash-binary with
R123 as the working fluid and was calculated to be 11.81%.
A modified exergoeconomic model was proposed for geothermal power plants using exergy and cost accounting
analyses, and a case study was in this regard presented for the Tuzla geothermal power plant system (Tuzla
GPPS) in Turkey to illustrate an application of the currently modified exergoeconomic model by Coskun et. al
(2011). Tuzla GPPS had a total installed capacity of 7.5 MW and electricity was generated using a binary cycle.
Exergy efficiency values varied between 35% and 49% with an average exergy efficiency of 45.2%, respectively.
Ganjehsarabi et al. (2012) performed an exergy analysis of the Dora II geothermal power plant with 9.5 MW net
power output by using actual plant data to evaluate plant performance and pinpoint the locations of exergy
destructions/losses. The overall energetic and exergetic efficiencies of the plant were calculated to be 10.7% and
29.6%, respectively. The largest exergy destruction rate took place in brine reinjection as 10.3 MW. The second
and third highest exergy destruction rates occurred in cooling tower and turbine with 4.7 MW and 3.4 MW,
respectively.
The present study aims at contributing to the area of binary geothermal power plant through energy and exergy
analyses. In this regard, the main objectives of this study are to: (i) calculate the energy and exergy efficiencies for
both the entire plant and the individual plant components, (ii) identify exergy destruction rates, and (iii) investigate
the parameters affecting energy and exergy efficiencies.

SYSTEM DESCRIPTION

828
There are several sources of medium-temperature geothermal water in the range of 90–140ºC which it is suitable
to produce electricity by binary plants using ORC. In this study, the data were taken from the literature (Guzović et
al., 2012) on a case study using a binary power plant as apparently modified for analysis. The geothermal fluid is
taken as 1600 kPa and 140 ºC and 64.87 kg/s from production well and then it is injected at 1600 kPa and 80ºC,
respectively.

Fig. 2. Process flow chart (modified from Guzović et al. (2012)).

The geothermal fluid transfers its heat to the working fluid – isopentane – inside the evaporator and the preheater.
In the preheater, the condensed isopentane is heated to the boiling point and transforms into dry saturated steam
within the evaporator. The dry saturated steam expands in the turbine, providing mechanical work to drive the
electric generator. After expansion in the steam turbine, the steam is fed to the air-cooled condenser. The
condensation heat is transferred to the environment by the forced convection of the air. The condensate is brought
to the initial pressure by the feed pump and returns to the preheater and the evaporator. Isentropic efficiencies for
the turbine and the feed pump are given as 0.85 and 0.8, respectively. Also, average annual air temperature is 15
°C. Isopentane operates between 700 kPa and 95 kPa with a mass flow rate of 36.04 kg/s.

ANALYSIS
The following assumptions are made in this study (Kanoğlu and Bolatturk, 2008; Yari, 2010).
 The operation of the geothermal power plant is considered in steady-state.
 The pressure drops throughout the heat exchangers and pipelines are considered negligible.
 The turbines and pumps have isentropic efficiencies.
 The changes in kinetic and potential energies are considered negligible.
 The geothermal fluid is treated as water.

The mass, energy and exergy balances for any control volume at steady state with negligible kinetic and potential
energy changes can be expressed, respectively, by

 m  m
in 0ut (1)

 E  E
in 0ut (2)

Q  W   m0ut hout   minhin (3)

and exergetically, it becomes

829
Exheat  W   Exin   Exout  I (4)

where the subscriptions in and out represent the inlet and exit states, E , Q and W are the energy rate, the heat
rate and the work input rate, m is the mass flow rate, h is the specific enthalpy. Also, Ex , I are the exergy rate
and the irreversibility (or exergy destruction) rate. Exheat is the net exergy transfer by heat at the temperature T
and the subscript 0 stands for the restricted dead state, which is given by

 T 
Exheat    1  0 Q (5)
 T 

The specific flow exergy is given by

ex  (h  h0 )  T0 (s  s0 ) (6)

where s is the specific entropy. Multiplying the specific exergy by the mass flow rate of the fluid gives the exergy
rate:

Ex  m ex (7)

The energy and exergy efficiencies are generally defined as

Eout
 (8)
Ein

Exout
 (9)
Exin

To determine the exergetic availability of a geothermal resource, the following equation can be used (Franco and
Villani, 2009):

Ex (T  Trej )  T0 ln(Tgeo / Trej )

 geo (10)
Ex0 (Tgeo  T0 )  T0 ln(Tgeo / T0 )

To compare geothermal plants, another parameter is the specific brine consumption, which is given by

mgeo
 (11)
Wnet

In addition, energetic and exergetic relations for components are given in Table 1.

Table 1. Energetic and exergetic relations for the subsystems.

Subsystem Energy relations Exergy relations
Preheater mgeo (h2  h3 )  mwf (h6  h5 ) IPH  (Ex2  Ex3 )  (Ex6  Ex5 )
Evaporation mgeo (h1  h2 )  mwf (h7  h6 ) IEV  (Ex1  Ex2 )  (Ex7  Ex6 )
Feed pump Wp  mwf (h5  h4 )  mwf (h5s  h4 ) / ps IP  Wp  (Ex5  Ex4 )
Turbine WT  mwf (h7  h8 )  mwf (h7  h8s )Ts IT  (Ex7  Ex8 )  WT
Condenser Q  mwf (h8  h4 )  mair (h10  h9 ) IC  Ex8  Ex4
Cycle Wnet Wnet
cycle   cycle 
mgeo (h1  h3 ) Ex1  Ex3

830
System Wnet Wnet
sys   sys 
mgeo (h1  h0 ) Ex1

RESULTS AND DISCUSSION

In this study, the brine has a mass flow rate of 64.87 kg/s at 1600 kPa and 140 °C. Due to the medium
temperature, binary geothermal power plant is feasible to produce electricity. The rejection temperature is selected
based mainly on the chemical composition of the geothermal fluid; in particular, it is necessary that the rejection
temperature should be high enough to avoid silica oversaturation, which could lead to silica scaling and serious
fouling problems in recovery heat exchangers, and in mineral deposition in pipes and valves. Considering that low-
enthalpy geothermal fields have temperatures between 110°C and 160°C and that it is difficult to use rejection
temperatures lower than 70°C–80°C (Franco and Villani, 2009). So, the rejection temperature is 80°C.
Isopentane is recommended as the working fluid in the ORC for the medium-temperature geothermal resources
because among all the refrigerants and the hydrocarbons that come into consideration, it shows the best
properties: low specific volumes, high efficiency (net power), moderate pressures in the heat exchangers, low cost,
low toxicity, low ozone depletion potential, low global warming potential and a low pinch-point temperature
(Guzović et al., 2012).
In the study, upper cycle pressure is defined according to the pinch-point temperature difference, which is a
difference between the brine pinch-point and the vaporization temperatures of isopentane. In the thermodynamic
calculations, a special attention is given to the values of the pinch points that are not below 5°C. In this study, the
upper pressure is 700 kPa and its saturated temperature is 98.62°C. Using Table 1 for evaporation, the
temperature of state 2 is defined as 104°C. According to this, the pinch-point temperature difference is calculated
as 5.41°C. A process flow chart is given in Fig. 2 where the main devices are preheater, evaporation, turbine,
condenser with air-cooled condenser and feed pump. According to this figure, energy analysis of each points can
be made using Eqs. (1)-(3) and the relations given in Table 1. Here, the thermodynamic properties of the working
fluid are determined using the EES software package and geothermal fluid is considered as water. While the
binary cycle is considered as only cycle, geothermal energy and binary cycle are considered as a system.
Thermodynamic properties of the fluids are calculated and are given in Table 2 for all the points. According to
these data, the turbine produces 2229 kW and the feed pump consumes 44.34 kW. The net power is 2184 kW.
Also, it is estimated that the parasitic power is about 16.6% of the net power generated in the cycle (Kanoglu and
Bolatturk, 2008). The parasitic power includes the brine production pumps, the condenser fans and other
auxiliaries. Subtracting the parasitic power, 362.6 kW, from the net power generated in the cycle, the net power
output becomes 1822 kW. The input energy rates to the cycle and the system are calculated as 16470 kW and
34189 kW. Using Eq. (8) and the data in Table 1 for the cycle and the system, the energy efficiencies are
calculated as 11.06% and 5.34%, respectively.

Table 2. Thermodynamic properties of fluids and exergy rates of the system.

State Fluid m P T h s ex Ex
no (kg/s) (kPa) (°C) (kJ/kg) (kJ/kg) (kJ/kg) (kW)
0 Geothermal 101.325 15 63.01 0.2242
0’ İsopentane 101.325 15 -372.4 -1.766

1 Geothermal 64.87 1600 140 590 1.738 91.05 5906

2 Geothermal 64.87 1600 104 437.2 1.351 49.66 3221
3 Geothermal 64.87 1600 80 336.6 1.074 28.33 1838

4 Isopentane 36.04 95 25 -350 -1.69 0.382 13.77

5 Isopentane 36.04 700 25.32 -348.8 -1.689 1.375 49.55
6 Isopentane 36.04 700 98.62 -167 -1.146 26.86 968.1
7 Isopentane 36.04 700 98.62 108.1 -0.406 88.85 3202
8 Isopentane 36.04 95 54.33 46.3 -0.373 17.33 624.7

9 Air 1414.12 101.325 15 288.5 5.661 0

10 Air 1414.12 101.325 25 298.6 5.695 0.308 435.55

In the analysis of such systems, energy analysis has generally been a common tool for analysis. The first law of
thermodynamics provides energy analysis, which identifies only externally energy wastes and losses. In this
regard, exergy comes out of the second law to be a powerful and effective tool for (i) designing and analyzing

831
energy systems by combining the conservation of mass and energy principles with the second law of
thermodynamics, (ii) furthering the goal of more efficient energy resource use by assessing meaningful efficiencies
and enabling the locations, types and true magnitudes of wastes and losses to be determined, (iii) revealing
whether or not, and by how much, it is possible to design more efficient energy systems by reducing the
inefficiencies in existing systems, (iv) addressing the impact on the environment of energy resource utilization, (v)
helping to achieve sustainable development (see Rosen and Dincer, 2003 for details).

Table 3. Exergetic investigation of the subsystems.

Component Exin (kW) Exout (kW) I (kW)  (%)
Feed pump 44.34 35.78 8.56 80.69
Preheater 1383 918.55 464.45 66.42
Evaporator 2685 2223.90 461.10 82.83
Turbine 2577.30 2229.00 348.30 86.49
Condenser 610.93 435.55 610.93 71.29

 cycle (%) 44.78

 sys (%) 30.84

Fig. 3. Energy and exergy loss diagrams of the system based on a) brine energy input rate (34189 kW) b) brine
exergy input rate (5906 kW).

The specific exergy values and exergy rates of all the points are calculated using Eqs. (4)-(7) and the values
given in Table 2. While the input exergy rate to the system is 5906 kW, the rate is 4068 kW for the cycle at 1822
kW power output. According to this, the exergy efficiencies of the cycle and the system are defined as 44.78% and
30.84%, respectively. After a general investigation, it can be focused on the subsystems and the defined exergy
destruction rates. According to the input and output exergy rates of the equipment, the exergy destruction rates
are calculated using Table 1 and the results obtained are listed in Table 3. The maximum exergy destruction rate
is due to the air cooled condenser as 610.93 kW. In this study, th exergy destruction rate of the air cooled
condenser is calculated according to the difference between States 8 and 4 because the waste heat rate cannot
be transformed to the useful work by air. The exergy efficiencies vary between 66.42% and 86.49%, which belong
to the preheater and the turbine, respectively. In addition, the energy and exergy losses of the system based on
the brine energy and exergy input rates are shown in Fig. 3. The energy input rate to the system is 34189 kW.
While 51.83% of the energy is reinjected, 41.78% of it is lost in the air cooled condenser, so, the net power is only
5.33%. This is an important argument subject whether geothermal resources should be used in power generation.
From the exergetic point of view, the results are similar to the energy analysis results. The biggest exergy loss is
31.12%, which belongs to the brine reinjection. Other exergy losses can be seen in the same figure.

The results of parametric study

832
A large number of design variables and operation parameters should be considered in geothermal power plants as
temperature, pressure, chemical composition of brine, rejection temperature, the ambient temperature and the
maximum rate of energy extraction.

Fig. 4. Reference values for exergy efficiency.

The inlet and rejection temperatures of the brine affect geothermal plants. Eq. (10) represents the ratio of the
theoretical work that can be extracted from the geofluid for given inlet geofluid temperatures and assumed
rejection temperatures, and the maximum theoretical work that can be extracted for given inlet geofluid and dead-
state temperatures. It provides an upper limit to the exergy efficiency. Fig. 4 shows the dependence of the
Ex / Ex0 ratio on the temperature of the source ( Tgeo ) and the rejection temperature of the brine ( Trej ). The ratio is
defined as 0.698 for T0, Tgeo and Trej as 15 °C, 140 °C and 80 °C, respectively. In addition, using Eq. (11), the
specific brine consumption is calculated as 35.64 kg/MJ. The value is used a comparative parameter for
geothermal power plants, generally.

Net power Mass flow rate

2100 39 30 40

Exergy Efficiency (%)

Energy efficiency 38
Mass Flow Rate (kg/s)

Energy Efficiency (%)

1900 38
25 Exergy efficiency 36
Net Power (kW)

1700 38
20 34
1500 37 32
1300 37 15 30
1100 36 28
10 26
900 36
5 24
700 35 22
500 35 0 20
500 550 600 650 700 750 500 550 600 650 700 750

Cycle Upper Pressure (kPa) Cycle Upper Pressure (kPa)

(a) (b)
Fig. 5. Comparison of the system according to the cycle upper pressure ( Trej = 80 °C) a) variation of net power and
mass flow rate b) variation energy and exergy efficiencies.

While the data for the brine are taken to be constant, others parameters can be investigated. At this point, the
cycle upper pressure and turbine inlet temperature are important. As can be seen in Fig. 5a, more power is
generated by higher pressure in spite of decreasing the working fluid mass flow rate. A limit to the pressure
increase is represented by the minimal temperature difference between the geothermal fluid and the isopentane at
the heat exchangers pinch point. The optimum pressure is considered to be the maximum pressure, at which the
pinch-point difference reaches an agreed minimum of 5°C (Guzović et al., 2012). While the temperature difference
is 5.41 °C for 700 kPa, it is 3.09 °C for 750 kPa. So, 750 kPa and more pressures are not convenient. Fig. 5b gives
a variation of energy and exergy efficiencies. Both of them increase with a higher pressure due to more turbine
power.
Also, the heat exchanger pinch point can be taken 5.41°C as constant. To ensure this, the mass flow rate of the
working fluid and the rejection temperature of the brine are considered as unknown parameters. As can be seen
in Fig. 6, more pressure means more rejection temperature of the brine. That is, less heat is picked up from
geothermal brine to isopentane. At this point, it can be considered that the waste heat of the brine can be used at

833
low temperature applications (e.g., district heating, industrial heating, green house). With decreasing the mass flow
rate and the net power, both energy and exergy efficiencies decrease.

2100 Net power Trej 90

30 40

Exergy Efficiency (%)

1900 80 Energy efficiency
38

Energy Efficiency (%)

Net Power (kW)

1700 70 25 Exergy efficiency 36

60 34
1500 20

Trej (°C)
50 32
1300 15 30
40
1100 28
30 10
900 26
20
5 24
700 10 22
500 0 0 20
500 550 600 650 700 500 550 600 650 700

Cycle Upper Pressure (kPa) Cycle Upper Pressure (kPa)

(a) (b)
Fig. 6. Comparison of the system according to the cycle upper pressure ( Tpp = 5.41 °C) a) variation of net power
and Trej b) variation of energy and exergy efficiencies.

1900 Net power Trej 90 30 40

Energy efficiency
38
1700 88

Exergy Efficiency (%)

25
Energy Efficiency (%)

Exergy efficiency 36
Net Power (kW)

1500 86 34
20
1300 84 32
Trej (°C)

15 30
1100 82 28
10 26
900 80
5 24
700 78
22
500 76 0 20
100 120 140 160 100 120 140 160

Turbine Inlet Temperature (°C) Turbine Inlet Temperature (°C)

(a) (b)
Fig. 7. Comparison of the system according to the turbine inlet temperature ( Tpp = 5.41 °C) a) variation of net
power and Trej b) variation of energy and exergy efficiencies.

Generally, in power plants, the superheated steam enters to turbine to take more power and avoid erosion of the
turbine blades. If the working fluid is heated to the superheated region in the evaporation equipment, its mass flow
rate decreases from 35.63 kg/s at 100 °C to 23.77 kg/s at 160 °C. The situation brings less net power and less
energy and exergy efficiencies, as can be seen in Fig. 7. Dry fluids, such as isopentane, are characterized by a
positive slope of the saturated vapor curve in the T–s diagram. Because of this characteristic, dry fluids do not
need to be superheated. The reason for this is that after expansion, the saturated vapor remains in the
superheated region. Additionally, problems related to the flow of wet steam through a turbine are excluded. Thus,
preventing problems, such as erosion of the turbine blades, droplet separation, condensate draining and similar
problems, is not necessary, which simplifies the turbine design (Desai and Bandyopadhyay, 2009; Shengjun et al.,
2011; Vanslambrouck et al., 2011; Guzovic et al., 2012). While the situation is good for the turbine, the heat
content of the superheated vapor is dissipated in the condenser. Therefore, a regenerator can be used to increase
the cycle efficiency.

Conclusions
We have investigated a binary geothermal power plant using the data taken from the literature through energy
and exergy efficiencies in this study. We have also calculated exergy destruction rates for main components of
the system. Furthermore, we have studied the effects of some important parameters (e.g., the inlet/rejection
temperature of the brine, the cycle upper pressure, the turbine inlet temperature) on the energy and exergy
efficiencies and indicated in figures.
We can extract some concluding remarks from this study as follows:

834
  The system generates a net power of 2184 kW for energy and exergy input rates of 34189 kW and 5906
kW, respectively.
 Due to the big energy input rate, the system has a less energy efficiency of 5.34%. Also, exergy efficiency
is calculated as 30.84%. There is still an argument at this point that geothermal sources should be used
in power generation. In addition, 51.83% of the brine energy is rejected. To increase the energy/exergy
efficiencies, the waste heat can be used for lower temperature applications.
 Operation parameters affect the net power seriously. Some of them can be taken as constant for the plant
(e.g., brine inlet temperature, brine mass flow rate, maybe rejection temperature). According to the
decision, other ones (e.g., turbine inlet temperature, cycle upper pressure, working fluid mass flow rate)
are investigated to increase the efficiencies and the net power. It is important that dry fluids, as
isopentane, do not need to be superheated. Also, after the turbine, a regenerator can be considered.
 Another advantage of the binary technology is that the geothermal fluids (or brines) do not contact the
moving mechanical components of the plant (e.g. the turbine), assuring a longer life for the equipment.
The results are expected to be beneficial to the researchers, government administration, and engineers
working in the area of the geothermal power plants.

Acknowledgement
The authors gratefully acknowledge the support provided by Dokuz Eylul University, Yasar University, University of
Ontario Institute of Technology as well as the Natural Sciences and Engineering Research Council and The
Scientific and Technological Research Council of Turkey.

NOMENCLATURE
ex specific exergy, kJ/kg
E energy rate, kW
Ex exergy rate, kW
h specific enthalpy, kJ/kg
I irreversibility or exergy destruction rate, kW
m mass flow rate, kg/s
P pressure, kPa
Q heat rate, kW
s specific entropy, kJ/kg K
T temperature, ºC or K
W Work rate or power, kW
Greek letters
 exergy or second law efficiency, %
 energy or first law efficiency, %
 interval (-)
 specific brine consumption, kg/MJ
Subscripts
C condenser
EV evaporator
geo geothermal
in inlet
out outlet
P pump
PH preheater
pp pinch point
rej rejection
s isentropic
sys system
T turbine
wf working fluid
0 reference index
Superscripts

835
Over dot quantity per unit time
Abbreviations
AC air condenser
EES engineering equation solver
EV evaporator
FN fan
FP feed pump
GN generator
GPPS geothermal power plant system
GPPs geothermal power plants
HP high pressure pump
IHE internal heat exchanger
IP injection pump
IW injection well
ORC organic Rankine cycle
P power
PH preheater
PW production well
TB turbine

REFERENCES
Clauser, C. 2006. Geothermal energy. In: Heinloth K, editor, Landolt-Bornstein. Group VIII: advanced materials
and technologies, vol. 3: Energy technologies, Subvol. C: renewable energies. Springer Verlag, Heidelberg-Berlin;
p. 493–604. <www.geophysik.rwth-aachen.de/IHFC/books.html> [accesed 15.09.09].
Coskun, C., Z. Oktay and I. Dincer. 2011. Modified exergoeconomic modeling of geothermal power plants. Energy
36:6358-6366.
Desai, N. B. and S. Bandyopadhyay. 2009. Process integration of organic Rankine cycle. Energy 34:1674–86.
DiPippo, R. 2007. Ideal thermal efficiency for geothermal binary plants. Geothermics 36:276–285.
Franco, A. and M. Villani. 2009. Optimal design of binary cycle power plants for water-dominated, medium-
temperature geothermal fields. Geothermics 38:379–391.
Ganjehsarabi, H.,, A. Gungor and I. Dincer. 2012. Exergetic performance analysis of Dora II geothermal power
plant in Turkey. Energy 46:101-108.
Geothermal energy facts. 2012. http://geothermal.marin.org/pwrheat.html#Q1, [accessed in 12.11.2012].
Ghafghazi, S., T. Sowlati, S. Sokhansanj and S. Melin. 2010. Techno-economic analysis of renewable energy
source options for a district heating project. International Journal of Energy Research 34:1109-1120.
Guzovic´, Z., B. Majcen and S. Cvetkovic´. 2012. Possibilities of electricity generation in the Republic of Croatia
from medium-temperature geothermal sources. Applied Energy 98:404–414.
Guzovic´, Z., D. Loncar and N. Ferdelji. 2010. Possibilities of electricity generation in the Republic of Croatia by
means of geothermal energy. Energy 35:3429-3440.
Kanoglu, M. and A. Bolatturk. 2008. Performance and parametric investigation of a binary geothermal power plant
by exergy. Renewable Energy 33:2366–2374.
Kilkis, B. A. 2011. Lignite-geothermal hybrid power and hydrogen production plant for green cities and sustainable
buildings. International Journal of Energy Research 35:138-145.
Kose, R. 2007. Geothermal energy potential for power generation in Turkey: A case study in Simav, Kutahya.
Renewable and Sustainable Energy Reviews 11:497–511.
Ratlamwala, T. A. H., I. Dincer and M. A. Gadalla. 2012. Performance analysis of a novel integrated geothermal-
based system for multi-generation applications. Applied Thermal Engineering 40:71-79.
Rosen, M. A. and I. Dincer. 2003. Exergy-cost-energy-mass analysis of thermal systems and processes. Energy
Conversion and Management 44:1633-51.
Shengjun, Z., W. Huaixin and G. Tao. 2011. Performance comparison and parametric optimization of subcritical
organic Rankine cycle (ORC) and transcritical power cycle system for low-temperature geothermal power
generation. Applied Energy 88:2740–2754.
Vanslambrouck, B., I. Vankeirsbilck, S. Gusev and M. De Paepe. 2011. Turn waste heat into electricity by using an
organic Rankine cycle. In: Proceedings of 2nd European conference on polygeneration, Tarragona, Spain; 1–14.

836
Yari, M. 2010. Exergetic analysis of various types of geothermal power plants. Renewable Energy 35:112-121.

837
 EXERGETIC EVALUATION OF A HIGH-PRESSURE HYDROGEN PRODUCTION
SYSTEM
1 2 3
Selcuk Inac , Adnan Midilli and Ibrahim Dincer
1
Energy Technologies, Marine Engineering, Turgut Kiran Maritime College,
Recep Tayyip Erdogan University, 53900, Rize, Turkey
2
Energy Division, Mechanical Engineering Department, Faculty of Engineering,
RecepTayyipErdogan University, 53900, Rize, Turkey
3
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology,
2000 Simcoe Street North, Oshawa, Ontario L1H 7K4, Canada
E-mails: selcuk.inac@erdogan.edu.tr, adnan.midilli@erdogan.edu.tr, Ibrahim.Dincer@uoit.ca

ABSTRACT
This study presents exergetic assessment of a high pressure hydrogen production system (HPS). The system
considered in this study includes a high-pressure-PEM electrolyser, two heat exchangers, a motor-pump
system, a circulation pump, and a hot water tank. In order to perform the main objective of this study, the
following important parameters are taken into consideration: (i) the operating pressure (ranging from 1 to 200
bar), (ii) the operating temperature (ranging from 70oC to 80oC), (iii) dead state temperature (assumed to be
25oC), (iv) the energy efficiency of the PEM electrolyser (ranging from 0.5 to 1), (v) the pump efficiency
(assumed to be 0.8), (vii) mass flow rate of hydrogen from PEM electrolyser (3 kg per hour) and (viii) mass
flow rate of pure water supplied to the PEM electrolysis (27 per hour). Considering all these parameters,
exergy analysis is performed for the HPS. The results show that exergy efficiency of HPS increases with the
rise of the operating pressure and temperature. It thus requires a higher amount of energy input. In this regard,
there is strong need to optimize the process.

Keywords: Hydrogen, exergy, energy, irreversibility, high-pressure-PEM electrolyser.

1. INTRODUCTION
Energy is a key factor in social, economic and sustainable development of countries (Ay et al., 2006;
Dincer, 1999; Midilli and Dincer, 2008; Ozsaban et al., 2011). However, many environmental issues causing
in global climate change are originated of energy production, consumption, and conversion (Midilli et al.,
2005). Today, fossil fuels (i.e., petroleum, natural gas and coal) are quickly consumed to provide of the
world energy demand. Particularly, utilization of the non-renewable fuels has caused to global warming due
to the carbon dioxide, methane, and nitrous oxide and chlorofluorocarbon emissions. In this regard,
hydrogen seen as an energy source of sustainability, unlimited clean energy, zero emission of greenhouse
gases and high energy efficiency has been accepted to be leader energy supply for the energy industry in
future (Dincer, 2012; Kalinci et al., 2009; Tolga Balta et al., 2010; Zheng et al., 2012).

One of the most effective methods of hydrogen production from water is electrolysis method. In this method,
electrolyser is a unique instrument for pure hydrogen and oxygen production from the water. There are
many electrolyser types but PEM electrolyser is very simple and compact among these electrolysers
(Gorgun, 2006). PEM (proton exchange membrane) electrolyser is a promising important technology.
However, PEM electrolyser is effective and safe technology for hydrogen production (Biaku et al., 2008).

Under these considerations, a detailed literature review is performed, based on some studies over high
pressure hydrogen production technologies available in the open literature (e.g., Armandi et al., 2012; Kim
et al., 2013; Laoun, 2007; Marangio et al., 2011; Marangio et al., 2009; Ni et al., 2008; Nieminen et al.,
2010; Pérez-Herranz et al., 2010; Santarelli et al., 2009). In this study, a parametric study on high pressure
hydrogen production is conducted to investigate how the exergy efficiency changes with system operating
conditions such as, temperature and pressure. Furthermore, in this study is also investigated the energy
requirement of PEM electrolyser and the HPS, the total irreversibility of all main components of the HPS of
the depending on operation temperature and pressure of the system (Inac, 2009; Kacar, 2009; Ozsaban,
2009).

2. SYSTEM DESCRIPTION
The following system is designed for high pressure hydrogen production and analyzed in this regard for
practical applications. The working principle of the HPS can be explained as follows.

838
 WPEM H2Opw
Wp
(T0,P0)
H2Opw H2Opw
(TPEM,PPEM) (TPEM,P0)
High Pressure PEM
Pump Heat Exchanger-I
Electrolyser

H2Ocw H2Ohw
O2
(TPEM,PPEM) Wcp
C.Pump
To industrial Heat Exchanger-II
H2
use
(TPEM,PPEM)
H2Ohw
H2Ocw H2
(To,PPEM)

To compression Hot Water Tank

unit

Hot water coming from

the compression unit

Fig.1. Schematic representation of the HPS [modified from (Ni et al., 2008; Oi and Sakaki, 2004)]

3. EXERGY ANALYSIS
In order to perform the parametric study, the following assumptions are made for the exergy analysis:
 Inlet temperature and pressure of pure water are 20oC and 1 atm, respectively.
 Hydrogen gas production flow rate is 3 kg/h.
 Operating temperatures of PEM electrolyser are taken as 70, 75, and 80oC
 Operating pressures of PEM electrolyser are taken as 1, 10, 20, 30, 40, 50, 100, 150 and 200 bar.
 Energy efficiencies of PEM electrolyser are taken as 0.5, 0.6, 0.7, 0.8, 0.9 and 1.
 Mass flow rate of pure water provided in system is 27 kg per hour.
 The plate heat exchangers in the system operate on an adiabatic process.
 Hydrogen and oxygen gases in system are real gas.
 Energy efficiencies of the pumps are 0.8.
 Thermodynamics properties corresponding operating temperature and pressure in system are taken
from the NIST (NIST, 2013). Depending on the values taken from other sources may vary results. In this
study, NIST have been selected as data source.
 Ambient temperature and pressure in system are taken to be 25oC and 1 atm, respectively.
 Temperature of heating water is assumed to be 90 oC.
 Temperature of cooling water is assumed to be 20 oC.
 The chemical exergy values are taken from the literature (Kotas, 1985).

In this regard, the energy and exergy balances of the HPS are shown as in Fig. 2, by considering a key
literature work (Dincer and Rosen, 2007). Thus, the exergy balance for a system can be written as follows:

(1)

where for a steady-state system, is zero, and are amount of total exergy input and output,
respectively, is total irreversibility of a system.

Here, the total irreversibility for a system can be expressed as

(2)

which is used for each of the main components in the HPS. However, total exergy ( ) can be calculated
from:

839
 (3)

where and are the physical and chemical exergy, respectively.

(4)
where changes of specific kinetic energy and specific potential energy are assumed to be zero in this study.

a) b)
Pure water PEM electrolyser
input work input To =298.15 K
𝑝𝑤 ExPEM To =298.15 K

Po = 1 atm Po = 1 atm

Heating
water Pump 𝑤 𝑝𝑢 𝑝
input work input
HIGH PRESSURE HIGH PRESSURE
HYDROGEN PRODUCTION HYDROGEN PRODUCTION
SYSTEM SYSTEM
Cooling Circulation
water pump 𝑐𝑤 𝑐.𝑝𝑢 𝑝
output work input

Oxygen Hydrogen 𝑂2 𝐻2
gas output gas output

Fig.2. Energy and exergy flow diagram of the HPS.

The physical exergy can be written as below:

(5)

where and s are enthalpy and entropy, ho and so are enthalpy and entropy at the reference (dead) state of
and , respectively.

The exergy flow rate can be calculated from

(6)

The total exergy balance of a system can be obtained by

(7)

where and are inlet and outlet mass flow rate, is the exergy transfer by heat, is the
exergy transfer by work.

Exergy transfer by heat and work can be expressed as

(8)

and

(9)

The exergy efficiency of a system can be written as (Baniasadi and Dincer, 2011):

(10)

where is the total irreversibility in a system and is the total exergy input to a system.

840
Firstly, the energy requirement of high pressure PEM electrolyser must be theoretically determined for high
pressure hydrogen production. Requirement of energy necessary for high pressure hydrogen production is
calculated as (Ni et al., 2008).

𝐻 (11)

where electricity requirement of PEM electrolyser, ; requirement of thermal energy of PEM

electrolyser. Chemical reaction in PEM electrolyser can be written as below (Ni et al., 2007).

𝐻𝑂 𝑂 𝐻 (12)

The electricity and heat requirement of PEM electrolyser can be defined as (Newell, 2000):

𝑐 𝑐

𝑐 𝑐

𝑐 𝑐 (13)

𝑐 𝑐

𝑐 (14)

The total energy requirement of PEM electrolyser ( ) can be defined as follow:

(15)

where

(16)

where is amount of total heat requirement in order to the realized chemical reaction, is total
exergy input as heat and work to PEM electrolyser.

Exergy efficiency of PEM electrolyser has been calculated by using Eq. (13) (Baniasadi and Dincer, 2011).

(17)

where is the total exergy input to PEM electrolyser, is the pure water exergy input to PEM
electrolyser.

The total irreversibility of PEM electrolyser can be defined as follow

(18)

The theoretic work of pump ( ) can be calculated from (Cengel and Boles, 2006)

(19)

where is the specific volume of pure water.

The real work of pump ( ) can be written as

841
 (20)

and
𝑝 𝑐𝑝
and (21)

where is electrical efficiency of pump, and are the total exergy input to pump and circulation
pump, respectively.

Under these considerations, the exergy efficiency of the HPS can be calculated by

- (22)

where is exergy efficiency of system, is total exergy input to the HPS (kW), is total
irreversibility of all components in the HPS (kW).

(23)

The total irreversibility of system is calculated by

(24)

The exergy balances of other main components in the HPS are showed in Table 1. As a result, the exergy
efficiency of the HPS can be written as follows:

𝑐𝑝 𝑝
(25)
𝑝𝑤 𝑤 𝑝𝑢 𝑝 𝑐𝑝

3. RESULTS AND DISCUSSION

In this study, an exergetic assessment of the newly developed HPS is performed. Fig. 3 shows the energy
consumption amount of PEM electrolyser as a function of the operating temperature and pressure of
system. In this study, the flow rate of hydrogen production is taken to be 3 kg/h. As shown in Fig. 3, the
energy requirement of PEM electrolyser increases with the rise of operating pressure and decreases with
the increase of operating temperature of the system.
164,4

164,3
Total energy requiremet of PEM electrolyser

164,2

164,1
(kWh/3kgH2)

164,0
Operating temperature of the system70 oC
Operating temperature of the system 75 oC
163,9 Operating temperature of the system 80 oC
Efficiencies of the pumps: 0.8
Efficiency of PEM electrolyser: 0.7
Flow rate of hydrogen produced: 3 kg/h
163,8

163,7
0 20 40 60 80 100 120 140 160 180 200
10 30 50 70 90 110 130 150 170 190
Operating pressure of the system (bar)
Fig.3. Variation of total energy requirement of PEM electrolyser depending on the operating pressure and
temperature of the system.

842
Furthermore, Fig.4 illustrates the exergy efficiency of PEM electrolyser as a function of the operating
temperature and pressure of system. As shown in Fig. 4, the exergy efficiency of PEM electrolyser goes up
with the increase of the operating temperature and pressure of the system.
0,66

0,65
Exergy efficiency of PEM electrolyser

0,64

0,63

0,62
Operating temperature of the system70 oC
Operating temperature of the system 75 oC
Operating temperature of the system 80 oC
0,61
Efficiencies of the pumps: 0.8
Efficiency of PEM electrolyser: 0.7
Flow rate of hydrogen produced: 3 kg/h
0,60

0,59
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
Operating pressure of the system (bar)
Fig.4. The variation of the exergy efficiency of PEM electrolyser depending on the operating pressure and
temperature of the system.

Fig. 5 indicates the exergy efficiency and total irreversibility of the HPS based on the operating temperature
and pressure of the system. As shown in Fig. 5, the exergy efficiency of the HPS increases with the rise of
the operating temperature and pressure of the system. However, the total irreversibility of the HPS
decreases with the increase of the operating pressure and temperature of the system.
0,66 67

66
0,65
65

0,64 64
Total irreversibility of the HPS
Exergy efficiency of the HPS

o
Operating temperature of the system70 C
63
0,63 Operating temperature of the system 75 oC
(kWh/3kgH2)

Operating temperature of the system 80 oC

Efficiencies of the pumps: 0.8 62
Efficiency of PEM electrolyser: 0.7
0,62 Flow rate of hydrogen produced: 3 kg/h
61

0,61 60

59
0,60
58

0,59 57
0 20 40 60 80 100 120 140 160 180 200
10 30 50 70 90 110 130 150 170 190
Operating pressure of the system (Bar)
Fig. 5. The variation of the exergy efficiency and total irreversibility of the HPS depending on the operating
pressure and temperature of the system.

843
4. CONCLUSIONS
In this study, a parametric study on the exergetic evaluation of the high pressure hydrogen production has
been presented. In this regard, exergy analyses both the HPS and PEM electrolyser are discussed
according to operating temperature and pressure of the system. The following two concluding remarks are
made.
 Total energy requirement of the HPS system and PEM electrolyser increases with the rise of operating
pressure of the system. In this regard, PEM electrolyser has maximum power consumption in the HPS.
 The overall system exergy efficiency increases and the total irreversibility of the HPS decreases with an
increase in operating pressure and temperature of the system, respectively.
Note that in order to find out the optimum operating pressure and temperature and to determine the
optimum exergetic efficiency and irreversibility of the system, the exergetic optimization should be
performed as to be done in the future.

ACKNOWLEDGEMENT
The authors acknowledge the technical support of Recep Tayyip Erdogan University in Turkey and
University of Ontario Institute of Technology, and Natural Sciences and Engineering Research Council of
Canada.

REFERENCES
Armandi, M., Drago, D., Pagani, M., Bonelli, B., Santarelli, M., 2012. Direct coupling of H2 production
through a high pressure PEM electrolyzer and its storage by physisorption on microporous materials.
International Journal of Hydrogen Energy 37, 1292-1300.

Ay, M., Midilli, A., Dincer, I., 2006. Investigation of hydrogen production from boron compounds for pem fuel
cells. Journal of Power Sources 157, 104-113.

Baniasadi, E., Dincer, I., 2011. Energy and exergy analyses of a combined ammonia-fed solid oxide fuel
cell system for vehicular applications. International Journal of Hydrogen Energy 36, 11128-11136.

Biaku, C., Dale, N., Mann, M., Salehfar, H., Peters, A., Han, T., 2008. A semiempirical study of the
temperature dependence of the anode charge transfer coefficient of a 6kW PEM electrolyzer. International
Journal of Hydrogen Energy 33, 4247-4254.

Cengel, Y.A., Boles, M.A., 2006. Thermodynamics an Engineering Approach, New York.
Degiorgis, L., Santarelli, M., Calì, M., 2007. Hydrogen from renewable energy: A pilot plant for thermal
production and mobility. Journal of Power Sources 171, 237-246.

Dincer, I., 1999. Environmental impacts of energy Energy Policy 27, 845-854.

Dincer, I., 2012. Green methods for hydrogen production. International Journal of Hydrogen Energy 37,
1954-1971.

Dincer, I., Hussain, M.M., Al-Zaharnah, I., 2005. Energy and exergy utilization in agricultural sector of Saudi
Arabia. Energy Policy 33, 1461-1467.

Dincer, I., Rosen, M.A., 2007. Exergy, environment and sustainable development. Exergy, 36-59.

Gorgun, H., 2006. Dynamic modelling of a proton exchange membrane (PEM) electrolyzer. International
Journal of Hydrogen Energy 31, 29-38.

Inac, S., 2009. T er modynamıc modelling and exergetic sustainability analysis of a system production high-
pressurised-hydrogen gas, Graduate School of Natural and Applied Science Department of Mechanical
Engineering,MSc Thesis. Nigde University, Turkey.

Kacar, O., 2009. Thermodynamic modeling of power generation via a run of river type hydropower plant,
Graduate School of Natural and Applied Science Department of Mechanical Engineering,MSc Thesis.
Nigde University, Turkey.

Kalinci, Y., Hepbasli, A., Dincer, I., 2009. Biomass-based hydrogen production: A review and analysis.
International Journal of Hydrogen Energy 34, 8799-8817.

Kim, H., Park, M., Lee, K.S., 2013. One-dimensional dynamic modeling of a high-pressure water electrolysis
system for hydrogen production. International Journal of Hydrogen Energy 38, 2596-2609.

844
Kotas, T.J., 1985. The exergy method of thermal plant analysis. Butterworth Publishers,Stoneham, MA,
United States.

Laoun, B., 2007. Thermodynamics Aspect of High Pressure Hydrogen Production by Water Electrolysis.
Revue des Energies Renouvelables 10, 435-444.

Marangio, F., Pagani, M., Santarelli, M., Calì, M., 2011. Concept of a i g pressure PEM electrolyser
prototype. International Journal of Hydrogen Energy 36, 7807-7815.

Marangio, F., Santarelli, M., Cali, M., 2009. Theoretical model and experimental analysis of a high pressure
PEM water electrolyser for hydrogen production. International Journal of Hydrogen Energy 34, 1143-1158.

Midilli, A., Ay, M., Dincer, I., Rosen, M.A., 2005. On hydrogen and hydrogen energy strategies II: future
projections affecting global stability and unrest. Renewable and Sustainable Energy Reviews 9, 273-287.

Midilli, A., Dincer, I., 2008. Hydrogen as a renewable and sustainable solution in reducing global fossil fuel
consumption. International Journal of Hydrogen Energy 33, 4209-4222.

Newell, T.A., 2000. Thermodynamic analysis of an electrochemical refrigeration cycle. INTERNATIONAL

JOURNAL OF ENERGY RESEARCH 24, 443-453.

Ni, M., Leung, M., Leung, D., 2007. Parametric study of solid oxide steam electrolyzer for hydrogen
production. International Journal of Hydrogen Energy 32, 2305-2313.

Ni, M., Leung, M.K.H., Leung, D.Y.C., 2008. Energy and exergy analysis of hydrogen production by a
proton exchange membrane (PEM) electrolyzer plant. Energy Conversion and Management 49, 2748-2756.
Nieminen, J., Dincer, I., Naterer, G., 2010. Comparative performance analysis of PEM and solid oxide
steam electrolysers. International Journal of Hydrogen Energy 35, 10842-10850.

NIST, 2013. Nist Web Sitesi<http://webbook.nist.gov/chemistry/fluid/>Last accessed:January 2013.

Oi, T., Sakaki, Y., 2004. Optimum hydrogen generation capacity and current density of the PEM-type water
electrolyzer operated only during the off-peak period of electricity demand. Journal of Power Sources 129,
229-237.

Ozsaban, M., 2009. Thermodynamic modeling and energy sustainability analysis of high pressurized
hydrogen gas storage process, Graduate School of Natural and Applied Science Department of Mechanical
Engineering,MSc Thesis. Nigde University, Turkey.

Ozsaban, M., Midilli, A., Dincer, I., 2011. Exergy analysis of a high pressure multistage hydrogen gas
storage system. International Journal of Hydrogen Energy 36, 11440-11450.

Pérez-Herranz, V., Pérez-Page, M., Beneito, R., 2010. Monitoring and control of a hydrogen production and
storage system consisting of water electrolysis and metal hydrides. International Journal of Hydrogen
Energy 35, 912-919.

Santarelli, M., Medina, P., Calì, M., 2009. Fitting regression model and experimental validation for a i g -
pressure PEM electrolyzer. International Journal of Hydrogen Energy 34, 2519-2530.

Tolga Balta, M., Dincer, I., Hepbasli, A., 2010. Energy and exergy analyses of a new four-step copper–
chlorine cycle for geothermal-based hydrogen production. Energy 35, 3263-3272.

Zheng, J., Liu, X., Xu, P., Liu, P., Zhao, Y., Yang, J., 2012. Development of high pressure gaseous
hydrogen storage technologies. International Journal of Hydrogen Energy 37, 1048-1057.

845
 POLYPYRROLE/CARBON SUPPORTED PLATINUM CATALYSTS FOR PEM FUEL CELLS

A. Bayrakçeken*, E. Das, G. Bozkurt and F. Memioglu

Atatürk University, Faculty of Engineering, Department of Chemical Engineering, 25240, Erzurum, Turkey
ayse.bayrakceken@gmail.com, das.elif@gmail.com, bozkurtgamze@ymail.com, fulya.memioglu@gmail.com

ABSTRACT
In this study, firstly, PPy/C composites were synthesized by in situ chemical oxidative polymerization of pyrrole
monomer on carbon black by changing the carbon amount in the synthesis in between 5-35%. Pt was incorporated
over the composites having different carbon ratios by using microwave irradiation technique. The composites and the
Pt loaded catalysts were characterized by using surface area analysis, four point probe conductivity analysis, XRD,
TGA and TEM techniques. The prepared catalysts were tested in PEM fuel cell.

INTRODUCTION
The heart of a proton exchange membrane (PEM) fuel cell is membrane electrode assembly (MEA) which is
comprised of the proton conducting membrane and the anode and cathode electrodes where the half cell oxidation
and reduction reactions occur (Bayrakçeken et al., 2008). The electrodes used in PEM fuel cells are mostly carbon
supported platinum (Pt) catalysts. Carbon supports are used in order to increase the surface area of the Pt used. As
an alternative conducting polymers and carbon (C) composites are used as the supports in PEM fuel cells in order to
decrease the fuel cell performance loss caused by the carbon corrosion of the carbon material (Bruijn et al., 2008)
during the long term operations of PEM fuel cells. Polypyrrole (PPy) is a promising conducting polymer material that
can be used as the support material (Memioğlu et al., 2012).

The support material properties including surface area, pore diameter and pore size distribution are very significant
that affect the properties of the catalysts (A. L. Dicks, 2006). The structural properties directly affect the particle size
and distribution of the metal nanoparticles on the support material. In order to obtain highly dispersed nanoparticles
on the support material the structural properties of the support material has to provide the requirements.

One of the parameters that affect the catalyst properties is the catalyst preparation technique. An energy efficient and
time saving catalyst preparation technique which is gaining an increasing attention is microwave irradiation technique
(Liu et al., 2004). This technique provides an efficient way of obtaining highly dispersed nanoparticles on the support
material.

In this study, PPy/C composites were synthesized by using chemical oxidative polymerization of pyrrole monomer
over carbon black. The carbon amount in the composite was varied in between 5-35%. The catalysts were prepared
by using microwave irradiation technique.

EXPERIMENTAL
PPy/C composites were synthesized by in situ chemical oxidative polymerization of pyrrole monomer on carbon
black (Memioğlu et al., 2012). Required amounts of the carbon black (Vulcan XC 72, C) was taken and then mixed
with ethanol. After stirring for half an hour distilled pyrrole solution was added to this mixture and stirred for 30 min
further. At this step, the dopant p-toluenesulfonic acid ( p-TSA) was added. Ammonium persulphate (APS) as the
oxidant was added to this mixture and stirred for 4 h. After stirring, the mixture was filtered, washed with ethanol,
distilled water and acetone, respectively. The carbon amount in the synthesis was varied as 5, 12, 20 and 35%.

Pt/composite catalysts were prepared by using microwave heating of ethylene glycol solutions of the Pt precursor
with a similar procedure given elsewhere (Liu et al., 2005). Briefly, required volume from 0.05 M aqueous solution of
Pt precursor (H2PtCl6) and 50 ml ethylene glycol (EG) were mixed in a beaker and then 0.1 g composite was added
to this solution. After ultrasonication for half an hour, the mixture was put in the microwave oven and heated for 60 s
by means of a 800 W microwave power. The resulting suspension was cooled immediately, and then filtered and the
residue was washed with acetone. The solid product was dried at 373 K overnight in a vacuum oven.

Synthesized composites were characterized by using BET and four point probe technique. Nitrogen adsorption-
desorption analyses were carried out by a Quantachrome Autosorb-1-C/MS instrument. Multipoint BET surface area,
pore size distribution and pore volume of the composites were analyzed by this technique. Electrical conductivities of
the composites were determined by using four point probe technique by using a custom manufactured set up
connected to a potentiostat (Princeton Applied Research, Versastat 3). XRD data of the Pt loaded catalysts were
obtained by a Rigaku Miniflex X-ray diffractometer. Thermogravimetric analysis of the catalysts were studied in a
Netzsch Thermal analyzer by heating the samples in flowing air from room temperature to 1000˚C at a heating rate of
10˚C/min in order to determine the Pt loading. JEOL JEM 2100F STEM instrument was used for TEM analysis.

PEM fuel cell tests were obtained in a 5 cm2 fuel cell hardware in Henatech fuel cell test station. The temperatures for
the fuel cell and the anode and cathode humidifiers were all set to 70ºC during the experiments. Anode and cathode

846
flow rates of hydrogen and oxygen were set to 0.25 slpm. The fuel cell was operated after feeding the gases at 0.5 V
at least one hour and then periodically performance curves were taken until no more change were observed.

RESULTS AND DISCUSSION

Multipoint surface areas of the composites were determined by using nitrogen adsorption-desorption isotherms. Table
1 summarizes the multipoint surface areas of the composite materials used as the catalyst support. As can be seen
from table that the surface area is increasing with an increase in the carbon amount within the composite.

Table 1. Multipoint BET surface areas of the synthesized composites.

2
Composite Carbon percentage (%) BET surface area (m /g)
PPy/C 5 49
PPy/C 12 80
PPy/C 20 77
PPy/C 35 116

Electrical conductivities of the composites prepared with different carbon ratios were obtained by four point probe
technique. The effect of pressing pressure of the pellets is also a parameter that affects the electrical conductivity of
the composites. The effect of pellet pressing pressure on the electrical conductivity of the composite having 20% C
was investigated as given in Figure 1. Sánchez-González et al. (2005) performed similar experiment for different
carbon materials and got similar results. It was seen that the electrical conductivity of the composite was increased
with an increase in pressing pressure. This increase can be attributed to the decrease in contact resistance which
caused the electrical conductivity to increase.

Fig.1. Effect of pressing pressure on the electrical conductivity of PPy/C composite having 20%C.

In order to get comparable results all the other composites were pelleted under 14 ton press pressure. The electrical
conductivities for the prepared composite materials obtained by four point probe technique are given in Table 2.

Table 2. Electrical conductivities of the composites.

Material Carbon percentage (%) Electrical Conductivity (S/cm)
PPy/C 5 0.0026
PPy/C 12 5.7636
PPy/C 20 34.422
PPy/C 35 70.197

TGA is used in order to determine the Pt loading over the composite materials. Figure 2 shows the TGA results for
the composite materials. The targeted nominal Pt loading for all the catalysts was 20%. It was observed that although
the surface areas of the composites changed significantly Pt loading in between 12-18 % was obtained for different
support materials.

847
 120
5% 12%
20% 35%
100

80
Mass loss, %

60

40

20

0
0 200 400 600 800 1000 1200
Temperature, oC

Fig.4. TGA results of the catalysts prepared over different support materials having carbon percentage in between 5-35%.

XRD results of the prepared catalysts are given in Figure 5 for the support materials having different carbon
percentages. The characteristic peaks for the fcc Pt is obtained for all catalysts.
Intensity (au)

35%

20%

12%

0 20 40 60 80 100
o
2 Theta,

Fig.5. XRD results of the catalysts.

TEM results of the catalysts are given in Figure 6. As can be seen from figure that a highly dispersed Pt nanoparticles
were obtained over the support materials. The nominal Pt loading provided for all catalysts were same, 20%, but the
dense of the nanoparticles over the support was also increased as the carbon amount increased.

848
(a) (b)

(c) (d)

(e) (f)

849
 (g) (h)

Fig.6. TEM results of the catalysts (a)-(b) 5% Pt/PPy:C (c)-(d) 12% Pt/PPy:C (e)-(f) 20% Pt/PPy:C (g)-(h) 35% Pt/PPy:C.

PEM fuel cell performance tests of the catalyst are given in Figure 7. As can be seen from figure that, the PEM fuel
cell performance, especially at higher current densities are increasing with an increase in the carbon amount. The
lowest performance was obtained for the 5% PPy/C support material which also has the lowest electrical conductivity.

1.0 0.08
5% Pt/PPy:C 12% Pt/PPy:C
20% Pt/PPy:C 35% Pt/PPy:C

0.8
0.06
Power Density, W/cm 2

0.6
Voltage, V

0.04

0.4

0.02
0.2

0.0 0.00
0 50 100 150 200 250

Current Density, mA/cm2

Fig.7. PEM fuel cell performance tests.

CONCLUSIONS
Long term operations of the PEM fuel cells are of great interest in order to widespread utilization in different areas.
The elimination of the nondurable materials will improve the way to the commercialization. There has to be an
optimization between the PEM fuel cell performance and the durability of the materials. The recommended structures
in this study are important to reduce the carbon corrosion occurred in the carbon supported catalysts during the long
term operations. The improvement and the optimization in the structures will improve the PEM fuel cell performance
further which also may affect the durability.

Acknowledgement
The authors gratefully acknowledge the financial support by Turkish Scientific and Technological Research Council
(TÜBİTAK) through grant number 110M081 and Atatürk University BAP projects through grant numbers 2009/256
and 2011/144.

850
REFERENCES
Bayrakçeken, A., Erkan, S., Türker, L. and İ. Eroğlu. 2008. Effects of membrane electrode assembly components
on proton exchange membrane fuel cell performance. International Journal of Hydrogen Energy 33 (1): 165-170.

Bruijn, F. A., Dam, V. A. T. and G. J. M. Janssen. 2008. Review: durability and degradation issues of PEM fuel cell
components. Fuel Cells 1: 3-22.

A. L. Dicks. 2006. The role of carbon in fuel cells. Journal of Power Sources 156: 128-141.

Liu, Z. L., Lee, J. Y., Chen, W., Han, M. and L. M. Gan. 2004. Physical and electrochemical characterizations of
microwave-assisted polyol preparation of carbon-supported PtRu nanoparticles. Langmuir 20: 181-189.

Liu, Z., Gan, L. M., Hong, L., Chen, W. and J.Y. Lee. 2005. Carbon-supported Pt nanoparticles as catalysts for
proton exchange membrane fuel cells. Journal of Power Sources 139:73-78.

Memioğlu, F., Bayrakçeken, A., Öznülüer, T. and M. Ak. 2012. Synthesis and characterization of
polypyrrole/carbon composite as a catalyst support for fuel cell applications. International Journal of Hydrogen
Energy 37:16673-16679.

Sánchez-González, J., Macίas-Garcίa, A., Alexandre-Franco, M. F. and V. Gómez-Serrano. 2005. Electrical

conductivity of carbon blacks under compression. Carbon 43: 741–747.

851
 COMPARISON OF NATURAL GAS FIRED AND INDUCTION HEATING FURNACES

U. Unver, H. M. Unver
1
Yalova Universitesi, Kazim Karabekir Mah. Rahmi Ustel Cad. No:1, Yalova Turkey
2
Kırıkkale Universitesi, Muhendislik Fakültesi Yahsihan Kirikkle Turkey
umit.unver@yalova.edu.tr, unver@kku.edu.tr

ABSTRACT
Generally, in steel processing systems, steels are heated via induction or combustion furnaces. Natural gas fired
furnaces have installation costs advantage and induction furnaces have the advantages of less scale formation on the
surface of the workpiece as well as environmental reasons. The operation cost of both systems is a vital argument to be
solved. In this paper, a natural gas fired and induction-heating furnace of a forging process was studied. Thermodynamic
analysis was performed for the furnaces which heats the work piece up to 1300 C. The energy performances of both
systems were analyzed and compared. The heating process is optimized and a hybrid process is suggested.

INTRODUCTION
To bequeath habitable environment, today’s world has to pay attention to energy sources and usages. This well-known
truth forced to change the energy policies of entire world. Today many countries encourage the tendency to the
renewable resources. However, fossil fuels cannot be given up suddenly due to both the number of clean power plants
and their low efficiencies. So, the critical energy demand cannot be individually supplied by renewable resources. Since
the installed powers of renewables are not sufficient yet, the energy saving applications or saved energy is also assumed
to be renewable by most academic communities.

INDUCTION HEATING
Induction heaters are used to provide alternating electric current to an electric coil (the induction coil). The induction coil
becomes the electrical (heat) source that induces an electrical current into the metal part to be heated (called the
workpiece). No contact is required between the workpiece and the induction coil as the heat source, and the heat is
restricted to localized areas or surface zones immediately adjacent to the coil (1). It provides faster and more precise
heating of local areas, consumes less energy and is considered environmentally friendlier than other methods. Other
advantages also include lower labor cost for device operators, easy maintainability of the equipment, quality
assurance, automation capability and high reliability. Induction heating is a complex process including
electromagnetic, thermal and metallurgic phenomena. In this process an alternating electric current induces
electromagnetic field, which in turn induces eddy currents in the workpiece. The induced eddy currents release energy
in the form of heat, which is then distributed throughout the workpiece (2).
Laborelec (3) indicated that the principle of induction heating is mainly based on two well-known physical
phenomena, electromagnetic induction and the Joule effect. The energy transfer to the object to be heated occurs by
means of electromagnetic induction. It is known that an alternating current is induced in a loop of conductive material
when this loop is placed in an alternating magnetic field. The formula is as follows:

(1)

Were U is voltage (V), φ is magnetic flux (Wb) and t is time (s). When the loop is short-circuited, the induced voltage U
will cause a current to flow that opposes its cause-the alternating magnetic field. This is the Faraday-Lenz law. If a
massive conductor (e.g. a cylinder) is placed in the alternating magnetic field instead of a short-circuited loop, eddy
currents (Foucault currents) will be induced (see Figure 1).

Figure 1: Faraday’s Induction law.

The eddy currents heat up the conductor according to the Joule effect. When a current I [A] flows through a conductor
with resistance R [O], the power is dissipated in the conductor. In most applications of induction heating, the
determination of resistance R is not a simple matter due to the non-uniform distribution of current in the conductor.

Haimbaugh stated that power requirements are related to the amount of energy required to heat a workpiece and to
the induction heating system power losses. The energy or heat content required to heat the workpiece can be
852
calculated when the material, its specific heat, and the effective weight of material to be heated per hour are known.
Higher power densities provide the ability to heat surfaces more rapidly. However, there may be limitations to the
amount of power that an individual induction coil can handle (1). Power induced in the workpiece can be determined
as:

P = U . I = R . I2 (2)

Where P is the power (kW). In literature there are several studies focused on heating of metal. In this study, since our scope is
the heating of steels for hot forging, we allowed for data of the papers that are related to our subject.

Table 1. Power consumptions of various metals for hot forging heating. (4)
Steel Aluminum Copper Brass
(CuZn 70/30)
Forging Temperature (ºC) 1250 500 900 750
Power Induced (kWh/ton) 240 136 105 90
Power Consumed (kWh/ton) 350 – 400 280 – 300 230 – 250 180 – 190
Frequency (Hz) 50 – 10.000 50 – 4.000 50 – 4.000 50 – 4.000

Table 2. Average Power Requirements For Induction Heat Processing Of Common Metals (5) (kWh/tons).
Process Carbon Magnetic Nonmagnetic Brass
Steel Stainless Stainless
Steel Steel

Hot Forging 440,9 413,4 474,0 440,9

Hardening/Aging 275,6 286,6 - 358,3

Annealing/Normalizing 248,0 231,5 413,4 413,4

Warm Forming 192,9 - 275,6 -

Stress Relieving 165,3 55,1 220,5 220,5

Tempering 77,2 77,2 110,2 -

Curing of Coatings 55,1 55,1 82,7 121,3

The efficiency should also consider .The electrical efficiency is defined as follows (3, 9):

(3)

Where e is the electrical efficiency, is the required energy for heating material (kW) and is the electric energy
consumed (kW).
In the Table 3 Laborelec (3) has given the efficiencies of induction systems as to frequency converters

Table 3. Induction installations general aspects with various frequency converters (3)
thyristors transistors tubes
Efficiency 90-97% 75-90% 55-70%
Frequency range: 100 Hz - 10 kHz up to 500 kHz up to 3000 kHz
Power range: up to 10 MW up to 500 kW up to 1200 kW

The data was provided from the workpiece that has widest diameter passing through the conductor. Efficiency
reduces with diameter reduction, by means of the increment of air gap between the workpiece and inductor. The
operating costs are compared in Table 4 by Poncin (4). He was indicated that, heating a steel workpiece to 1250 oC
consumes 350 kWh/ton. In the table it can be seen that, overall operating cost of a gas fired heating furnace is the
cheapest method.

In addition, installation costs of induction and gas fired furnaces were given by Mortland (5) in Table 5. In the
calculations, Annual Energy costs, scale losses, scrap losses, labor requirements were taken in to consideration. As
to the results ofMortland (5), the Total Annual Operating Cost of Induction Furnaces are more economical than the
Gas Fired Furnaces.

As it can be seen from Table 4 and Table 5 the results are not coherent. Thus, it must be noted that, these calculated
853
costs vary as to the place, plant and the operation.

Table 4. Actual heating costs of various processes given by Poncin (4)(Euro/ton).

Induction Gas LPG

Labor 3,85 7,7 7,7

Fuel 28,02 13,27 56,04
M&O 2,98 5,95 5,95
Scale Losses 2,23 11,15 11,15
Amortisation Costs 15,37 7,68 7,68
TOTAL 52,45 45,75 88,52

Table 5. Installation costs of induction and gas fired furnaces given by Mortland (5).
Item Induction Furnace Gas-Fired Furnace

Installed Cost $600,000 $200,000

Heating Efficiency 60% 15%
Annual Energy Cost $720,000 $540,000
Scale Loss 1/2% 2%
Scrap Loss 1/4% 1%
Annual Scrap and Scale Loss Cost $150,000 $600,000
Labor Requirement 1 Operator 2 Operators
1/4 Maintenance 1/2 Maintenance
Annual Labor Cost $60,000 $120,000
Total Annual Operating Cost $930,000 $1,230,000

The advantages of induction furnaces are aligned as follows (3-6)

Technical process
 Maintenance costs and spare parts costs are fair.
 High power density provides a compact installation and realize a quick heating.
 Floor space requirements are less.
 Induction offers the possibility of reaching very high temperatures
 Induction heating can be applied to specific area of workpiece
 Induction installations are suited for automation
 Recovers time and heat losses during feeding and receiving of work piece.
 No need to stock fuel.
 A significant portion of the heat losses can be recuperated
 Extreme purity is possible by working in a vacuum or in inert atmospheres
 The precise location of heating can be determined accurately
 The heating can be regulated precisely
 Environment and working conditions
 No production of flue gasses
 Induction installations generally have good efficiency although this efficiency also depends upon the
characteristics of the material to be heated

Stefan and Günter recommended a hybrid furnace system that includes both with natural gas fired and induction
furnaces. They advised that the material would heated up to 700-800 oC via the gas fired furnace and to 1200-1300
o
C via induction furnace to avoid scale formation (7).

NATURAL GAS FIRED FURNACES

Natural gas fired furnaces (NGF) are somewhat simpler than the induction furnaces. You can see and hear the
process in the furnace. The elements are familiar and known. These furnaces can be assumed as bigger scaled
furnaces of our domestic ones. Therefore, most of the systems which are used in NGF are conventional e.g. loading
system, walls, gas systems.

In NGF, main energy input is the combustion of natural gas. The electric, which is used to run the burners and air
fans, may be considered as auxiliary energy. The total energy input via natural gas combustion can be calculated as;

(kW) (4)

Where is the mass flow rate of natural gas (kg/s) and LHV is the Lovest Heating Value. Energy of flue gases may
be calculated as;

854
 (kW) (5)

The workpiece that is loaded to furnace may be at environment temperature or any pre-heat temperature. Even case
the temperature difference can be calculated as the difference between inlet and exit temperatures. Thus, the total
energy transferred to the work piece in the furnace can be calculated from;

(kW) (6)

Where is the mass of workpiece loaded in 1 hour (kg/h), Cw is the specific heat of work piece in (kJ/kg oC) and Tw
is the final temperature of the work piece (K). The radiation heat transferred from the hot surfaces simply can be
calculated from;

(kW) (7)

The heat transfer via turbulent flow natural convection from the hot surfaces can be calculated from;

(kW) (8)

The efficiency of a NGF can be obtained via direct and indirect methods. In the Indirect method, percentage of all
losses through the total energy input must be calculated. In the direct method, proportion of the heat transferred to the
workpiece to the total energy input via natural gas combustion can be calculated via the following equation.

(9)

Results and Discussion

The NGF considered in this study was installed 35 years ago, which has an old technology (Figure 1). The furnace
has no exhaust system and naturally, has no recuperator or regenerator. The flame was blow out from the openings
of the hatches. Thus, the thermal losses from the NFG were not based on only the radiation from the hot surfaces and
exhaust gases. The opening losses were highly effective because of more than 1300 oC flame leakage. That means,
the energy of flame with its radiation potential is inconsumable.

Figure 2 An operating Natural Gas Fired Furnace

In NGF systems, burners provide a stable combustion. By this means, it is easy to evaluate the volumetric or mass
flow rate of natural gas. When natural gas flow rate read, total energy consumption can be easily calculated via the
Eq. 4. In this study, Lowest Heating Value (LHV) of natural gas was taken from IZGAZ (Official Gas Distribution
Company).

855
Table 7. Design Data and Energy Analysis of NGF
Fuel Consumption per Unit Time (sm3/h) 49,5
Lowest Heating Value of Natural Gas (kWh/Sm3) 10,64
Energy Consumption per Unit Time (kWh/h) 527
Combustion Time to Attain Steady State (h) 3
Natural Gas Consumption Until Steady State (sm3/h) 149
Total Energy Consumption Until Steady State (kW) 1580

Energy Requirement of 1 Tons of Steel (Design Capacity) (kW/h) 181

Energy Requirement of Steel Forged in Unit Time (Actual Data) (kW/h) 30
Qex Exhaust Losses (kW) 162
Qhs Radiation Losses (kW) 25
Qn Natural Convection Losses (kW) 19
Qo Opening Losses (kW) 140
QT Total Heat Losses (kW) 205
Efficiency (Design Data) % 34,44
Efficiency (Actual Data) % 5,74

The results of the calculations and design data of the system, which was handled with in this study, were given in
Table 7. The exhaust losses were obtained via Eq. 5. It was considered that all the 1300 oC hot stack gases were lost
with their energy potential, since the NGF system has no exhaust, recuperator or regenerator. In the system, even
some of the flame was thrown out from the openings of the furnace. In most of the applications, the opening losses
obtained 5% of the total energy input. However, in this study it was calculated about 26% of total energy input. The
main reason of this is, not only the loss of high temperature energy potential of the exhaust gases but also the energy
potential of very high temperature flame and flames radiation potential. In this type of heating systems, the flame is
wanted to be kept inside the heating zone because of its very high temperature, which is useful for convection and
radiation. It is desired to finish the burning of natural gas inside the heating zone, but never outside. Because of this,
the opening losses had a high percentage through the total energy input.

Hot surface radiation losses were calculated with Eq. 7. The mean temperatures of surfaces of the NGF were
achieved via thermal camera (See Figure 3). In the system, the insulations were also not optimum. At some points on
the furnace, the wall temperature was reached to 200 oC. This wall temperature is definitely not allowable
temperature, because it means we are using the wall as a resistant heater to heat environment. Comparing to
opening losses, hot surface losses are not seem to be serious. Nevertheless, the wall temperature should still
reduced to To+40 oC.

Figure 3. Thermal camera view of a Natural Gas Fired Furnace

Natural convection losses were calculated with Eq. 8, which is given by Tütünoğlu (8). At first glance, the natural
convection may not seem to be high in proportion to other losses. However, if the dimensions of the furnace are
considered, we can conclude that it has to be reduced also. The high wall temperature causes high natural convection
losses as well as radiation losses.

856
Finally, the efficiency of both induction and NGF systems were calculated via Eq. (9). The data were collected from
meters for natural gas and electric. It must be noted that, the NGF system starts up 3 hours before shifts. Because of
this, additional operating hours of combustion system were considered. However, the effect of combustion to attain
steady state is various. It depends on many parameters such as, the difficulty level of the work piece, operating time,
production quantity, diameter of the work piece etc.

Table 8 Validation of Analysis and Comparison

Induction Ref (4) Induction Ref (5) Induction Analyzed NGF Analyzed
Total Consumption (kWh/ton) 375 441 755 527
Energy Requirement (Kwh) 188 188 299 30
Losses (kW) 187 253 361 722
Efficiency % 50,23 42,72 52,19 5,74
Heating Cost ($/ton) 52,08 61,24 69,91 158,94

In table 8 validation of the analysis results and comparison of the subject furnaces of this study were given. In the
analysis the result obtained from the induction furnace is in a good agreement with (4 and 5). Mainly efficiency agrees
with the literature. A slight difference may cause from Cos  of the systems, but NGF system has a dramatic
difference between induction furnace as well as the other natural gas fired furnaces given in literature. The most
important and basic reason is; the analyzed heating system has no exhaust. If there were, we could have a chance to
measure the combustion and recover the waste energy. The efficiency, which is given in Table 8, is the mean value of
calculation results. A detailed analysis of efficiency of NFG is given in Figure 4. As it is mentioned, the efficiency of
NGF depends on many parameters. If the order can be produced in 3 hours then the efficiency differs between 2,4-
9,5% as to the production mass in one shift. If the mass of the production in one shift increases, efficiency of the
system also increases. Let’s consider that 500 kg will be forged, when it is forged in 3 hours the efficiency of the
system calculated 2,4%. However, when production time extends to 8 hours, the efficiency of the system decreases to
1,5 %.

Figure 4 Efficiency of natural gas fired furnace considering load-per-shift and operating hours.

The efficiencies given in figure 4 are all very low. The system is old and ineffective. Because of this, a project proposal
was given to East Marmara Development Agency (MARKA) to develop the NGF system and it is accepted. In the
scope of the project, the system replacement of the old system with a modern NGF system was proposed. Taking into
account that the most efficiency detractive topics are exhaust losses and opening losses, the projected system will
have regenerators and economizers, tight lids. Since the most energy destructing section of the existing NGF are lids,
to avoid the opening losses and leakage of the flame, a useful lid system is designed. In the designed NGF, loading
system will minimize the load/unload losses as well as leakage. The wall insulation is considered to stabilize the wall
temperature at (To+40) oC, so the radiation and natural convection losses will also decrease. With the projected NGF,
efficiency is supposed to increase up to 40%.

It is also decided that the next project will be about hybrid system that heat the workpiece up to 900 oC by NGF and to
1300 oC by induction furnace.

Conclusion
In this study, a thermodynamic analysis has performed for two types of furnaces of a forging facility as well as recent
studies. The data are collected from an induction furnace and a natural gas fired furnace. It is seen that the efficiency
857
of induction furnace is about 52%, which agrees with the literature. However, the NGF is very ineffective since the
losses of the system are very high. Depending on the lot size and operating hours, the NGF efficiency was calculated
max. 9.5% and minimum 1.4%. It is identified that the highest energy losses are exhaust (162 kW) and opening
losses (140 kW) in order. In the current situation, heating the material with induction furnace is clearly profitable.
Finally, the inefficient NGF should be remade with current technology.

Nomenclature
U : Voltage (V)
φ : Magnetic flux (Wb)
t : Time (s)
P : Power (kW)
I : Current (a)
R : Resistance (O)
e : Electrical efficiency
: Required energy for heating workpiece (kW)
: Electric energy consumed (kW)
NGF : Natural Gas Fired Furnace
: Total energy input via combustion of natural gas (kW)
: Mass flow rate (kg/s)
LHV : Lowest Heating Value (kW/kg)
: Exhaust losses (kW)
: Exhaust flow rate (kg/s)
: Specific heat of exhaust gasses (Kj/kgK)
: Exhaust temperature (C)
: Atmospheric temperature (C)
: Total energy transfer to the work piece (kW)
: Mass heated per unit time (kg/s)
: Specific heat of work piece (Kj/kgK)
: Temperature of the work piece (C)
: Radiation heat transferred from the hot surfaces (kW)
: Stefan Boltzmann Constant (Kw/m2K)
ε : Emmisivity
A : Area (m2)
: Surface temperature (C)
: Heat transfer due to Natural Convection (kW)
: Overall efficiency

References
1. Haimbaugh, Richard E. 2001. Practical Induction Heat Treating, ASM International, ISBN: 978-0-87170-743-7
2. Kranjc M Et. Al. Numerical analysis and thermographic investigation of induction heating, 2010. International
Journal of Heat and Mass Transfer 53:3585–3591
3. Laborelec JC. 2011. Inductıon Heatıng, Leonardo Energy Applıcatıon Note, Publication No 0123 December 2011.
ECI Available at: www.leonardo-energy.org/drupal/node/1693
4. Poncin R., (Translator : Sabahattin ÖZEN) 2004. İndüksiyonla Takoz Isıtma, TMMOB Metalürji Mühendisleri
Odası Dergisi 138:49-52 Haziran.
5. Mortland, Jack. 1993. A Cost-Effective And Energy-Efficient Method For Process Heating. TechCommentary
Vol. 2, No. 1 EPRİ Center for Materials Fabrication.
6. DaWei Induction Heating Machine Co.,Ltd electronic publication available @: http://www.induction-
heating.com.cn/uploadfile/20121008/20121008214904637.pdf
7. Beer, S. and Günter U. 2012. Combined Log Heating Furnace Inlineconcept For Aluminyum Extruders,
Induction Technolohy. Reports, Heat Processing. 2:77-80
8. Tütünoğlu Y, Güven A. and Öztürk İ.T. 2011. Cam Temperleme Fırınında Enerji Analizi, III. Enerji Verimliliği
Kongresi Bildirileri Kocaeli

858
 THE USE OF HAZELNUT OIL ETHYL ESTER AS A FUEL
IN PRE-CHAMBER DIESEL ENGINE
a b
Murat Cetin Ahmet Tandiroglu
a
Department of Engine Vehicles and Transportation Technologies,
b
Department of Mechanical Engineering Technology
Vocational High School, Erzincan University,24100 Erzincan, Turkey
mcetin@erzincan.edu.tr, atandiroglu@erzincan.edu.tr

ABSTRACT
In this study, briefly reviews the use of hazelnut oil ethyl ester which was produced by hazelnut oil as an
alternative fuel in pre-chamber diesel engines, and compares it with pure hazelnut oil and diesel fuel. In the
experiments, a naturally aspirated, pre-chamber indirect injection, four cylinders, four stroke diesel engine was
used. During the test, power, torque, emissions values and fuel consumption were measured with respect to the
engine speed. As a viscosity and density of hazelnut oil ethyl ester was found to closely similar to those of diesel
fuels. Test results obtained from hazelnut oil ethyl ester were compared to hazelnut oil and standard diesel fuel
test results. CO, NOx, smoke density of hazelnut fuel and exhaust temperatures were lower than of diesel fuel.
The biodiesel NOx emissions higher than diesel fuel all engine rpm and NOx emission was measured 15% higher
from the diesel fuel. The specific fuel consumption increased, at low-speed 6%, medium speed 9%, high-speed
12%. The average increase in specific fuel consumption 9.7% determined. As a result of this study; it is concluded
that hazelnut oil biodiesel could be used as an alternative fuel in diesel engines.

Keywords: Hazelnut oil; Hazelnut oil ethyl ester; Alternative fuel; Combustion; Diesel engine

INTRODUCTION
Diesel fuel is largely consumed by the transportation sector. Diesel engines are usually employed by heavy duty
vehicles used for transportation in the world and contribute to the prosperity of the worldwide economy. Due to
increased environmental problems and fast depletion of fossil fuel resources is demanding an urgent need to carry
out research work to find out the viable alternative fuels and sustainable energy systems. Studies on internal
combustion engines have been recently concentrated on alternative fuels. In general, they agree that of plenty oil
will be used as an energy carrier. Over the past decades, since the invention of internal combustion engines, the
development of integral combustion engines has been based on the availability of petroleum derived fuel, which in
turn has been tailored to meet the needs of current engines. However, alternative fuels for diesel engines have
become increasingly important due to the gradual depletion of world petroleum reserves, increasing fuel prices
and the impact of environmental pollution due to increasing exhaust emissions and several socioeconomic
aspects. As public concern about environmental pollution and energy security increases, alternative diesel fuels,
such as biodiesel, diesel and ethanol–diesel fuel blends, are receiving more attention. The increase in alternative
fuel investigations is caused by two main factors; a rapid decrease in world petroleum reserves and important
environmental concerns originating from exhaust emissions. Besides them, some other aspects such as
increasing fuel price, the idea of supplying the fuel demand from local sources, for reduce the import of the crude
oil and creating new employments have been promoted these investigations. A number of researchers have
investigated the effect of using many different kinds of biodiesel and biodiesel blended with ethanol or methanol on
using in diesel engines (Karaosmanoglu, et al., 2000. Cetin et al., 2007). The fuels produced by transesterification
of the oils are called biodiesel. The advantages of biodiesels as diesel fuel are the minimal sulfur and aromatic
hydrocarbons content, higher flash point, lubricity, cetane number, biodegradability and not contain any metals or
crude oil residues and non-toxicity. Although biodiesel cannot entirely replace petroleum-based fuels, biofuels and
diesel fuel blends can be used on existing engines to achieve both environmental and energy benefits. Biodiesel
has properties similar to those of traditional fossil diesel fuel such that it can be substituted for diesel fuel with little
or no engine modification. Studies clearly indicate that the use of biodiesel may potentially reduce the dependence
on petroleum diesel fuel and improve air quality. Substantial reduction in particulate emissions can be obtained
through the addition of biodiesel to diesel fuel. Vegetable oils, the main source of biodiesel, have considerably
higher viscosity and density compared to diesel fuel. Despite transesterification process, which has a decreasing
effect on the viscosity of vegetable oil, it is known that biodiesel still has some higher viscosity and density when
compared with diesel fuel (Huzayyin, et al., 2004.Kalam et al., 2003). The viscosities of fuels have important
effects on fuel droplet formation, atomization, vaporization and fuel/air mixing process, thus influencing the
exhaust emissions and performance parameters of the engine. The vegetable oil ester-based biodiesel has long
been used as fuel for diesel engines. Many studies about the use of biodiesel fuels in diesel engines have been
done and some of them have been reviewed. Studies clearly indicate that the use of biodiesel may potentially
reduce the dependence on diesel fuel and improve air quality. Substantial reduction in particulate emissions can
be obtained through the addition of biodiesel to diesel fuel. B20(a mixture of 20% biodiesel and 80% diesel) has

859
become the most popular biodiesel fuel blend used and this blend level has been studied in different countries.
Biodiesel blended with ethanol the performance and emissions on and it found reduction of CO and NOx
emissions using 20% blended fuel but an increase in HC emission. Compared the emissions of biodiesel and
biodiesel with emulsified and fumigated methanol and found that the disadvantage of increased NOx emission of
biodiesel. Compared to fossil diesel fuel, biodiesel has several superior combustion characteristics. Biodiesel fuel
can effectively reduce engine-out emissions of particulate matter, CO, and hydrocarbons in modern four-stroke
compression–ignition engines. The exhaust emissions of diesel engines fuelled with biodiesel or its blends with
diesel fuel have been studied by many investigators. It has been usually reported that there are reductions in CO,
HC and smoke emissions while NOx emissions increase. It has been shown that the oxygen content in biodiesel is
the main factor for reducing pollutant emissions and increasing NOx as a result of better combustion (Karabektas
et al., 2008. Murugesan et al., 2009. Mbarawa., 2010. Buyukkaya., 2010).

Turkey's energy demand has risen rapidly as a result of economic and social development over the past two
decades. As in many other countries, Turkey is dependent on fossil fuels to meet its energy requirements. Fossil
fuels account for approximately 88% of the country's total primary energy consumption. Turkey imports three major
sources of energy, and its dependence on imported fossil fuels is expected to increase even further. At present,
Turkey's oil production met only 10% of demand; the remaining 90% is being imported every year. In spite of
Turkey's heavy dependence on fossil fuels for energy demand, the country has a large potential for development
of renewable resources of every type. Bio-fuels can provide an opportunity for Turkey to decrease its dependence
on foreign oil, eliminate irregularities in agriculture, create new employment opportunities, decrease rural
depopulation, and sustainable energy development. In the 2020s, it is estimated that gasoline consumption will
remain at its current levels of around 2.5 million tones. However, it is likely that diesel consumption will rise from
13.9 million tonnes (in 2010) to 18.3 million tones (Erkut., 2010). Turkey has a large area of suitable agricultural
land for the production of bio-fuel crops. Turkey's bio-diesel production is only around 60000 tons per year. In this
study, this main idea has been taken as the research aim and the evaluation of the direct use of hazelnut oil and
hazelnut oil biodiesel as a diesel fuel has been investigated and compared with that of diesel fuel. Hazelnut is
especially important in Turkey, because Turkey is the biggest hazelnut producer country in the world and hazelnut
oil was selected for the study because of its wide use in Turkey. Yields of 2200 kg/hectare annually are common
with good cultural practice, or 20 barrels of petroleum equivalent per hectare-year (Cetin et al., 2007). In Turkey,
studies about the use of hazelnut oil biodiesel fuels in diesel engines have been done. The hazelnut oil methyl
ester was tested in a four cylinder, turbocharged, direct injection diesel engine and test results compared to
standard diesel fuel test results. As a test results, hazelnut oil methyl ester are very close to diesel fuel and CO,
NOx, smoke density of hazelnut oil methyl ester and exhaust temperatures were lower than of diesel fuel (İçingür
et al., 2006). Biodiesel from a hazelnut soapstock/waste sunflower oil mixture was investigated the effects of the
methyl ester addition in different proportions to Diesel No.2. The CO emissions of the blend were higher at low
speed and lower at high speeds, and higher CO2 emissions in the experimental speed range and NOx emissions
slightly increased due to the higher combustion temperature and the presence of fuel oxygen with the blend at full
load were observed (Usta et al., 2005).

OBTAINING OF HAZELNUT OİL ETHYL ESTER

The transesterification is a widely applied, convenient and most promising method for reduction of viscosity and
density of vegetable oils. A detailed description of the transesterification process can be found in the literature.
Although neighbor and boundary to Europe a limited amount of biodiesel is produced from rapeseed oil in Turkey,
cottonseed oil offers a great potential as the source of biodiesel because cotton is the second produced seed plant
after sunflower in terms of the cultivated area (Karaosmanoglu et al., 2000. Usta et al., 2005). Beginning of the
study; the natural hazelnut oil used as a fuel and it was produced by The Union of Hazelnut Sales Cooperative.
Transesterification is an efficient method to decrease and convert high viscosity vegetable oil into a fuel with
chemical properties similar to those of diesel fuel. In the present study, base catalyzed transesterification NaOH is
used to prepare biodiesel from hazelnut oil. The production of hazelnut oil ethyl ester was made in Erzincan
University Vocational High School, Department of Engine Vehicles and Transportation Technologies, Automotive
Technology Program of diesel engines laboratory by using biodiesel production system. The transesterification
process of hazelnut oil was performed using 1.25 liter ethyl alcohol and 50 grams sodium hydroxide as catalyst per
5 liter pure refined hazelnut oil. Biodiesel production process, purity of 99.8% sodium hydroxide, 0.910 kg/l density
of pure refined hazelnut oil, purity 99.5%, and 99.8% purity, 0.79 kg/l density of ethanol was used. First, the
hazelnut oil was heated to about 65 ºC in a reactor with a capacity of about 8 liter. Ethanol and catalyst were
mixed in a separate glass containers. Then, the catalyst and ethanol mixture were to dissolve and added to the
heated hazelnut oil in the reactor. The mixture was stirred for 1 hour at a fixed temperature of about 65 ºC to
esterification reaction. After esterification process; the glycerol layer was settled down, the ester layer formed at
the upper part of the container and the products which are divided in separate containers. A temperature of 65±1
ºC was maintained for 4 hours, and the reaction products were allowed to settle under gravity for 6 hours in a

860
separating funnel. After the production; washing of the biodiesel thus produced is essential for removal of the
impurities, water and the residual catalyst, which may be harmful for combustion engines. Produced biodiesel was
heated at 100 ºC, for removing water contained in the esterified hazelnut oil. The mixture was again allowed to
settle under gravity for 6 hours, and the lower layer of water containing impurities was drained out. Finally, the
produced hazelnut oil biodiesel was left to cool down. The production process of is presented schematically in Fig.
1 and the overall reaction is represented in Figure 2.

Figure 1. The flow chart of the biodiesel production process

C17H33COOCH2 HOCH2
Catalyst
C17H33COOCH + 3C2H5OH HOCH + 3C17H33COOC2H5

C17H33COOCH2 HOCH2
Triolein Ethanol Glycerol C18 Ethylester
(Biodiesel)
Figure 2. The transesterification of triolein (glycerol trioleate) to C18 ethyl esters (Li et al., 2009).

Produced hazelnut oil biodiesel and hazelnut oil were analyzed at the laboratory of chemical and used fuels
properties are given as a comparison of in Table1. As shown in Table 1, there is significant difference for
specific gravity and heating value between vegetable oils and diesel fuel. Viscosity of hazelnut oil is more than
11 times than that of diesel fuel and hazelnut oil has about 27% lower heating value. Viscosity of hazelnut oil
biodiesel is more than 1.75 times than that of diesel fuel and hazelnut oil biodiesel has about 14% lower
heating value. From the technological point of view, the fuel property of hazelnut oil and hazelnut oil biodiesel
seems to meet the fundamental requirements of a diesel engine. Therefore, use of hazelnut oil and hazelnut
oil biodiesel as a fuel substitute for conventional diesel fuel in a diesel engine is reasonable and prospective.

Table 1. Property of Hazelnut oil biodiesel, hazelnut oil and diesel fuel
Property Hazelnut oil biodiesel Hazelnut oil Diesel oil
Density (kg/L) 0.82 0.91 0.86
Flash point (°C) 50-170 245 60
Cetane 45-50 41–43 50 min
Sulfur (%) 0.00-0.0024 0.030 0.050
Heating value (kJ/kg) 38724 33028 45390
Colour Light green Amber yellow –
Kinematic viscosity (mm2 s-1) 1.9-5.6 30.97 2.9
Water and sediments (%) 0 0 Trace
Ash (%) 0.01-0.02 0.004 0.001
Density (kg/L) 0.82 0.91 0.86
Flash point (°C) 50-170 245 60
Cetane 45-50 41–43 50 min

861
EXPERIMENTAL STUDY
The layout of the experimental setup is shown in Fig.3. The experimental setup consists of a naturally
aspirated IDI diesel engine, an engine test bed, a gas analyzers and thermocouples. The experiments were
performed with a 2403 cm3 displacement, four cylinders, four-stroke, water cooled, 21:1 compression ratio pre-
chamber diesel engine of Mercedes-Benz OM 616. The maximum torque was 138 Nm at 2400 rpm, and the
maximum engine power was 44 kW at 4200 rpm. For such a proposal, at the conventional diesel engine not
modification of structure is unnecessary as has been confirmed. A hydrokinetic dynamometer was coupled to
the engine and the engine’s power was measured. A simple mechanism has been attached to the engine
speed by controlling the throttle position, and the engine was examined at partial and full loads. In order to
satisfy the standard conditions for each fuel at the experimental engine, it was possible to keep constant test
process operating parameters of the engine circuits, the oil and cooling water temperatures, which can
influence mainly the exhaust emission levels, thus increasing the credibility for comparing the measured
values. The same procedure was repeated for three type fuels by keeping the same engine operating
conditions. The differences in the measured performance and exhaust emission parameters from the baseline
operation of the engine, when working with conventional diesel fuel, hazelnut oil and hazelnut oil biodiesel
were determined and compared. The series of tests were conducted using of 3 different type fuels, with the
engine working at a different speed, and the static injection timing of was kept at 24o crank angle before top
dead center. The fuel piping system allows three separate pipe lines for diesel fuel; hazelnut oil and hazelnut
oil biodiesel in order to avoid mixing between the tested fuels and the engine was run for about 20 minutes to
stabilize at its new condition in each fuel change. Beginning of the study; the investigation effort was directed
to study the engine performance, emission characteristics of diesel engine using diesel fuel. In the second and
third phases of the study; the investigation effort was directed to study the engine performance, emission
characteristics of diesel engine using diesel fuel and 100% refined hazelnut oil and hazelnut oil biodiesel. K
type thermocouples and sensors are installed in the system in order to measure the parameters: pressure and
temperature of lubricant oil, fuel consumption, fuel temperature, temperature of cooling water, pressure, speed
and temperature of intake air, exhaust gas temperature, ambient temperature, atmospheric pressure and
humidity. A Bosch ETT008.55EU model and Bosch BEA370 gas analyzer were used to measure the CO, CO2
in percentages by volume and NOx and HC ppm by volume. Smoke was measured with Bosch 3.011 exhaust-
gas measuring module for diesel engine emissions. The initial data were taken after the stabilization of the
engine in running. The results of this study that has been performed with pre-combustion chamber diesel
engine have indicated that hazelnut oil and hazelnut oil biodiesel are employed as a substitute for diesel fuel.

Fig. 3. The Experimental apparatus.

CALCULATION OF THE ENGINE PERFORMANCE PARAMETERS

The experiments were performed without any modification on the engine. For each fuel testing, the volumetric
flow rates of air and fuels were measured and based on which the mass consumption rates of the fuel were
calculated. The air/fuel ratio was calculated, based on the stoichiometric air/fuel ratio and the actual air/fuel
ratio. During the experimental study, indicated performance parameters, such as the brake power (Pe), the
brake specific fuel consumption(be), the brake thermal efficiency (bte), the brake mean effective
pressure(bmep) can be determined by using the following equations respectively, according to the engine
torque(Me), the engine speed(ne) and the mass consumption rate(ṁf) of the fuel. The brake power, Pe (in kW)

Me  ne
Pe  (1)
9549
The brake mean effective pressure, bmep(in bar),

862
bmep  4Me / VH   105 (2)
3
As known, VH( in m ) the engine total (i.e. of all four cylinders) displacement volume. The brake specific fuel
consumption, be (in g/kWh)
.
B f  m f . f .103.3600 / Pe (3)

3600
be  (4)
H u .e

where be is the ṁf is the mass consumption rate of the fuels in kg/s; Hu is the lower heating value of the fuels
in kJ/kg and is a constant which depends on the kind of the fuel. The primary evaluation calculated for pre-
chamber diesel engine burning of diesel fuel, hazelnut oil and hazelnut oil biodiesel is its thermal efficiency ( e
). The relationship between thermal efficiency (  ) and specific fuel consumption (be) is represented by
e
equations 3-4. For make easy and understand, we used the specific fuel consumption as an estimate of the
thermal efficiency and, on account of the inverse relationship between them, a decrease in specific fuel
consumption would reflect an increase in thermal efficiency. Thus, in the following tests, specific fuel
consumption will be taken as the main evaluation parameter.

RESULTS
The engine performance at various loads using diesel fuel, hazelnut oil and hazelnut oil biodiesel was
observed in terms of brake power (Pe), brake specific fuel consumption (bsfc), and exhaust emissions (CO,
CO2, HC, NOx and O2). The viscosities of biodiesel fuels have important effects on fuel droplet formation,
atomization, vaporization and fuel/air mixing process, thus influencing the exhaust emissions and performance
parameters of the engine.

Engine performance
The results have been compared with the basic engine performance using 100% diesel fuel at different load
conditions over the engine speed range. For three test fuels, the torque variations with engine speed were
show with Fig. 4(a), the maximum torque values were observed around 2250 rpm of engine, for all test fuels
similar to the torque values, the torque for diesel fuel was higher than hazelnut oil and hazelnut oil biodiesel
fuels. Fig. 4(b) shows that the values of brake power variations of the test engine for three test fuels. It is clear
from the Fig. 4(b) that the power of diesel fuel was found higher than hazelnut oil and hazelnut oil biodiesel
fuels. It can be observed that the engine torque and power increase with increase in engine speed. Using
100% hazelnut oil and hazelnut oil biodiesel, for all of engine load and speed conditions, the engine power and
torque are generally slightly lower than the corresponding values using diesel fuel. There are two reasons for
taking this result: firstly the higher calorific value of diesel fuel is higher than hazelnut oil and hazelnut oil
biodiesel. Secondly the viscosity of hazelnut oil and hazelnut oil biodiesel is higher than diesel fuel and
combustion has been affected and deteriorated from the viscosity of hazelnut oil and hazelnut oil biodiesel.
The maximum torque values for diesel fuel 138 Nm, for hazelnut oil128 Nm and for hazelnut oil biodiesel 133
Nm were measured respectively at 2155 rpm. The lowest torque value was obtained for hazelnut oil fuel for all
speeds of engine. The maximum engine power values for diesel fuel 43 kW, for hazelnut oil 37.4 kW and for
hazelnut oil biodiesel 41 kW were measured respectively at 4200 rpm. The lowest power value was obtained
for hazelnut oil fuel for all speeds of engine.

150 50
Engine torque (Nm)

40
Engine power(kW)

120
30
90 Diesel fuel
Diesel fuel
Hazelnut oil 20
Hazelnut oil
Hazelnut biodiesel Hazelnut biodiesel
60 10
1000 2000 3000 4000 rpm 5000 1000 2000 3000 4000 rpm 5000
Fig. 4(a). Variation of engine torque and (b). engine power parameters versus engine speed for fuels.

Specific fuel consumption

Fig. 4(c) shows that the fuel consumption rates of the three fuels, the specific fuel consumption(be) curves
were optioned for best torque all engine speeds. As shown the curves; the specific fuel consumption(be) of for
the three fuels decreased slightly with the increase in to medium engine speed and increase in engine torque.

863
Hazelnut oil and hazelnut oil biodiesel consumption increased roughly in proportion with engine speed under
engine loads conditions. This is primarily due to the fact that diesel has the highest heating value from than
other two fuels, and needed the lowest fuel consumption rate for achieving the same engine brake horsepower
output as the other fuels. Hazelnut oil biodiesel consumed less than hazelnut oil in attaining the same engine
horsepower output, for that its heating value was higher than hazelnut oil. Depending on the high engine
speeds, the friction power increased according to the mechanical efficiency dropped and the engine torque
output decreased, this situation leading to an increase in the fuel consumption rate for all fuels. Results from
the tests of the engine fuelled by diesel fuel, hazelnut oil and hazelnut oil biodiesel indicated that specific fuel
consumption of the engine fuelled by hazelnut oil is approximately 15% higher and hazelnut oil biodiesel is
approximately 10% than that diesel fuel. The increase of the specific fuel consumption(be) can be explained
by the lower heat values of hazelnut oil and hazelnut oil biodiesel compared with diesel fuel. For the lower heat
values of hazelnut oil and hazelnut oil biodiesel more fuel has to be injected to obtain the same power output.

375
Diesel fuel
325 Hazelnut oil
Hazelnut biodiesel
be(g/kWh)

275

225

175
1000 2000 3000 4000 rpm 5000

Fig. 4(c). Variations of specific fuel consumption parameters versus engine speed for fuels.

Exhaust gas emissions

Fig. 4(d) shows the CO emissions for the two alternative fuels under different engine speeds. The minimum
CO emissions were obtained with hazelnut oil biodiesel and the maximum CO emissions were obtained with
hazelnut oil. The diesel fuel CO emission is between from hazelnut oil biodiesel and hazelnut oil. Due to the
high viscosity, the air/fuel mixing process is affected by the difficulty in atomization of the hazelnut oil. The
higher of load, the richer of fuel mixture burned, and thus, more CO is produced. This situation is typical of
behavior for a homogenous charge engine as CO emission is usually affected by fuel viscosity and
combustion quality. The increase in the CO levels is same for the same increase of the supplied hazelnut oil
biodiesel, diesel fuel and hazelnut oil at maximum engine rpm region. In addition, hazelnut oil biodiesel were
found to emit significantly lower CO concentrations compared with diesel fuel and the hazelnut oil. The
hazelnut oil and hazelnut oil biodiesel are contain 10% oxygen on a mass base and have lower carbon content
than the diesel fuel. Because of the oxygen content in the biodiesel enhances, thereby reducing the formation
of CO emission in combustion process. Therefore, the hazelnut oil biodiesel can be used as a clean, lower
carbon alternative fuel to replace fossil fuel for diesel engines. As it is clear from the figure that the curves of
CO emissions for hazelnut oil biodiesel remain under the curve of pure diesel but increase on use of the
hazelnut oil all engine speeds.

0,25
Diesel fuel
Hazelnut oil
0,2 Hazelnut biodiesel
CO(volume %)

0,15

0,1
1000 2000 3000 4000 rpm 5000
Fig. 4(d). Variations CO emissions parameters versus engine speed for fuels.

Fig. 4(e) shows that the HC emissions of for all fuels are lower in partial load and decreases at medium engine
speed but tend to increases at higher engine speed and engine loads. The emissions of unburned
hydrocarbons HC for hazelnut oil biodiesel higher than diesel fuel at all engine speeds. HC emissions in
hazelnut oil and hazelnut oil biodiesel are higher than the diesel fuel. The high viscosity of hazelnut oil and
hazelnut oil biodiesel were caused to deterioration, the combustion tend to generate high levels of unburned

864
hydrocarbons. Factors causing combustion deterioration (such as high latent heats of vaporization) could be
responsible for the increased HC production. Although the CO and HC emission of hazelnut oil was higher
than diesel fuel, but the HC emission of hazelnut oil biodiesel was higher and the CO emission was lower than
diesel fuel. This is because of better combustion of hazelnut oil biodiesel inside the combustion chamber due
to oxygen content in biodiesel.

70
Diesel fuel
68
Hazelnut oil
66 Hazelnut biodiesel
HC (ppm)

64
62
60
58
56
1000 2000 3000 4000 rpm 5000
Fig. 4(e). Variations HC emissions versus engine speed for fuels.

Fig. 4(f) shows O2 concentrations respectively, for diesel fuel, hazelnut oil and hazelnut oil biodiesel. The
hazelnut oil and hazelnut oil biodiesel O2 emissions values are show that almost the same when operating with
diesel fuel. It can be seen that oxygen concentration available for reactions decrease the engine speed and O2
emissions are higher than diesel at all speed range. It can be said that O2 concentration was high in the
exhaust gazes and similar for fuels. This high oxygen concentration could be explained as due to the oxygen
content of hazelnut oil biodiesel and hazelnut oil and excessive air in the combustion. This is due to the lack of
oxygen resulting from the operation at higher air/fuel ratios. O2 emissions and CO2 emissions are opposite
situation to as shown in Fig. 4(f) and Fig 4(g). This situation can be explained by the detailed chemical
structures of hazelnut oil, hazelnut oil ester and combustion reactions.

18
Diesel fuel
17 Hazelnut oil
O2 (volume%)

16 Hazelnut biodiesel

15
14
13
12
1000 2000 3000 4000 rpm 5000

Fig. 4(f). Variations O2 concentrations parameters versus engine speed for fuels.

Fig. 4(g) shows CO2 emissions, respectively, for all test fuels under varied engine speeds. The CO 2 emission
is increased with increasing engine speed for three fuels. Diesel fuel produced the highest amount of CO 2 than
the hazelnut oil and hazelnut oil biodiesel fuels for all engine speeds. The hazelnut oil, hazelnut oil biodiesel
are lower CO2 emission values indices than diesel fuel. The reason for this, the fact that hazelnut oil and
hazelnut biodiesel are a low carbon fuel and has a lower carbon/hydrogen ratio than diesel fuel. This is
probably due to the fact that this hazelnut oil, hazelnut oil biodiesel has the lowest air/fuel ratio and also the
fact that CO2 emission concentration decreases with the decrease of the air/fuel ratio. In addition, the carbon
contents of diesel fuel higher than hazelnut oil and hazelnut oil biodiesel. The burning of the hazelnut oil and
hazelnut oil biodiesel appeared to emit the lowest CO2 emission, primarily owing to its having the lowest
carbon content from diesel fuel. Generally, the amount of carbon dioxide is proportional to the amount of fuel
burned and the CO2 emissions increase with increasing engine speed and the mass of fuel injected using the
fuel. It is an indication of efficient combustion of biodiesel due to its oxygenated nature which helps for more
complete combustion.

865
 8
7
CO2 (Volume%)

6
5
4 Diesel fuel
3 Hazelnut oil
Hazelnut biodiesel
2
1000 2000 3000 4000 rpm 5000
Fig. 4(g). Variations CO2 emissions versus engine speed for fuels.

Fig. 4(h) shows the smoke values versus to engine rpm for all the used fuels. According to the Fig. 4(h), the
hazelnut oil and hazelnut oil biodiesel are produces less smoke than the diesel fuel at 2000 rpm engine
speeds of approximately. The smokes for hazelnut oil and hazelnut oil biodiesel are lower than that of the
diesel fuel under 2000 rpm engine speed and lesser amount of unburned hydrocarbons measured in the
engine exhausts gases. But in higher speed, hazelnut oil and hazelnut oil biodiesel are produced significantly
higher smoke than diesel fuel. It explained that the hazelnut oil viscosity and the hazelnut oil biodiesel viscosity
are higher than diesel fuel and this situation caused to deterioration of combustion and incomplete
combustion.

3
2,5
Smoke (BSU 1/m)

2
1,5
1 Diesel fuel
0,5 Hazelnut oil
Hazelnut biodiesel
0
1000 2000 3000 4000 rpm 5000
Fig. 4(h). Variations smoke concentrations versus engine speed for fuels

Fig. 4(i) shows, the NOx emissions for the all fuels under varied engine speeds and engine torques. The
hazelnut oil biodiesel and hazelnut oil were produced a higher amount of NOx, while diesel fuel produced the
low amount of NOx emissions at these engine speeds. The NOx emission values were observed to decrease
with the increase in engine speed for all fuels. The formation of NOx depends on a large extend on the flame
temperature and high compression ratio. The NOx emission decreases with the decrease in the fuel to air ratio
under fuel-lean burning conditions of high engine speed. The NOx emissions with hazelnut oil biodiesel
increase approximately 12-18% as compared to diesel fuel for all engine speed. The NOx emissions with
hazelnut oil biodiesel higher approximately 8-10% as compared to hazelnut oil and the NOx emissions with
hazelnut oil higher approximately 10-14% as compared to diesel fuel. This situation is probably, the relatively
higher oxygen content of hazelnut oil biodiesel, which produced a higher NOx formation from pure hazelnut oil
and diesel fuel. The increase in NOx emissions may be attributed to various reasons, such as improved fuel
spray characteristics, high viscosity, and better combustion of biodiesel due to its high oxygen content and
higher. Finally, it is seen that the hazelnut oil biodiesel and hazelnut oil operations usually yield higher NOx
emissions at all engine test speeds compared to diesel fuel operations.

866
 500
Diesel fuel
450
400
Hazelnut oil

350 Hazelnut biodiesel

NOx(ppm)

300
250
200
150
1000 2000 3000 4000 rpm 5000

Fig. 4(i). Variations NOx emissions versus engine speed for fuels

Fig. 4(j) shows the exhaust gas temperatures changes depending on engine rpm for different fuels. It can be
seen that hazelnut oil biodiesel and hazelnut oil exhaust gas temperatures lower than diesel fuel for all of the
engine speeds. The highest value of exhaust gas temperature was observed at 4350 rpm with diesel fuel
(555oC) at full load, whereas the corresponding value with the hazelnut oil biodiesel was measured to 515 oC.
The increase in exhaust gas temperature with using hazelnut oil reached maximum at 465oC. The reason for
this change in the exhaust gas temperature, hazelnut oil and hazelnut oil biodiesel are low calorific value than
diesel fuel and the differences of the cetane number. It can be said that depends on the high cetane number of
biodiesel, the ignition delay shortening and as a result of the final temperature of combustion temperature is
lower than diesel fuel combustion.

600
Exhaust gas temperat.C

500

400

Diesel fuel
300
Hazelnut oil
Hazelnut biodiesel
200
1000 2000 3000 4000 rpm 5000
Fig.4(j). Variations the exhaust gas temperature versus engine speed for fuels

CONCLUSIONS
The experiments are conducted for diesel fuel and pure hazelnut oil and hazelnut oil biodiesel(hazelnut oil
ethyl ester) for different fuels at various loads. The following conclusions may be drawn from the present
study:
-It was shown that hazelnut oil biodiesel as an alternative fuel can be used successfully to run pre-chamber
diesel engine without modifications to the engine and subsystems.
-Specific fuel consumption with hazelnut oil biodiesel is higher than diesel fuel. This is attributed to the lower
heating value of the hazelnut oil biodiesel.
-Hazelnut oil biodiesel emissions are higher (HC, NOx and smoke) except for CO and CO2 when compared
with diesel fuel emissions.
-The increase in NOx emissions was proportional to the amount of biodiesel. In the case of biodiesel, the
increase in NOx emission was 15% compared to the diesel fuel, respectively.
-Hazelnut oil biodiesel physical and chemical properties are similar, and hazelnut oil biodiesel heat value close
to diesel fuel, but it is lower.
-In the specific fuel consumption increased, at low-speed 6%, medium speed 9%, high-speed 12%. Specific
fuel consumption increased by an average rate of 9.7% was determined.
-Hazelnut oil biodiesel is more expensive than diesel fuel at the moment. If hazelnut is not used as a food, with
local availability of the fuel it is favorable to adopt in the economic point of view.

Based on these results and the local availability of hazelnut oil biodiesel, it may be concluded that the fuel
derived from hazelnut is a potential candidate for internal combustion engines.

867
Nomenclature
IDI Indirect diesel injection
% vol. Percentages of volume
NaOH Sodium hydroxide
CO Carbon monoxide
CO2 Carbon dioxide
HC Hydrocarbon
O2 Oxygen
NOx Nitrogen oxides
Kg Kilogram
PPM Parts per million
ne Engine rpm
Me Effective Torque, Nm
Pe Brake power kW
 Density, kg/m3
BSU Bosch Smoke Unit
BSU l/m
L Liter
ºC Degree Centigrade
Nm Newtonmeter
kW Kilowatt
rpm Revolution per minutes
be Specific fuel consumption, kg/kWh
bte Brake thermal efficiency,%
bmep Brake Mean Effective Pressure
ṁf Mass flow of fuel, kg
Hu Lower heating value, kJ/kg
e Thermal efficiency,%

REFERENCES

Buyukkaya E. 2010. Effects of biodiesel on a DI diesel engine performance, emission and combustion
characteristics. Fuel 89 3099–3105.

Cetin M. and F. Yuksel. 2007. The use of hazelnut oil as a fuel in pre-chamber diesel engine. Applied Thermal
Engineering 27; 63–67.

Erkut Y. 2010. TÜPRAŞ to invest $2 billion to increase diesel production, The Turkish Petroleum Refineries
Corporation (TÜPRAŞ).

Huzayyin A.S., A.H. Bawady, Rady M.A. and Dawood A. 2004. Experimental evaluation of diesel engine
performance and emission using blends of jojoba oil and diesel fuel. Energy Conversion and Management 45;
2093–2112.

İçingür Y. and Koçak M.S. 2006. A Study on the performance and emission parameters of hazelnut oil methyl
ester as an alternative diesel fuel. Journal of Polytechnic 9; 119-124.

Kalam, M.A., Husnawan M. and Masjuki H.H., 2003. Exhaust emissions and combustion evaluation of coconut
oil-powered indirect injection diesel engine. Renewable Energy 28; 2405–2415.

Karabektas M., Ergen G. and Hosoz M. 2008. The effects of preheated cottonseed oil methyl ester on the
performance and exhaust emissions of a diesel engine. Applied Thermal Engineering 28; 2136–2143.

Karaosmanoglu F., Kurt G. and Ozaktas T. 2000. Long term CI engine test of sunflower oil. Renewable Energy
19; 219–221.

Li E., Zhi Ping Xu and Victor Rudolph. 2009. MgCoAl–LDH derived heterogeneous catalysts for the ethanol
transesterification of canola oil to biodiesel. Applied Catalysis B: Environmental 88;1–2, 42–49.

Mbarawa M. 2010. The effect of clove oil and diesel fuel blends on the engine performance and exhaust
emissions of a compression-ignition engine. Biomass and Bioenergy 34; 1555-1561.

Murugesan A., Umarani C., Subramanian R. and Nedunchezhian N. 2009. Bio-diesel as an alternative fuel for
diesel engines-A review. Renewable and Sustainable Energy Reviews 13; 653–662.

Usta, N., Ozturk E., Can O., Conkur E.S., Nas S., Con A.H., Can A.C. and Topcu M. 2005. Combustion of
biodiesel fuel produced from hazelnut soapstock/waste sunflower oil mixture in a diesel engine. Energy
Conversion and Management 46; 741–755.

868
 DEVELOPMENT OF A REDUCED MECHANISM FOR N-HAPTANE FUEL IN HCCI ENGINES
1 2 3
Keyvan Bahlouli , R. Khoshbakhti Saray , Ugur Atikol

1
Mechanical Engineering Department, Faculty of Engineering, Eastern Mediterranean University,
Famagusta, N. Cyprus, Via Mersin 10, Turkey.
2
Faculty of Mechanical Engineering, Sahand University of Technology, Tabriz, Iran.
3
Mechanical Engineering Department, Faculty of Engineering, Eastern Mediterranean University,
Famagusta, N. Cyprus, Via Mersin 10, Turkey.

e-mails: Keyvan.bahlouli@emu.edu.tr, khoshbakhti@sut.ac.ir, ugur.atikol@emu.edu.tr

ABSTRACT
Due to the lack of direct control of combustion timing in the homogeneous charge compression ignition
engines the chemistry of the in-cylinder charge is responsible for the ignition. Thus, obtaining knowledge
about fuel chemistry has a significant importance in engine design and optimization. Large numbers of
detailed chemical kinetics mechanisms have been suggested to predict various fuels oxidation. However,
employing comprehensive chemical kinetics mechanisms in predictive models results in the high demand for
simulation time which makes the use of these mechanisms impractical. Consequently, reduced mechanisms
of smaller sizes are needed. The objective of this study is to produce reduced mechanism of n-heptane fuel,
to be applicable for CFD simulation, by utilizing a two-stage reduction process. This work is performed by
using a validated single zone combustion model. To remove unimportant species at the first stage, the
directed relation graph with error propagation (DRGEP) is applied. In the second stage, the principal
component analysis (PCA) method is used to eliminate insignificant reactions and species. Peak pressure,
maximum heat release and CA50 have been selected as representative parameters for the performance of
engine. For the generated reduced mechanism at each reduction step, these parameters would be
calculated and the deviations from the corresponding value obtained by applying detailed mechanism to the
model will be evaluated until user-specified error tolerances violate. This combination of two methods
successfully reduced the detailed Golovichev’s n-heptane mechanism (57 species and 290 reactions) to a
reduced mechanism with size of 40 species and 95 reactions, while maintaining small errors (less than 1
present) compared to the detailed mechanism.

1. INTRODUCTION
Reduction of transportation emissions and fuel consumption are important reasons motivating the
development of alternative energy technologies for the IC engines. Homogeneous charge compression
ignition (HCCI) combustion engine is a potential candidate for higher thermal efficiency and lower emissions
(Onishi et al. 1979, Najt and Foster 1983, Thring 1989). However, lack of direct control for combustion timing
is one of the main challenges associated with HCCI combustion engine application.
Accurate fuel oxidation chemistry models offer great potential for HCCI engine design and optimization. The
Golovichev’s mechanism (Golovitchev 2000), including 57 chemical species and 290 reactions, and more
comprehensive mechanism of Curran with 560 species and 2539 reactions (Curran et al. 1998), as
examples, were introduced for n-heptane combustion.
However, for developing predictive combustion models, incorporating detailed chemistry into computational
fluid dynamics (CFD) calculations is commonly known as essential. A large system of nonlinear stiff ordinary
differential equations (ODE) is produced by using detailed chemical kinetics mechanism. The numerical
solution of the large number of such systems within the CFD framework results in exceedingly long CPU
times. Consequently, with most of the comprehensive kinetic mechanisms developed for hydrocarbon fuels,
large scale three-dimensional reactive flow simulations become computationally unaffordable (Liang et al.
2009a). Consequently, further reduction of the size of detailed mechanism, keeping its essential feature, is
desired.
There are two main classes of reduction procedure for detailed chemical kinetic mechanisms, namely time
scale analysis and skeletal reduction. Both approaches target identifying and eliminating the unnecessary
species and reactions and aim at producing computationally efficient reduced mechanisms that are still able
to reproduce the main features of their corresponding detailed mechanisms over the conditions of interest.
Examples of time scale analysis include computational singular perturbation (CSP) (Lam and Goussis 1994)
and intrinsic low-dimensional manifolds (ILDM) (Maas and Pope 1992) methods whereas skeletal reduction

869
contains sensitivity analysis (Hwang 1982), principal component analysis (PCA) (Turányi et al. 1989),
optimization-based techniques (Bhattacharjee et al. 2003, Oluwole et al. 2006), element flux-based method
(Androulakis et al. 2004), directed relation graph (DRG) (Lu and Law 2005) and also directed relation graph
with error propagation (DRGEP) [(Pepiot-Desjardins and Pitsch 2008). Furthermore, dynamic methods like
the work of He et al. (He et al. 2010) and Liang et al. (Liang et al. 2009b, Liang et al. 2009a) update the
reduced mechanism dynamically at each time step, based on the local conditions and develop locally
accurate mechanism. However, as mentioned in ref. (He et al. 2010), the discontinuity in species conversion
rate, when mechanisms switch during the simulation in the fly scheme, may cause species composition
oscillation and the possibility of ODE solver failure. Multi-stage reduction schemes which are the integration
of two or more reduction methods could be employed for increasing the extent of reduction such as DRGASA
(Niemeyer et al. 2010), DRGEP-PCA (Shi et al. 2010), and DRGEP-CSP-DRGEP (Bahlouli et al. 2012) . The
present work utilizes the integration of DRGEP and PCA methods. The procedure used in this work for
generating the reduced mechanism is almost similar to the previous paper of the author (Bahlouli et al. 2012)
except for using PCA instead of CSP.
In this study a reduced mechanism of n-heptane, which is applicable for CFD simulation, is produced by
utilizing two stage reduction processes. To remove unimportant species, at the first stage, the directed
relation graph with error propagation (DRGEP) is applied. In the second stage, the principal component
analysis (PCA) method is used to eliminate insignificant reactions. At each reduction step, the performance
of the reduced mechanism is compared with the original one to guarantee that the generated mechanism
does not exceed user-specified error tolerances.

2.1. Directed Relation Graph with Error Propagation (DRGEP)

The DRGEP method suggests that the effect of the error established by altering the concentration of a
species or by eliminating the species entirely is damped as it propagates along the graph to reach the target
species, a set of species deemed of interest to the investigator. Generally speaking, the species do not have
equal importance, and the species directly linked to the target is of relatively high importance than those that
are farther from the targets. In order to take into account this error propagation process, a geometric
damping has been introduced by Pepiot and Pitsch (Pepiot-Desjardins and Pitsch 2008, Pepiot-Desjardins
and Pitsch 2005) in the selection procedure as follows:

m 1
rAB , P   rsi si 1
(1)
i 1

 
RAB  max rAB ,P 
(2)
All paths P
 

A new definition of the direct interaction coefficient is introduced by Pepiot and Pitsch, (Pepiot-Desjardins
and Pitsch 2008) which is motivated by the shortcomings of earlier formulations, namely,

v
j 1
A, j  j  Bj
(3)
rAB 
max PA , C A 

Here

1 if the jth elementary


 Bj   reaction involve species B,
0 otherwise


PA   max 0 , v A, j  j 
n

j 1
(4)

870
C A   max 0 ,  v A, j  j 
n
(5)
j 1

2.2. Principal Component Analysis (PCA) Method

In the simulation of a combustion process a set of ordinary differential equations is used,

 f k , c 
dc
dt (6)

Here c (t) is the concentration of any species and k is the kinetic parameter. Any change in the kinetic
parameters at time t1 , where k  k and c  c , results in a change in the solution at time t 2 ( where
 

t1  t 2 ). Regarding this fact, Turanyi et al. (Turányi et al. 1989) introduced a reaction rate sensitivity gradient,
which is the derivative of the deviation in the concentration of the species with respect to the rate constant as
follows:

f i t2 
~

F k  , c , t2  ij 
k j (7)

Non-dimensional sensitivity matrix Eq. (7) can be written as:

f i t 2 

~

F k  , c , t2  ij

kj
f i t 2  k j (8)

Since f i is given by

n n
f i k , c    vij R j   vij k j r j c 
(9)
j 1 j 1

~
The elements of the log-normalized sensitivity matrix F can be written as:

~ kj f i k , c  vij R j k , c  vij R j
Fij   n 
f i k , c  k j (10)
 vij R j k , c 
fi
j 1

~
In which F is considered as a ratio of the rate of formation or consumption of species i in reaction j and the
~
net rate of the concentration change of species i. If the magnitude of F is equal to zero it means that species
i does not exist in reaction j. As mentioned by Vajda et al. (Vajda et al. 1985) the kinetic information inherent
~
in the matrix F is extracted by principal component analysis. The response function, which is the basic
concept in the principal component analysis, is reformulated for reaction rate consideration as follows:

Q , c    
 f j  , c   f j   , c
n   2

(11)

j 1 
 f j  ,c   

871
Q , c  is a measure of deviation in a reaction rate caused by a parameter perturbation,  j  ln k j and
 j  ln k j . Vajda et al. (Vajda et al. 1985) suggested that Eq. (11) can be approximated by the simple
quadratic expression:

(12)
Qˆ     T F T F  
~ ~

Here    Q  , c in the neighborhood of  . Kinetic information comes by performing eigenvalue-



~T ~
eigenvector of the matrix F F . The important reactions can be defined as the significant eigenvector
elements of reactions which are characterized by large eigenvalues. With providing the user-specified
tolerances for these parameters, unnecessary reactions can be identified.

3. RESULTS AND DISCUSSION

The Golovichev’s mechanism (Golovitchev 2000) consisting of 57 species and 290 reactions, is chosen
for the oxidation of n-heptane fuel. Five different engine operating conditions are selected for combustion
analysis of n-heptane fuel listed in Table 1. The reduction process which is used in the present work is based
on a two-stage reduction method, which utilizes DRGEP and PCA methods successively. Also, the
combustion system considered in this study is HCCI combustion modeled by a single zone combustion
model. Discussion of how well the simulations reproduce experimental data is beyond the scope of this
paper. More specifically, DRGEP identifies and eliminates unimportant species and reactions and then, in the
second stage, the PCA method is applied to improve the process by eliminating redundant reactions. For
DRGEP reduction, like Liang et al. (Liang et al. 2009a) and Shi et al. (Shi et al. 2010) fuel, HO2, and CO are
selected as the target species at each sampling point. Species that are reachable from the target species are
identified at each sampling point and the collection of all of these species sets constitutes the final important
species set for a specific operating condition. The rest of the species are considered as unimportant species
and reactions that include any of these species are eliminated from the final mechanism.
An initial tolerance value to start the DRGEP reduction process is 10 5 and the tolerance value for PCA
reduction process that follows the DRGEP process is 103 . The selected representative parameters for
investigating the validity of the reduced mechanism in this work are the predicted CA50, peak pressure, and
maximum heat release. At each generation, these parameters are calculated using the HCCI engine
simulation code and compared with the results obtained by using the detailed mechanism for the considered
cases. The generated reduced mechanism is considered as valid one if the errors in predicted CA50, peak
pressure, and maximum heat release using reduced mechanism do not exceed 1° CA, 1%, and 1%
respectively, with those of the detailed mechanism. Therefore, for each operating conditions mentioned in
Table 1 the developed reduced mechanisms are in different final sizes and final algorithm tolerances as can
be seen in Table 2.

Table 1.Operating conditions for considered cases of the n-heptane fueled HCCI engine.
Case 1 2 3 4 5
Equivalence ratio (Ф) 0.68 0.41 0.43 0.38 0.26
n-heptane mass rate
103.43 75.52 89.69 92.40 79.23
(mg/s)
Air mass rate (g/s) 2.28 2.77 3.11 3.66 4.57
TIVC (K) 333 333 413 413 413
PIVC (bar) 1.54 1.55 1.56 1.54 1.57
% EGR 51.01 40.69 31.66 19.79 0.0
Compression Ratio 11.5 11.5 11.5 11.5 11.5

872
 Table 2. Comparison of n-heptane skeletal mechanisms sizes generated at each operating conditions.
Allowable error
Case Species Reactions tolerances are
exceed?
Case 1 38 65 No
Case 2 40 95 No
Case 3 36 104 No
Case 4 38 74 Yes
Case 5 38 85 No

Fig. 1 and Fig.2 show the mechanism size and the interfered errors in calculation of CA50, peak
pressure, and maximum heat release because of elimination of insignificant species and reactions and
algorithm error tolerances at each generation during DRGEP and PCA reduction processes, which are
distinguished with a vertical line. The considered case in these figures is test 2. For the first reduction stage
which utilizes DRGEP reduction method, insignificant species and their corresponding reactions are
identified and eliminated effectively from the mechanism. Following the first stage, PCAF reduction is applied
to the mechanism to remove further reactions in the second stage. By applying this two-stage reduction
method, a detailed mechanism of n-heptane including 57 species and 290 reactions is reduced to a smaller
mechanism containing 40 species and 95 reactions.

Fig. 1. Mechanism size and the corresponding error values at each reduction stage (case 2).

Fig. 2. Algorithm error tolerances for case 2 of the NG fueled HCCI engine.

873
By utilizing the specified reduction processes for each of the operating conditions, the corresponding
reduced mechanism is generated. However, it is required to have a single reduced mechanism. Normally,
this final reduced mechanism can be developed by the combination of all the generated mechanisms of each
case. Another alternative solution for this goal, as mentioned by Shi et al (Shi et al. 2010), and also can be
seen in Fig. 3, is to evaluate the performance of each of the reduced mechanisms at different operating
conditions. It can be seen that most of the generated mechanisms can be used for all other cases while
predicting the representative parameters in the error tolerance limit. For example, the generated reduced
mechanism for case 1 for n-heptane fuel is used to simulate the combustion phase of the n-heptane-fueled
HCCI engine in different considered cases in Table 1. The calculated error in predicting peak pressure,
maximum heat release, and CA50 between reduced and detailed mechanisms are less than the user
specified error tolerance values. However, the developed reduced mechanism of case 4 for n-heptane is not
able to accurately predict some of these parameters for operating conditions of case 1, case 2, and case 3,
therefore it is not considered as a valid one.

(a) case 1 (b) case 2 (c) case 3

(d) case 4 (e) case 5

Fig. 3. Performance of each generated reduced mechanism for n-heptane fuel at different operating conditions

To verify the ability of the reduced mechanism in predicting the pressure trace and heat release rate
histories, a comparison of these parameters between reduced and detailed mechanisms are depicted in Fig.
4 for different operating conditions of the n-heptane-fueled HCCI engine. As indicated in this figure,
simulation using reduced mechanism is in good agreement with the simulation utilizing detailed mechanism.

Fig. 4. Comparison of pressure traces and heat release rate histories by applying the detailed n-heptane mechanism and
its reduced mechanism generated for case 2 at different operating conditions

874
 In addition to matching pressure and heat release rate traces, the start of the combustion (SOC)
calculated from the reduced mechanism is compared with the corresponding detailed results. The definition
of the SOC adopted here was the point at which10 percent of total heat is released. Fig. 5 shows the
difference in prediction of SOC by applying the reduced chemical kinetics mechanism and the detailed
chemical kinetics mechanism in HCCI engine cycle calculations. In HCCI engines, combustion strongly
depends on the chemical kinetics (Onishi et al. 1979). It is observed that the SOC is predicted accurately for
all considered cases.

Fig. 5. Error in prediction of SOC for reduced mechanism relative to the detailed one at all considered cases

Finally, a comparison of the simulation results for the mass fraction profiles of some major species such
as fuels and CO using the reduced mechanism and the detailed mechanism are shown in Fig. 6. The results
show that the calculated mass fraction of these species using the reduced mechanism agrees very well with
the simulation results utilizing the detailed mechanism.
The mechanism reduction achieved reduction ratios of 30 percent for the number of species and also 67
percent for the number of reactions.

Fig. 6. Comparison of mass fraction for some selected species between the detailed n-heptane mechanism and its
reduced mechanism generated for case 2

4. CONCLUSION
In the current study, a reduced mechanism of n-heptane fuel based on the Golovichev’s n-heptane
mechanism, was developted. Reduction procedure to develop the reduced mechanism was based on an
integrated method that utilizes DRGEP and PCA reduction methods. The mechanism reduction achieved
reduction ratios of 30 percent for the number of species and also 67 percent for the number of reactions,
within tight error limits for prediction of CA50, peak pressure, and maximum heat release.

NOMENCLATURE
v stoichiometric coefficient t time

 reaction rate n total number of reactions

PA net production rate of Greek

species A

875
CA net consumption rate of  Total equivalence ratio
species A
F sensitivity matrix Abbreviations

~ Non-dimensional sensitivity IVC inlet valve closing

F
matrix
Rj reaction rate for reaction j rpm revolutions per minute

r j c  product of reactants EGR exhaust gas recirculation

concentration for reaction j
k rate constant SOC start of combustion

Q  , c  response function SZCM single zone combustion model

c concentration of species

REFERENCES
Androulakis, I. P., Grenda, J. M. and Bozzelli, J. W. 2004. Time-integrated pointers for enabling the analysis of
detailed reaction mechanisms. Aiche Journal, 50(11), 2956-2970.
Bahlouli, K., Saray, R. K. and Atikol, U. 2012. Development of a Reduced Mechanism for n-Heptane Fuel in HCCI
Combustion Engines by Applying Combined Reduction Methods. Energy & Fuels, 26(6), 3244-3256.
Bhattacharjee, B., et al. 2003. Optimally-reduced kinetic models: reaction elimination in large-scale kinetic
mechanisms. Combustion and Flame, 135(3), 191-208.
Curran, H. J., et al. 1998. A comprehensive modeling study of n-heptane oxidation. Combustion and Flame, 114(1-
2), 149-177.
Golovitchev, V. I., 2000. Mechanisms (Combustion Chemistry) [online]. Available from:
http://www.tfd.chalmers.se/~valeri/MECH.html.
He, K., Androulakis, I. P. and Ierapetritou, M. G. 2010. On-the-fly reduction of kinetic mechanisms using element
flux analysis. Chemical Engineering Science, 65(3), 1173-1184.
Hwang, Y. T. 1982. On the proper usage of sensitivities of chemical kinetics models to the uncertainties in rate
coefficients. Proceedings of the National Science Council B. ROC, 6, 270-278.
Lam, S. H. and Goussis, D. A. 1994. CSP method for simplifying kinetics. International Journal of Chemical
Kinetics, 26(4), 461-486.
Liang, L., Stevens, J. G. and Farrell, J. T. 2009a. A dynamic adaptive chemistry scheme for reactive flow
computations. Proceedings of the Combustion Institute, 32 I, 527-534.
Liang, L., et al. 2009b. The use of dynamic adaptive chemistry in combustion simulation of gasoline surrogate
fuels. Combustion and Flame, 156(7), 1493-1502.
Lu, T. and Law, C. K. 2005. A directed relation graph method for mechanism reduction. Proceedings of the
Combustion Institute, 30(1), 1333-1341.
Maas, U. and Pope, S. B. 1992. Simplifying chemical kinetics: Intrinsic low-dimensional manifolds in composition
space. Combustion and Flame, 88(3-4), 239-264.
Najt, P. M. and Foster, D. E. 1983. Compression-ignited homogeneous charge combustion. SAE, 790501.
Niemeyer, K. E., Sung, C. J. and Raju, M. P. 2010. Skeletal mechanism generation for surrogate fuels using
directed relation graph with error propagation and sensitivity analysis. Combustion and Flame, 157(9),
1760-1770.
Oluwole, O. O., et al. 2006. Rigorous valid ranges for optimally reduced kinetic models. Combustion and Flame,
146(1-2), 348-365.
Onishi, S., et al. 1979. Active thermo-atmosphere combustion (ATAC) - A new combustion process for internal
combustion engines.
Pepiot-Desjardins, P. and Pitsch, H., 2005. Systematic Reduction of Large Chemical Mechanisms. Presented at
the 4th Joint Meeting of the U.S. Sections of the Combustion Institute. Drexel University, Philadelphia, PA.
Pepiot-Desjardins, P. and Pitsch, H. 2008. An efficient error-propagation-based reduction method for large
chemical kinetic mechanisms. Combustion and Flame, 154(1-2), 67-81.
Shi, Y., et al. 2010. Automatic chemistry mechanism reduction of hydrocarbon fuels for HCCI engines based on
DRGEP and PCA methods with error control. Energy and Fuels, 24(3), 1646-1654.
Thring, R. H. 1989. Homogeneous charge compression ignition (HCCI) engines. SAE, 892068.
Turányi, T., Bérces, T. and Vajda, S. 1989. Reaction rate analysis of complex kinetic systems. Int. J. Chem. Kinet.,
21(2), 83-99.
Vajda, S., Valko, P. and Turanyi, T. 1985. Principal Component Analysis of Kinetic-Models. International Journal of
Chemical Kinetics, 17(1), 55-81.

876
 THREE DIMENSIONAL NUMERICAL MODELLING OF HYDROGEN COMBUSTION IN A
SPHERICAL MODEL COMBUSTOR

Mustafa Ilbaş*, Zehra Gokalp Ozturk**, Serhat Karyeyen*

*Gazi University, Faculty of Technology, Department of Energy Systems Engineering

Teknikokullar/ANKARA/TÜRKİYE
**Gazi University, Institute of Science and Technology, Ankara/TÜRKİYE
*ilbas@gazi.edu.tr, **zgokalp5263@hotmail.com, *serhatkaryeyen@gazi.edu.tr

ABSTRACT
In this study, combustion and emission characteristics of different fuels have been investigated numerically
in a spherical model combustor. These investigations have been performed by using Ansys Fluent CFD
code. The numerical modelling of turbulent non-premixed diffusion flames have been performed in this
combustor. Mathematical models studied in this study included the k-ε model of turbulent flow, the
PDF/mixture fraction model of non-premixed combustion and P-1 radiation model. Three dimensional
modelling of the model combustor was carried out for combustion of hydrogen, methane, propane and
butane at the same combustion condition (40 kW thermal power and λ=1.0 stoichiometric mixture ratio).
According to the modelling results, the maximum temperature was predicted during the hydrogen
combustion, Similarly, the minimum temperature in this model combustor was also predicted during the
butane combustion. Moreover, More NOX emissions were formed in high temperature zones. Hence, the
maximum NOX formation was predicted for hydrogen combustion among these fuels.
Keywords: Combustion, CFD Modelling, Emission, Spherical Combustor

INTRODUCTION
Energy is an important parameter in human life for many years and fossil fuels comprise high energy.
Required energy is mainly supplied by fossil fuels. But, fossil fuels cause formation of green house gases
and these gases, especially carbon monoxide, are responsible from global warming (Boran et al., 2012). In
general, fossil fuels cannot be used directly. Its energy can be converted into the heat. Because of that,
Combustors are necessary for this process. There are a lot of known combustors. Cylindirical combustors
are the most known combustor type and there are a lot of studies interested in cylindirical combustors in the
literature.

Keramida et al. (Keramida et al., 2000) investigate effects of the different radiation models on combustion
modelling in a natural gas-fired furnaces. Ilbas et al. (Ilbas et al., 2005) investigate numerically combustion
and NOX emission characteristics of pure hydrogen and hydrogen-hydrocarbon composite fuel in a
cylindirical combustor. Similarly, Tomczak et al. (Tomczak et al., 2002) investigate combustion of the
hydrogen-methane mixture in a gas turbine combustion system. Combustion of hydrogen-methane mixtured
fuels was performed by Yılmaz and Ilbas (Yılmaz and Ilbas, 2008). Kamnis et al. (Kamnis et al., 2006) takes
place numerical modelling of propane combustion in a cylindirical combustor. In an other study related with
propane combustion, Large Eddy Simulation has been used to predict temperature ans species
concentrations by Paul et al. (Paul et al., 2010) in a cylindirical combustor. Qi et al. (Qi et al., 2008) studies
combustion of premixed butane-air laminar flame jet and they observe the combustion with Mach-Zehnder
Interferometry.

In this study, 3-D numerical modellings of hydrogen, methane, propane and butane combustion were
numerically investigated in a spherical model combustor.

CFD MODELLING

Description of the Spherical Model Combustor

Three-dimensional CFD modelling of spherical model combustor was developed and physical domain of
this combustor is shown in Figure 1. Dimensions of the combustor are 10 mm of fuel inlet diameter, 200 mm
of air inlet diameter with 4 mm of wall thickness, 200 mm of combustor outlet and 600 mm of combustor
diameter. Lengthes of the inlet-outlet are also equal to 50 mm. All boundary conditions are 293 K and 1 atm
for this combustion system.

877
Fig. 1. The developed spherical model combustor

Mathematical and Combustion Model

The timeaveraged gas phase equation for steady turbulent flow can be written as (Ghenai, 2010):

(1)

where is the dependent variable that can symbolize the velocity . is the transport coefficient of
variable and is the source term of the transport equation for .

This study was modelled by using The Mixture Fraction/PDF combustion models. This combustion model
comprises the solution of transport equations for a single conserved scalar (the mixture fraction). Transport
equations for individual species are not solved in this model. Individual component concentrations for
species of interest are derived form the predicted mixture fraction distribution. The interaction of turbulence
and chemistry are accounted for with the help of a probability density function (PDF) (Ilbas, 1997).

The PDFcombustion model was specifically developed for the modelling of turbulent diffusion flames. For a
fuel/oxidant system, the mixture fraction, can be expressed in terms of the local fuel mass fraction as:

(2)

where and are mass fraction of fuel and oxidant.

The mixture fraction, is a conserved quantity whose value at each point in the flow domain is computed by
the solution of the following conservation equation for the time averaged value of the turbulent flow field
(Ilbas, 1997).

(3)

where is the source term.

In addition to solving for the mean mixture fraction, a conservation equation is solved for the mixture fraction
variance, which is used in the closure model describing turbulence-chemistry interactions (Ilbas, 1997):

(4)

where , and are constants used in the mixture fraction/PDF model.

The radiation heat transfer occurs at high temperature levels. The gas flame temperature is generally high
(1000-1600 °C) in combustors. Hence, the heat transfer from combustors is significant. Thus, It is essential

878
to include the radiation model for a more accurate prediction of the temperature profile in combustors (Ilbas,
1997).

NOx Formation
Generally, NO accounts more than 95% of total NOX formation in most combustion process. Because of
that, It is sufficient to examine the formation of NO. In gas mixture combustion, NO formation can be
attributed to three chemical kinetic processes which are thermal, prompt and fuel NOX. Thermal NOX is
formed by the oxidation of atmospheric nitrogen at high temperature (Ilbas et al., 2005). Prompt NOX is
formed by the reactions of intermediate species at the flame front. The fuel NOX mechanism defines the
oxidation of the nitrogen bound in the fuel molecules (Ilbas, 1997).

Well-known Zeldovich (Zeldovich, 1947) reaction determines the thermal NO:

O + N2 N + NO (5)

N + O2 O + NO (6)

N + OH H + NO (7)

The mass transport equation is solved for the NO calculation, taking in account convection, diffusion,
production and consumption of NO and related chemical species. The overall thermal NO formation rate
can be calculated as (Fluent Incorporated, 2011).

(8)

The prompt NOX formation mechanism was firstly notified by Fenimore (Fenimore, 1971). The prompt NOX
formation is significant in most hydrocarbon fuel combustion conditions especially, low temperature, short
residence times and fuel-rich conditions. In this study, the prompt NO contribution to the total NOX formation
considerable because methane, butane and propane have carbon in their structures. Actually, the actual
prompt NOX formation route is more complex and embodies many intermediate reactions and species. The
prompt NOX route is generally accepted as

CH + N2 HCN + N (9)

N + O2 NO + O (10)

HCN + OH CN + H2O (11)

CN + O2 NO + CO (12)

Formation rate of the prompt NO is calculated by using the De Soete (De Soete GG, 1971) global model as

(13)

The thermal and prompt NOX formations are calculated by using Eq. (8) and (13).

(14)

RESULTS AND DISCUSSIONS

In this part of study, combustion of hydrogen, methane, butane and propane in a spherical model
combustor are shown and commented. Initially, the predicted temperature distributions on y-z plane are
given in Figure 2a, 2b, 2c and 2d. According to the figure 2, the predicted maximum temperature level was
about 2160K for hydrogen combustion. Subsequent predicted maximum temperature levels were about
1930K for methane combustion, 1850K for butane combustion and 1880K for propane combustion. But,
when it is looked at carefully to figure 2c and 2d, it is clearly seen that the predicted temperature
distributions are much the same in both cases. The reason of this case, heating values of butane and
propane are similar. Temperature levels for all predictions are increasing from the inlet zone towards the
combustor exit. In fact, temperature level is generally decreasing towards the combustor exit considerably
(Ilbas et al., 2005). But, this situation may take place in the cylindirical and longer combustors.

879
Fig. 2a. Temperature distribution for hydrogen Fig. 2b. Temperature distribution for methane

Fig. 2c. Temperature distribution for butane Fig. 2d. Temperature distribution for propane

Figure 2a shows pure hydrogen combustion. Similarly, Yilmaz and Ilbas (Yılmaz and Ilbas, 2008)
investigate combustion of hydrogen-methane mixtured fuel in a cylindirical combustor. According to this
study, the maximum temperature level is determined as about 1700 K during M30H70 fuel combustion.
There is a minor difference between study of Yilmaz and Ilbas and the present studies. But, Experimental
and the predicted temperature levels are in good agreement. The one of the reason for some differences is
different fuel gas fraction and the basic reason is the use of different combustors (cylindirical and spherical).

Methane combustion results are presented in figure 2b. In a study carried out in the literature, Keramida et
al. (Keramida et al., 2000) predict combustion of methane in a cylindirical combustor and the maximum
temperature level is predicted as about 1600 K at lower thermal power of combustion condition. When this
study compared with Keramida’s study, It is seen that there is also small difference. There are two reasons
for this case. Thermal powers and combustors used in those studies were different. The combustion models
carried out in these studies were also different. Combustion performance is predicted by using eddy-
dissipation combustion model by Keramida et al., while predictions are performed by using The Mixture
Fraction/PDF combustion model in the present work.

The predicted temperature distributions in the combustor for butane and propane combustion are illustrated
in figure 2c and 2d. In the figure 2d, temperature distribution for propane is shown and Paul et al. (Paul et
al., 2010) predict the propane combustion by using Large Eddy Simulation (LES) method in a cylindirical
combustor. According to the study of Paul et al., the maximum temperature value is predicted as almost
1900 K and these results and present work results for propane combustion are very similar.

880
In combustion modellings, temperature distributions may be shown on different planes. By this way,
combustion characteristics are described in more detailed throughout combustor. Hence, temperature
distributions on x-y plane for all modellings are given in figure 3.

Fig. 3a. Temperature distribution for hydrogen on x-y plane at z=0.1m, z=0.3m and z=0.5m (left to right)

Fig. 3b. Temperature distribution for methane on x-y plane at z=0.1m, z=0.3m and z=0.5m (left to right)

Fig. 3c. Temperature distribution for butane on x-y plane at z=0.1m, z=0.3m and z=0.5m (left to right)

881
Fig. 3d. Temperature distribution for propane on x-y plane at z=0.1m, z=0.3m and z=0.5m (left to right)

As It is seen in the figure 3, the overall predicted flame temperature increases with proceeding towards the
combustor exit. In addition, the predicted wall temperature was about 1800 K for hydrogen combustion in
the figure 3a. Hydrogen has the lower density in comparison with other fuels and air. Because of that,
hydrogen may diffuse more productive in a spherical combustor. Moreover, It includes the highest heating
value among these fuels.

Fig. 4a. NOX distribution for hydrogen Fig. 4b. NOX distribution for methane

Fig. 4a. NOX distribution for butane Fig. 4b. NOX distribution for propane

882
NOX distributions for hydrogen, methane, butane and propane combustions are shown in figure 4. As it is
seen in the figure 4, NOX formation increases as the temperature level increases. In other words, NOx
formation in high temperature zones is higher and in low temperature zone is lower. Moreover, the NOX
formation increases from the combustor inlet to the combustor exit. The maximum NOX formation was about
1450 ppm for hydrogen combustion. The maximum NOX formations for the other fuels were almost 130 ppm
for methane combustion, almost 60 ppm for buthane combustion and almost 90 ppm for propane
combustion in combustor.

CONCLUSIONS
In this study, three-dimensional CFD analyses of combustion and emission characteristics of hydrogen,
methane, butane and propane were performed in a spherical combustor. The some conclusions obtained in
this study are given below:

 When CFD modelling results are compared with the literature, results show similarity mostly.
 The maximum NOX formation is for hydrogen combustion and the minimum NOX formation is for
butane combustion as expected.
 Although the excess NOX formation happens during the hydrogen combustion, the best combustion
performance is belong to hydrogen combustion. The hydrogen contains higher heating value in proportion
to the other fuels.
 Especially, density of air in comparison to the density of hydrogen is 15 times higher. Because of
that, the hydrogen is preferable in the spherical combustor.
 It is showed that the spherical combustor is an alternative to the cylindrical combustor, which is
used commonly, especially for hydrogen combustion

883
REFERENCES
Boran, F. E., Dizdar, E., Toktas, I., Boran, K., Eldem, C. and Ö. Asal. 2012. A multidimensional analysis of
electricity generation options with different scenarios in Turkey. Energy Sources, Part B: Economics,
Planning and Policy. 8(1): 44-55.

De Soete, GG. 1971. Overall reaction rates of NO and N2 formation from fuel nitrogen. 15th Symposium
(International) on Combustion, Pittsburg, PA, USA: The Combustion Institute, p. 1093.

Fenimore, CP. 1971. 13th Symposium (International) on Combustion, Pittsburg, PA, USA: The Combustion
Institute.

Fluent Incorporated. Fluent 13.0, User’s Guide, 2011.

Ghenai, C. 2010. Combustion of syngas fuel in gas turbine can combustor. Advanced in Mechanical
Engineering, Article ID: 342357 1-13.

Ilbas, M. 1997. Studies of Ultra Low NOX Burner, PhD Thesis, University of Wales, College of Cardiff, UK.

Ilbas, M., Yılmaz, İ. and Y. Kaplan. 2005. Investigations of hydrogen and hydrogen-hydrocarbon composite
fuel combustion and NOX emission characteristics in a model combustor. International Journal of Hydrogen
Energy. 30: 1139-1147.

Kamnis, S. and S. Gu. 2006. Numerical modelling of propane combustion in a high velocity oxygen-fuel
thermal spray gun. Chemical Engineering and Processing. 45:246-253.

Keramida, E. P., Liakos, H. H., Founti, M. A., Boudouvis, A. G. and N. C. Markatos. 2000. Radiative heat
transfer in natural gas-fired furnaces. International Journal of Heat and Mass Transfer. 43: 1801-1809.

Paul, S. C., Paul, M. C. and W. P. Jones. 2010. Large Eddy Simulation of a turbulent non-premixed
propane-air reacting flame in a cylindirical combustor. Computers & Fluids. 39: 1832-1847.

Qi, J. A., Wong, W. O., Leung, C. W. And D. W. Yuen. 2008. Temperature field measurement of a premixed
butane/air slot laminar flame jet with Mach-Zehnder Interferometry. Applied Thermal Engineering. 28:1806-
1812.

Yılmaz, I. and M. Ilbas. 2008. An experimental study on hydrogen-methane mixtured fuels. International
Communications in Heat and Mass Transfer. 35: 178-187.

Zeldovich, YAB. 1947. Oxidation of nitrogen in combustion. USSR Institute of Chemical Phsyics: Academy
of Science.

884
PRE-COMBUSTION CO2 CAPTURE BY PD BASED COMPOSITE MEMBRANE DEPOSITED ON
POROUS NICKEL SUPPORT

a,b a,b a a b a*
Sung-Wook Lee , Chun-Boo Lee , Dong-Wook Lee , Jong-Soo Park , Sung-Hyun Kim , Shin-Kun Ryi

a
Energy Materials and Convergence Research Department, Korea Institute of Energy Research (KIER), 102 Gajeong-ro,
Yuseong-Gu, Daejeon 305-343, South Korea
b
Department of Chemical and Biological engineering, Korea University, 5-Ga, Anam-Dong Sungbuk-Gu, Seoul 136-701,
South Korea
c
Department of Chemical and Biological Engineering, University of British Columbia, 2360 East Mall, Vancouver, BC,
Canada V6T 1Z3
*
Corresponding author: Tel.: +82-42-860-3155; Fax: +82-42-860-3309;
E-mail address: h2membrane@kier.re.kr

ABSTRACT
In this study, a combined test of the water gas shift (WGS) reactor and a Pd-based composite membrane was
carried out for pre-combustion CO2 capture in a coal gasifier. To increase the stability of the membrane module at
a high pressure and temperature, the high pressure module, which was composed of a unit-cell and a high
pressure chamber, was used in this study. The two series of WGS reactions, i.e., a high-temperature shift and a
low-temperature shift, were performed under a gas composition of 60% CO and 40% H 2 at 2100 kPa to imitate
coal gasification. After the two series of WGS reactions, the composition of the synthetic gas was 41.2% CO 2,
58.7% H2 and 0.1% CO with a CO conversion of ~99%, which was similar to the equilibrium value. The flow rate of
the dried synthetic gas after the removal of unreacted water was ~0.1 Nm3 h -1. The CO2 enrichment and H2
recovery tests with the high-pressure membrane module after the WGS reaction showed that the enriched CO2
concentration and H2 recovery ratios were ~92% and ~96%, respectively, at 673 K and 2100 kPa. The long-term
stability test showed that the CO2 concentration decreased to 78.2%, and CO was generated and reached to 8.8%
in the retentate stream after 47 h. The membrane and module analysis after the test showed that there were no
indications of any reactions on the membrane surface, while a large amount of carbon deposition was observed in
the high-pressure chamber, indicating that the RWGS and CO2 hydrogenation reaction took place on the surface
of a high-pressure chamber made of 316L stainless steel after ~50 h under CO2 and H2 mixtures. The stability test
for ~ 3137 h showed that these catalytic activities could be successfully prevented using steel with higher Cr and
Ni contents, such as 310S.

INTRODUCTION
The concentration of greenhouse gases in the atmosphere has increased. The capture and storage of CO2
(CCS) is the important objective for preventing global warming because CO2 is the main cause of the greenhouse
effect [1]. CCS involves integrated technologies to reduce the levels of CO2 directly. CO2 is generated during the
combustion of fossil fuels, such as coal, oil and natural gas [2]. The methods of CCS can be classified as post-
combustion capture [3], pre-combustion capture [4] and oxy-fuel combustion [5]. The methods for CO2 capture can
be accomplished by technologies such as solid adsorption, absorption into a liquid solvent or onto a membrane or
other methods of physical separation [4, 6]. During the process of power generation by the coal-based plants, the
integrated gasification combined cycle (IGCC) with pre-combustion CCS could enable the low-cost capture of CO2
in the long term [7, 8]. One advanced technology that offers increased energy efficiency and a reduced cost is the
use of hydrogen-selective membranes, as shown in Fig. 1 [4, 9]. In the IGCC, the coal is combined with oxygen
(and H2O) in the gasifier to produce the synthetic gas, which is composed mainly of hydrogen and CO. The
synthetic gas is further converted to CO2 and hydrogen through the water gas shift (WGS) reaction, and then CO2
capture is completed by CO2 enrichment from the hydrogen mixture by the H2 membrane [10, 11]. The gas stream
in IGCC has a considerably high pressure. The use of membranes allows for the selective permeation of hydrogen
to the permeate side, while the high-pressurised retentate stream consists primarily of CO2. The use of
pressurised CO2 in the retentate stream reduces the cost of the compression process required for CO 2 transport
and storage [12, 13].

885
Fig. 1. Simultaneous CO2 capture and H2 production using an H2 selective membrane for coal gasification.

In our previous studies, a hydrogen permeable Pd-based membrane was developed by DC/RF magnetron
sputtering followed by a polishing process [14, 15]. A unit-cell of the completed plate-type Pd-based membrane
was developed with a double O-ring sealing method and used to study the feasibility of separating hydrogen for
CO2 capture and hydrogen purification. We confirmed that the hydrogen membrane had a hydrogen permeation
flux of 1.4 mol m-2 s-1 with an H2/N2 perm-selectivity of 1000, and it could then enrich 40% of the CO2 at a flow rate
of 1.2 Nm3 hr-1 up to 87% with a hydrogen recovery ratio of 90% at 675 K and a total feed pressure of 1600 kPa
[10, 16]. We also studied the H2 membrane after the WGS reaction for pre-combustion CO2 capture in a coal
gasification process [17]. However, the flange-type membrane module was expanded at high pressures and
temperatures due to the very large expansion coefficient of the metal, which caused a decrease in the selectivity
for hours. Recently, we developed a high-pressure module composed of a unit-cell and a high-pressure chamber
[18]. The newly designed high-pressure module provided reliable leak-tightness for a long period of time under a
high temperature and pressure compared with the flange-type module. The hydrogen permeation test showed that
the hydrogen permeation flux increased up to 2.5 mol m-2 s-1 at a temperature of 673 K and a pressure difference
of 2000 kPa, and the N2 leakage test confirmed that the H2/N2 selectivity remained at 4000 after 300 hours.
In this study, the high-pressure membrane module was introduced after the WGS reactor for the pre-
combustion capture of CO2 in a coal gasifier. The goal was to study the feasibility of separating hydrogen for CO2
capture and hydrogen purification using Pd-based membranes after the WGS reaction and H2O removal using a
high-pressure chamber module. We also tested the long-term stability of the membrane and module for hydrogen
permeation flux, CO2 enrichment and H2 recovery. The gas composition of the synthetic gas, permeate and
retentate were analysed to confirm the purity of the hydrogen and the enrichment of CO2.

EXPERIMENTAL

Preparation of Membrane and Module

The Pd-based membranes of the plate type were deposited on a porous nickel support (PNS) developed by the
Korea Institute of Energy Research (KIER) [14-16]. The support was polished to reduce its surface roughness [16].
To further reduce the pore size, the PNS was modified with sub-micron sized nickel powder form Aldrich [10]. The
Pd-based membranes were deposited on the PNS using the DC/RF magnetron sputtering system at 723 K [10,
14]. The membrane was polished using an auto-polisher (Poli-300, GNP Tech.) to reduce the membrane thickness
and make the membrane free of defects [15]. The used membrane module was the one designed in KIER to be
used for hydrogen purification and CO2 capture under high pressure [18]. This high-pressure membrane module,
shown in Fig. 2, is composed of a unit-cell and a high-pressure chamber. The unit-cell and the high-pressure
chamber were made of 316L stainless steel. The unit-cell consisted of top and bottom plates, two metal O-rings,
metal tubes and a plate-type Pd-base membrane. The high-pressure chamber consisted of a cover plate, a
cylinder, a metal O-ring and metal tubes. The unit-cell was tightened with eight pairs of bolts and nuts for sealing.
The assembled unit-cell was mounted in the high-pressure chamber, and the tubes of the unit-cell were inserted
into the tubes of the chamber. The tightness of the tubes was achieved using graphite ferrules and the stainless
steel fittings.

886
Fig. 2. Photographs of the high pressure membrane module (a) composed of the assembled unit-cell (b) in the
high pressure chamber (c).

WGS Reaction and CO2 Capture

The placement of the hydrogen membrane system after the WGS reactor for pre-combustion CO2 capture is
shown in Fig. 3. This system is composed of two parts: a series of WGS reactors and a hydrogen membrane. The
WGS reactor consisted of a synthetic gas injection module, a water injection system, a gas and steam pre-heater
and a series of WGS reactors, i.e., a high-temperature shift (HTS) reactor and a low-temperature shift (LTS)
reactor. The WGS catalysts ware purchased from Shud Chemie and the physical and chemical properties are
shown in Table 1. The WGS reaction test was performed under gas after coal gasification, which was composed of
60.0% CO and 40.0% H2 [19]. The temperatures of the HTS and LTS ware controlled at 623-653 K and 453-493 K,
respectively. The reaction was carried out at 2100 kPa with a gas-hourly space velocity (GHSV) of 7,500 h-1 and a
steam-to-carbon (S/C) ratio of 3. The unreacted steam in the WGS reactor may be the result of the decrease of H 2
permeation [20]. Therefore, the steam was recovered with a heat exchanger and two condensers after the LTS
reactor. Then, the CO2 was enriched through H2 separation from the mixture of gas after the WGS with the high-
pressure module. The flow rate in the permeate and retentate streams ware measured using a digital mass flow
meter. The gas compositions of the WGS reactor and the permeate and retentate streams of the membrane
module were analysed with a gas chromatograph (GC 6890, Agilent). When testing was completed, the membrane
was cooled under N2 in the membrane module to prevent phase transition of the Pd membrane, which would cause
membrane destruction. Then the surface morphology and fundamental composition was characterised by
SEM/EDX.

887
Fig. 3. Schematic of the water gas shift/membrane hybrid system.

Table 1. Commercial catalysts for the water gas shift reaction

HTS(SHT-4) LTS(MDC-7)
Form Tablet Form Tablet
Shape 3.2 x 3.2 mm Shape 3.2 x 3.2 mm
Comp. Conc. (%) Comp. Conc. (%)
Iron Oxide 86-92 Copper Oxide 40-44
Chromium Oxide 6-10 Zinc Oxide 44-50
Chromium Trioxide 0.1-2.0 Aluminium Oxide 7.0-13
Copper Oxide 1.5-2.1

RESULT AND DISCUSSION

Pure Gas Permeation

The leak test of the membrane module was carried out using N2 at room temperature and a pressure difference
of 500 kPa. The leakage of N2 was <2.513 x 10-4 mol m-2 s-1, which is similar to the value of the previous study
[15]. The membrane was stabilised for 300 hours under H2 at 673 K and at a pressure difference of 100 kPa. The
permeation test of the Pd based membrane performed with pure hydrogen and N2 as a function of pressure
difference range from 100 to 2000 kPa at 673 K. As shown in Fig. 4, the hydrogen permeation flux increased with
the pressure difference. The hydrogen permeation flux was measured as 2.97 mol m-2 s-1 at a pressure difference
of 2000 kPa. The H2/N2 selectivity decreased from 6700 to 2910 as the pressure difference increased from 400 to
the 2000 kPa. This phenomenon was similar to the result shown in a previous study [18] and could be explained
by the different n-values of H2 permeation and N2 permeation through the membrane. N2 permeation is dependent
on the difference in pressure through defects in the membrane or module sealing, while hydrogen permeation
depends on the pressure exponent (n-value) of 0.5-1.0 [21]. From the Sevierts’ law, the pressure exponent n-value
is 0.5. However, sometimes, higher n-values were reported because of support resistance, surface reaction,
defects and impurities in the membrane. In this study, an n-value of 0.65 was observed.

888
Fig. 4. Hydrogen permeation flux and H2/N2 selectivity as a function of pressure difference: The temperature was
673 K.

WGS and CO2 Capture

The stream that emits from a coal gasifier contains ~60% CO (dry basis), which can be further converted to CO2
and H2 through the water gas shift (WGS) reaction (Eq. (1)).

CO + H2O ↔ CO2 + H2 ΔH0298 = -41 kJ mol-1 (1)

This reaction is exothermic and reversible such that CO conversion through the WGS reaction is
thermodynamically favoured at low temperatures. In traditional applications, CO conversion by the WGS involves
multi-stage reactors. The first one is a high-temperature shift (HTS) reactor operated at a high temperature (623-
653 K) under an Fe-Cr based catalyst, and the second one is a low-temperature shift (LTS) reactor operated at a
low temperature (453-493 K) under a Cu-Zn based catalyst [17, 19]. In this study, the WGS reactors were
composed of HTS equipped with Fe-Cr-Cu catalysts and LTS equipped with a Cu-Zn catalyst, as shown in Table 1.
A gas mixture containing 60% CO and 40% H2 was introduced into the HTS reactor at 2100 kPa to simulate a
high-pressure coal gasifier. At that time, steam was introduced with an S/C ratio of 3. As shown in Table 2, the
composition at the end of LTS was 41.2% CO2, 58.7% H2 and 0.1% CO, and the CO conversion reached ~ 99%,
i.e., the equilibrium value at 493 K, indicating that the WGS reaction was not affected by the pressure. As shown in
Eq. (1), the WGS reaction involves no variation of the mole number. Therefore, the pressure does not affect the
conversion.
Before introducing the WGS products into the membrane, the unreacted steam was removed by a heat
exchanger and two condensers after the LTS reactor. After steam removal, the flow rate of the dried synthetic gas
was ~0.1 Nm3 hr -1. The CO2 capture by separation of H2 with a high-pressure membrane module equipped with a
Pd-Au/PNS membrane was performed at 673 K and at the same pressure as the WGS reactor (2100 kPa). The
permeate side was ambient, meaning that the pressure difference across the membrane was 2000 kPa. As shown
in Table 2, the concentrations of CO2 in the retentate and H2 in the permeate stream reached ~92.1% and ~99.8%.
These values were higher than in a previous study using a flange-type module [11]. The use of a high-pressure
module increased the tightness and durability at high pressure and temperature [18].

Table 2. The gas concentration in the membrane module placed after the WGS reactor
Concentration. (%)
Before WGS After WGS After membrane After membrane
(reten. side) (perm. side)

CO 60 0.1 2.7
H2 40 58.7 5.1 99.8
CO2 41.2 92.1
CH4 0.1 0.2

889
 To observe the stability of the WGS combined membrane system, WGS and CO2 enrichment tests were carried
out for 47 hours. CO2 and CO in the retentate stream and H2 in the permeate stream were monitored during the
stability test. As shown in Fig. 5, the CO and CO2 concentrations in the retentate stream and the H2 concentration
in the permeate remained constant for ~ 30 h. However, the CO2 concentration decreased to 78.2%, and then the
CO concentration increased to 8.8% after 47 h. Meanwhile, the H2 concentration in the permeate stream remained
constant, demonstrating the stability of the gas-tightness of the membrane module and membrane. Furthermore, a
considerable amount of condensed water was observed in the Teflon tube located at the end of the pressure
regulator on the retentate side. This means that there was a reverse water gas shift (RWGS, Eq. (2)) reaction in
the retentate stream. The gas analysis showed that 13.9% of the CO2 was converted and 8.7% CO was
generated. The RWGS reaction involves no molar variation, meaning that 5.2% of the CO2 was consumed through
other reactions, such as the CO2 dehydrogenation reaction (Eq. (3)) and the methanation reaction (Eq. (4)). The
gas analysis showed that CH4 was ~ 0.1% in the retentate stream, meaning that the methanation reaction was
negligible.

(2)
(3)
(4)

Fig. 5. Long-term stability test of water gas shift/membrane hybrid system: The temperature was 673 K.

Li et al. reported CO2 decomposition on a Pd membrane surface, leading to carbon formation on the membrane
at > 673 K [22]. After testing the combined WGS-membrane for 47 h, a pure gas permeation test was carried out
with H2 and N2. The hydrogen permeation flux and the N2 leakage were similar to those of a fresh membrane,
indicating that deformation of the membrane occurred during the combined WGS-membrane test. After testing, the
membrane module was cooled under N2 and disassembled for membrane analysis. As shown in Fig. 6, SEM and
EDX analysis of the membrane surface confirmed that no carbon formation or membrane deformation occurred
during the test. On the other hand, we observed a considerable amount of carbon in the high-pressure chamber
made of 316L stainless steel, as shown in Fig. 7. Gielen et al. observed the reduction of H2 permeation flux when
CO2 was fed at 673 K [23]. The authors attributed the reduction of H2 permeation flux to carbon formation by
hydrogenation of CO2 and CO in the Pd or PdAg membrane layer. However, they did not show any analysis of the
membrane. The authors also observed the RWGS reaction over the 316L stainless steel powder when CO 2 was
fed with H2. Leffler [24] reported that there was carburisation in the 316L stainless steel in the atmosphere of the
H2 mixture with gases containing carbon, e.g., in the form of CO, CO2 or CH4.

890
Fig. 6. Surface SEM image (a) EDX profile (b) after the long-term stability test.

Fig. 7. Photographs of the high-pressure chamber after the long-term stability test.

We fabricated our membrane module using 310S stainless steel instead of 316L and carried out a long-term
stability test at 673 K and at a pressure difference of 100 kPa. As shown in Fig. 8, the H 2 permeation flux
drastically decreased after ~ 600 h in the 316L module, while the H 2 permeation flux remained constant for ~ 3137
h in the 310S module. After the long-term stability test, the two different modules were disassembled. While large
amount of carbon were observed in the 316L module, similar to the combined WGS-membrane test, no carbon
was observed in the 310S module. Table 3 shows the element composition of 316L stainless steel and 310S. The
310S has a chromium content of 25% and a nickel content of 20%, while 316L stainless steel has a chromium
content of 17% and a nickel content of 11%. Leffler [25] reported that carburisation was successfully prevented by
the addition of chromium, nickel and silicon. From the long-term stability test, we can conclude that the stability of
the membrane module can be achieved using high carburisation resistant material.

891
Fig. 8. Long-term stability test of the 316L module (a) and the 310S module (b): Feed: 40%CO2+60%H2;
Temperature: 673 K; Pressure difference: 100 kPa.

Table. 3. Elemental compositions of 316L and 310S stainless steel

Composition [%]

Mn Si Cr Mo Ni

316L ≤2.0 ≤1.0 16.5-18.5 2-2.5 10.0-13.0

310S ≤2.0 ≤1.5 24.0-26.0 - 19.0-22.0

CONCLUSIONS

The following results were obtained from the combined WGS and membrane tests for CO2 capture in the coal
gasification process:
• The high-pressure module composed of a unit-cell and high-pressure chamber provided more stable leak-
tightness than the flange-type module used in previous studies [11, 17].
• In the HTS and LTS reactions carried out under a very high pressure of 2100 kPa to simulate the coal
gasifier, ~99% CO conversion and a similar equilibrium value were achieved.
• The CO2 capture test using a high-pressure module equipped with a Pd-Au/PNS membrane was
performed at 673 K and at the same pressure as the WGS reactor (2100 kPa) and showed that the
membrane module enriched 41.2% CO2 to ~92.1% and 58.7% H2 to ~99.8% simultaneously.
• The stability test and SEM/EDX analysis of the membrane showed that no deformation was observed in
the membrane layer. Instead, 316L stainless steel was catalytic for the RWGS and CO2 dehydrogenation.
These catalytic activities could be successfully prevented using steel with higher Cr and Ni contents, such
as 310S.

ACKNOWLEDGEMENT

This work was supported by the Energy & Resource Technology Development Program (2008-C-CD11-P-07-
0-0000) under the Ministry of Knowledge Economy, Republic of Korea.

REFERENCES

[1] M.F. Irfan, M.R. Usman, K. Kusakabe, Coal gasification in CO2 atmosphere and its kinetics since 1948: A brief
review. Energy 36:12-40.
[2] B. Li, Y. Duan, D. Luebke, B. Morreale, Advances in CO2 capture technology: A patent review. Applied Energy
102:1439-1447.
[3] C. Kunze, H. Spliethoff, Assessment of oxy-fuel, pre- and post-combustion-based carbon capture for future
IGCC plants. Applied Energy 94:109-116.
[4] A.Y. Ku, P. Kulkarni, R. Shisler, W. Wei, Membrane performance requirements for carbon dioxide capture using
hydrogen-selective membrane in integrated gasficatiom combined cycle (IGCC). Journal of Membrane Science
367:233-239.
[5] Y. Huang, S. Rezvani, D. McIlveen-Wright, A. Minchener, N. Hewitt, Techno-economic study of CO2 capture and
storage in coal fired oxygen fed entrained flow IGCC power plant. Fuel Processing Technology 89:916-925.

892
[6] A. Padurean, C.-C. Cormos, P.-S. Agachi, Pre-combustion carbon dioxide capture by gas-liquid adsorption for
Integrated Gasification Combined Cycle power plants. International Journal of Greenhouse Gas Control 7 (2012 )
1-11.
[7] C.-C. Cormos, Evaluation of energy integration aspects for IGCC-based hydrogen and electricity co-production
with carbon capture and storage. International Journal Hydrogen Energy 35:7485-7497.
[8] A. Brunetti, F. Scura, G. Barbieri, E. Drioli, Membrane technology for CO2 separation. Journal of Membrane
Science 359:115-25.
[9] J.W. Dijkstra, J.A.Z. Pieterse, H. Li, J. Boon, Y.C. van Delft, G. Raju, G. Peppink, R.W. van den Brink, D.
Jansen, Development of membrane reactor technology for power production with pre-combustion CO2 capture.
Energy Procedia 4:715-722.
[10] S.-K. Ryi, J.-S. Park, K.-R. Hwang, C.-B. Lee, S.-W. Lee, Module configuration in CO2 capture using Pd-based
composite membranes. International Journal of Hydrogen Energy 36:13769-13775.
[11] S.H. Lee, J.N Kim, W.H Eom, S.-K. Ryi, J.-S. Park, I.H. Baek, Development of pilot WGS/multi-layer
membrane for CO2 capture, Chemical Engineering Journa 207-208: 521-525.
[12] S.-K. Ryi, J.-S. Park, K.-R. Hwang, C.-B. Lee, S.-W. Lee, The property of hydrogen separation from CO2
mixture using Pd-based membranes for carbon capture and storage (CCS). International Journal of Hydrogen
Energy 38:7605-7611.
[13] C.A. Scholes, K.H. Smith, S.E. Kentish, G.W. Stevens, CO2 capture from pre-combustion processes -
Strategies for membrane gas separation. International Journal of Greenhouse Gas Control 4:739-755.
[14] S.-K. Ryi, J.-S. Park, S.-H. Kim, S.-H. Cho, J.-S. Park, D.-W. Kim, Development of a new porous metal
support of metallic dense membrane for hydrogen separation. Journal of Membrane Science 179:439-445.
[15] S.-K. Ryi, J.-S. Park, K.-R. Hwang, C.-B. Lee, S.-W. Lee, Repair of Pd-based composite membrane by
polishing treatment. International Journal of Hydrogen Energy 36:13776-13780.
[16] K.-R. Hwang, S.-K. Ryi, C.-B. Lee, S.-W. Lee, J.-S. Park, Simplified, plate-type Pd membrane module for
hydrogen purification. International Journal of Hydrogen Energy 36:13769-13775.
[17] S.H. Lee, J.N. Kim, W.H. Eom, Y.D. Ko, S.U. Hong, I.H. Back, Development of water gas shift/membrane
hybrid system for precombustion CO2 capture in a coal gasification process. Energy Procedia 4:1139-1146.
[18] S.-K. Ryi, C.-B. Lee, S.-W. Lee, J.-S. Park, Pd-based composite membrane and its high-pressure module for
pre-combustion CO2 capture. Energy 51:237-242.
[19] D. Mendes, A. Mendes, L. M. Madeira, A. Iulianelli, J. M. Sousa, A. Basile, The water-gas shift reaction: from
conventional catalytic systems to Pd-based membrane reactors - a review. Asia-Pacific Journal of Chemical
Engineering 5:111-137.
[20] K.-R. Hwang, S.-W. Lee, S.-K. Ryi, D.-K. Kim, T.-H. Kim, J.-S. Park, Water-gas shift reaction in a plate-type
Pd-membrane reactor over a nickel metal catalyst. Fuel Process Technology 106:133-140.
[21] F. Guazzone, E.E. Engwall, Y.H. Ma, Effects of surface activity, defects and mass transfer on hydrogen
permeance and n-value in composite palladium-porous stainless steel membranes. Catalysis Today 118:24-31.
[22] B. Leffler, STAINLESS-stainless steels and properties, http://www.hazmetal.com.
[23] H. Li, A. Goldbach, W.Z. Li, H.Y. Xu, CO2 decomposition over Pd membrane surface. Journal of Physics
Chemistry B 112:12182-4
[24] F.C. Gielens, R.J.J. Knibbeler, P.F.J. Duyxinx, H.D. Tong, M.A.G. Vorstman, J.T.F. Keurentjes, Influence of
steam and carbon dioxide on the hydrogen flux through thin Pd/Ag and Pd membranes. Journal of Membrane
Science 279:176-85.

893
 MATHEMATICAL MODEL OF PETROLEUM DYNAMICS IN CLOSED CONDUIT
1 2
Sarunya Promkotra , Tawiwan Kangsadan
1,2
Assistant Professor, Dr.
1
Department of Geotechnology, Faculty of Technology, Khon Kaen University, Khon Kaen 40002, Thailand
2
Chemical and Process Engineering Program, Department of Mechanical and Process Engineering,
The Sirindhorn International Thai-German Graduate School of Engineering (TGGS),
King Mongkut’s University of Technology North Bangkok (KMUTNB), Bangsue, Bangkok 10800, Thailand
1 2
e-mail: sarunya@kku.ac.th , tawiwan.k.cpe@tggs-bangkok.org

ABSTRACT
Petroleum transport from a gas well to the gas production plant is relevant to the petroleum production of the
natural gas field. This migration affects movement direction on distance and chemical components under the
emphasis upon the relationship between condensate and other fluid hydrocarbons. The physical properties of
condensate are analyzed in the basis of API gravity (55.52 – 62.84o), density (0.73 – 0.76 g/cm3), specific gravity
(0.7362-0.7566), Reid vapor pressure and true vapor pressure. The highest hydrocarbon composition is methane,
96.70%. Hydrocarbon concentrations from the gas processing plant are found overall C 7+ (heptane, C7H16) to C16
(hexadecane, C16H34), and found most octane (C8H18). Petroleum dynamics is evaluated in the criteria of
petroleum transport in closed conduit. Mathematical model of petroleum dynamics is defined by the diffusion
coefficient of the specific chemical reactions depending on methane-ethane and methane-propane which are
15.71 cm2/sec and 9.20 cm2/sec, respectively. The quantities of molar fluxes gradually increase from one location
to another location when phases change from gas to liquid, and decrease to the condensate storage tank due to
the completion of liquid. The liquid phase preferably occurs because propane is slightly heavier than ethane. Their
physical properties have less affected to molar flux in the chemical reaction of propane. According to methane-
ethane combination, molar fluxes as a function of sampling distances of methane in reactants are higher than
methane in products. Besides, the molar fluxes of other components are expressed in the systematic trend.

INTRODUCTION
Condensate is a liquid hydrocarbon phase with low density and high API gravity which usually takes place in
correlation with natural gas. Its occurrence as a liquid phase relies on temperature and pressure states in petroleum
reservoir permitting condensation of liquid from vapor. Productivity of condensate from petroleum reservoirs is
complex due to the pressure sensitivity. Condensate becomes increasingly essential due to the growing world
dependence on the soaring of petroleum products in the marketplace. The recent comprehension of gas-
condensate flow behavior within reservoir rocks is inadequate because of the limited data accessible. Besides, this
fluid withdrawal from reservoir rocks to the gas well is required of accurate relative data for petroleum dynamics in
both gas-condensate reservoirs and the gas well. During condensate production, fluid transportation from the well
and the gas processing plant has to maintain pressure difference. If the controlled pressure drops below the dew
point during production, the condensate will condense from gas to liquid. According to the operation within the gas
processing plant in Num Phong District, Khon Kaen Province, the northeast of Thailand, mostly methane gas is
discovered approximately 96%mole (Sattayarak, 2005). Due to physical and chemical properties, this petroleum
reservoir indicates the dry gas reservoir (PTTEP, 2006). The pressure and temperature from a petroleum reservoir to
the ground surface change from 160 bar.g and 60oC to 74.54 bar.g and 31.67oC. While petroleum fluid flows to the
slug catcher, condensate mixture in this location is separated from natural gas. For the liquid processing, liquid
petroleum mixture consists of condensate and water. This mixture flows to the high pressure flash vessel (pressure
6.49 bars, temperature 24.28oC) which condensate is separated from water. Besides, condensate flows to the low
pressure flash vessel and evaporates. At last, it is kept in a storage tank (a condensate tank). Having a high liquid
hydrocarbon yield is often utilized for liquid petroleum reserves. Temperature and pressure in each location of the gas
processing plant are effective to create the chemical thermodynamic simulation and also its fluid dynamics (Fan et al.,
2006).

Petroleum dynamics is the main study of this research. This foremost scheme is aimed to identify the petroleum
flow in the closed conduit from the gas well to the gas processing plant. To investigate petroleum migration with
emphasis upon the relationship between condensate and other fluid hydrocarbons, mathematical fluid dynamics of
petroleum mixture in one direction is evaluated to the petroleum production in the gas production plant. This study
will provide a more understanding in petroleum migration from the gas well to cylindrical pipe and finally to process
production plant. The distance from the wells to the gas plant is approximately 62 kilometers. High efficiency well
for producing the natural gas can produce the natural gas about 60-140 MMscfd. Samples of natural gas and
condensate from the gas field are collected from the gas processing plant at this point. At present, petroleum
dynamics depended on petroleum migration has been studied by simulation, but not by mathematical modeling.
This modeling will scope down to the simply understanding in the specific location within the gas plant. Thus, these
studied results will be explicit the relationship of petroleum dynamics in specific direction and areas.

894
METHODOLOGY
Condensate and natural gas are collected during August to October 2012 from the gas processing plant at the inlet
slug catcher, high pressure flash vessel and condensate tank, as shown in Fig.1. Collecting sample is covered during
rainy season which affects on deeply groundwater. Generally, condensate, water and natural gas mix together by
ultimate condition. Thus, water from fluid mixture has to be removed from condensate at the outlet slug catcher,
where is condensate and water unblended. Analytical results of this water indicate remarkably fresh water, confirmed
by the total dissolved solids (TDS) with a value below 200 mg/L. Moreover, condensate sample from the inlet slug
catcher (upstream) has approximately the same composition likewise the natural hydrocarbon from the petroleum
reservoir. Therefore, this state is assumedly the representative of liquid petroleum from the petroleum well.
Expectation of the geochemical relation of condensate and natural gas would be gained (1) physical properties of
condensate and natural gas, (2) geochemical compositions related to physical properties and (3) conceptual stability
condition corresponding to condensate and natural gas.

Condensate and Natural Gas

Physical properties of condensate are examined for density, specific gravity, API gravity, followed by ASTM D5191.
Reid vapor pressure (RVP) and true vapor pressure (TVP) are followed by the analytical methods of the vapor
pressure of petroleum products (Mini Method). These parameters are always measured daily for maintaining the
quality of condensate. Hydrocarbon compositions of both condensate and natural gas are determined by the gas
chromatography (GC), GC-2010 Plus (Shimadzu), based on ASTM D1945. Its capillary column is HT-5 type with 25
meter in length and a 0.10 μm film thickness. Hydrogen gas is used as a carrier gas with the flow rate of 47 mL/min.
The column over temperatures sets at 170, 270 and 380°C, with a temperature ramp of 15, 8 and 15oC/min.
Detection of the analyses is performed using the flame ionization detector (FID). Internal standard is principally
methyl haptadecanoate (C18H36O2). It will be referenced in the calibration standard, and used to calculate
concentrations as a marker of analytical chromatograms.

7 8 Process in Gas Processing Plant

L W
4 6
L L
L, W
3 5
L, W L L
1 = Inlet slug catcher process
1*= Inlet coalescing filter separator
2 = Outlet slug catcher process
3 = High pressure flash vessel
4 = Low pressure flash vessel
(No water in process)
5 = Produced water separator
6 = Remaining water
7 = Condensate Tank
8 = Thermal oxidizer

2 Note
L = Liquid petroleum or condensate
L, W W = Water
G = Gas
G
1 1* = Collecting sample

Gas-Condensate-Water Mixture

Fig. 1. Flow chart of gas-condensate-water processes of the gas processing plant.

Hydrocarbon Transport in Conduit

Fluid hydrocarbon will be withdrawal from the reservoir and migrated from the gas well to the gas processing plant.
The distance between two locations is approximately 62 kilometer through cylindrical pipe. This pipe is lined down
under the surface around two meters, as shown in Fig. 2. The hydrocarbon components are assumable the same
compositions as situated in the reservoir rocks. This mathematical dynamics focuses on the migration of hydrocarbon
masses in one direction. Furthermore, the hydrocarbon transport is considered to be in the laminar flow regime,
without any angle in the cylindrical pipe.

895
 Gas Well Gas Processing Plant
62 km

r Ci (r, , z)
Ci (z) z

Fig. 2. Schematic diagram of cylindrical pipe where hydrocarbon migrated from the gas well to the gas processing plant.

Cylindrical Coordinate System

Three-dimensional coordinate system is designed by the cylindrical coordinate system in specific point based on
the position or distance from a selected reference axis. The direction from the axis relates to the reference
direction and distance from a selected reference perpendicular to the axis. The radial distance or the distance from
the axis refers to the angular position or as the azimuth. Two-dimensional polar coordinate system (both
combination of the radius and the azimuth) correspond to A in the plane through the point, analogous to the
reference plane. Longitudinal position (or axial position) is defined as the altitude as the third coordinate, if the
reference plane is considered horizontal. Cylindrical coordinates are practical in association to purposes and
occurrences that have some gyratory symmetry concerning the longitudinal axis, for example water flow in a
conduit or closed canal for circular cross-section. The following mathematical summary of cylindrical coordinate
system is derived by Kangsadan (2012).

Ci (t, r, , z,) = Ci(z) (1)

Assumption
1. At steady state, Ci is not a function of time (t).
2. No concentration distribution in radial direction; Ci is not a function of radius (r).
3. No concentration distribution in radial direction; Ci is not a function of angular (  ).

In which, Ci = Molar concentration of component i

z = Distance (m)

Mechanism of Mass Transport

Mass transport is the transport of the one constituent from a region of higher concentration to that of a lower
concentration. The net flow of each molecular species occurs in the direction of the negative concentration
gradient. Since mass transfer or diffusion occurs only in mixture, its evaluation must involve the examination of the
effect of each component.

Molecular Mass Transfer or Diffusion

Molecular of mass transfer, sometimes called diffusion mass, can be transferred by random molecular motion in
quiescent fluids independent of any convection within system. Convection mass transfer mass can be transferred
from surface into a moving fluid as follows:

Mass or MolarRate
Mass or molar fluxes: Flux = (2)
SurfaceArea

Mass or molar flux of a given species is a vector quantity denoting the amount of the particular species, in either
mass or molar units, that passes per given increment of time through a unit area normal to the vector. Flux must be
defined a reference to the following condition:

Stationary Reference: Coordinates are fixed in space; Naz = Molar flux of component A relative to a stationary.
CA
Equation of continuity:   NA  - RA = 0 (3)
t
In which,  = del or gradient operator, the vector derivative of a scalar

896
 CA
= molar concentration of component A gradient in time-direction
t
RA = rate of component A
C A
Generalize expression for fluxes; defined no reaction RA = 0 and at steady state: =0
t
dNAZ
Thus, = 0
dz
dCA
NAZ = - DAB + XA (NAX + NBX ) (4)
dz

In which, XA = mole fraction (mole of A divided by total mole)

DAB = mass diffusivity or diffusion coefficient for component A diffusing through component B (cm2.s-1)

Diffusion Coefficient
The Fuller correlation is estimating the diffusion coefficient for nonpolar, binary at low pressure (Fuller et al., 1996).
Empirical correlation recommended by

1
1 1 2
10-3 T1.75 ( + )
MA MB
DAB = 2
. (5)
 1 1
P   v A3 +   v B3 
 
 

In which, T = absolute temperature (K)

MA = molecular weight of diffusing species A (g/mol)
MB = molecular weight of diffusing species B (g/mol)
P = pressure (atm)
ΣvA,B = sum of the diffusion volume for component A, B

RESULTS AND DISCUSSION

These research studies are to exemplify parameters which are important for gas processing control in the gas
processing plant. Experimental results of condensate and natural gas obtain physical properties and chemical
compositions which are useful for the basic information of fabricating mathematical modeling. Mathematical fluid
dynamics applies to identify petroleum in a closed conduit related to the petroleum production. Parameters of
condensate, such as temperature and pressure are recorded in the production plant in each process. Both
parameters are automatically monitored by time. Detailed data are compared in Table 1. Besides, concentration of all
components (in mol%) of the natural gas and condensate are plotted all together, as illustrated in Fig. 3. Different
temperature and pressure depend upon various processes in the gas processing plant. Maximum and minimum
pressures are found approximately to be 70 and 0.02 bar in the inlet slug catcher and condensate tank, respectively.
Reduced pressure is subjected to control the condensate condition in liquid and to separate the natural gas. These
pressures related to the processes in the gas production plant will gradually be decreased in each location from the
inlet to condensate tank. Temperature gradients are more dominant than pressure which are concerning by the
controlling process. Both temperature and pressure play a critical role for phase change and phase equilibria from the
petroleum reservoir, gas well and gas plant. Thus, chemical thermodynamics and petroleum dynamics are the
nutshell for conceptual correlation.

Physical properties of petroleum-field waters are determined on site during sample collection, as indicated in Table 2.
Water samples obtain from fluid samples which consist of condensate and also petroleum-field waters. Some
parameters are not available for measurement, such as electrical conductivity (EC) and total dissolved solid (TDS). At
this present, water combined in the hydrocarbon system is found less and not enough for further collection. The
measurements of both EC and TDS indicate that these petroleum-field waters are in the condition of fresh water. In
general, water in any deep aquifers will present itself as brackish water (or saline). However, these water samples at
this time are not affected at all. This phenomenon are totally shown distinguished results of petroleum-field water in
2009 and 2011 (Kumchawong, 2009; Buaphai, 2011).

897
Table 1. Measuring parameters (average) in each location.
Inlet Slug Outlet Slug High Pressure Flash Low Pressure Flash Condensate
Parameters
Catcher Catcher Vessel Vessel Tank
Pressure (bar) 65.55 65.55 6.95 0.265 0.02
Temperature (oC) 29.03 29.03 42.78 54.65 29.5
Density (g/cm3) 0.7550 0.7400 0.7450 0.7300 0.7400
Specific Gravity 0.7561 0.7473 0.7379 0.7311 0.7414
API gravity (o) 55.65 57.85 59.93 62.06 59.36
RVP (psi) 0.83 0.98 1.55 2.51 1.42
TVP (psi) 1.43 1.59 1.735 3.105 2.05

Table 2. Physical properties of petroleum-field waters.

Inlet Slug Outlet Slug High Pressure Flash Low Pressure Produced Water Condensate
Parameters
Catcher Catcher Vessel Flash Vessel Separator Tank
pH 7.09 6.32 6.42 5.81 7.28 6.38
Temperature (oC) 41.10 27.90 27.65 33.50 28.95 26.30
EC (µS/cm) - 334.0 239.0 - 411.4 -
TDS (mg/L) - 149.9 105.0 - 186.7 -

Methane Gas
100
Condensate

10
Concentration (mol%)

1
Natural Gas

0.1

0.01
C1 C2 C3 iC4 nC4 iC5 nC5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 CO2 N2 Water

Chemical Components

Natural Gas Inlet Slug Catcher High Pressure Flash Vessel Condensate Tank

Fig. 3. Relationship of concentration (in mol%) and chemical components compared to natural gas and condensate.

Hydrocarbon Transport in Closed Conduit

Due to the chemical thermodynamics of fluid hydrocarbons, natural gas and condensate, the conceptual model for
mass transport is constructed by concentrations of hydrocarbon components. These fluid movements can present the
migration of hydrocarbon in closed conduit. Distance from petroleum reservoir to the gas well and transport to the
processing plant is very complex computation in mathematic scheme. Thus, this hydrocarbon transport is indicated
only for the gas well to the gas processing plant. Assumption for this transport depends on the chemical reaction in
equilibrium condition and one direction from the well to the gas plant. Prediction of parameters generates mainly from
temperature and pressure. Methane is dominantly a precursor in every chemical reaction. Moreover, other gases
related to the genesis within reservoir and exist in the same environment are combined as a reactant. The results of
GC from chemical laboratory indicates carbon dioxide, hydrogen gas and nitrogen in which reservoir are 10, 8, and 2
mole percent, respectively. Chemical equations are conducted by using petroleum components from condensate
which is collected in the gas processing plant. There are simply two chemical reactions creating based on following
criteria.

Computational Criteria
Variable chemical species depend on reactants and products, as indicated in two chemical reactions, number 1 and
2. Diffusion coefficients are indicated in Table 3 for calculation of multi-component gas mixtures and binary diffusion
coefficients (cm2/sec) at STP (temperature (273 K), pressure (1 atm)).

898
Reaction No.1: 4CH4 (g) + 2C2H6 (g) + 2CO2 (g) + H2 (g) + N2 (g)  2CH4 (g) + C7H16 (l) + CO2 (aq) + N2 (aq) + 2H2O (aq) + H2 (aq) (6)
Reaction No.2: 3CH4 (g) + C3H8 (g) + 3CO2 (g) + 4H2 (g) + N2 (g)  CH4 (g) + C7H16 (l) + CO2 (aq) + N2 (aq) + 4H2O (aq) (7)

Table 3. All compositions and criteria depending on Eq.1 and Eq.2.

Symbol and Components Concentration (CA) Diffusion Coefficient (cm2/sec) References for Diffusion Coefficient
H = H2 CH = concentration of H2 1.60 http://www.thermopedia.com/content/
C = CO2 CC = concentration of CO2 0.11 http://www.thermopedia.com/content/
N = N2 CN = concentration of N2 0.19 http://www.thermopedia.com/content/
W = H2O CW = concentration of H2O 0.28 http://www.thermopedia.com/content/
1 = CH4 C1 = concentration of CH4 0.18 http://www.thermopedia.com/content/
2 = C2H6 C2 = concentration of C2H6 1.52 x 10 -16 Witherspoon and Saraf (1965)
3 = C3H8 C3 = concentration of C3H8 1.21 x 10 -16 Witherspoon and Saraf (1965)
7 = C7H16 C7 = concentration of C7H16 3.67 x 10 -16 Leahy-Dios et al. (2008)

Calculation Outlines for Chemical Reaction 1

Mass (or chemical species) transports (NAz) can be calculated by Eq. (4). The minus sign of this diffusion equation
refers to the direction of mass from high to low concentration. The different concentrations have to be defined as
the direction, based on the Fick’s first law of diffusion. Also, diffusion coefficient (DAB) is calculated by using Eq. (5).
The subscript numbers refer to each hydrocarbon type and also their components indicating on the chemical
reaction (Reaction No.1), D1-2-C-H-N-1*-7*-C*-N*-W*-H* . Substitute the diffusion coefficient (v) of each species from
Table 3 into Eq.(8). Then,

D1-2-C-H-N-1*-7*-C*-N*-W*-H* = DAB
1
-3 1.75  1 1 1 1 1 1 1 1 1 1 1 2
10 (T)  + + + + + + + + + + 
  M 1 M 2 M C M H M N M 1* M 7* M C* M N* M W* M H* 
2 (8)
 1 1 1 1 1 1 1 1 1 1 1 
P   v 13 +   v 23 +   v C3 +   v H3 +   v N3 +   v 1*3 +   v 7*3 +   v C*  N*  W*  H*
3 + v 3 + v 3 + v 3 
 
 
1
-3 1.75  1 1 1 1 1 1 1 1 1 1 1 2
10 (273 K)  + + + + + + + + + + 
 4(16) 2(30) 2(44) 2 28 2(16) 100 44 28 2(18) 2 
 2
 1 1 1 1 1 1 1 1 1 1 1 
(1 atm)  0.18 13 + 1.5223 +  0.11C3 + 1.60 H3 +  0.19 N3 +  0.18 1*3 +  3.67 7*  C*  N*  W*  H*
3 + 0.11 3 + 0.19 3 + 0.28 3 + 1.60 3 
 
 

20.1420
= = 15.71 cm2/sec.
1.2820

Thus, diffusion coefficient (DAB) of this chemical reaction is 15.71 cm2/sec. This value will be used for Eq. (4). The
dN1z
term of N1Z in this equation is a molar flux of the methane reactant (CH4 (g). Define the = 0, then
dz
dC C * * * *
N1Z = - DAB 1 + 1 (N2Z + NCZ + NHZ + NNZ + N1Z + N7Z + NCZ + NNZ + N*WZ + NHZ
*
). (9)
dz CT

From the chemical reaction, the molar flux of N1Z from Eq. (9) is rearranged to other side in terms of reactants and
products. For example, methane is expressed as the N1Z in this reaction. The results of each term in Table 4 will be
replaced back in Eq.(9). Table 4 presents the relationship of the methane and other species in the chemical
equation. According to Eq.(9), the molar flux of methane can be completed as follows:

d(2,903.991) 2,903.991
N1Z = – (15.71) + (0.5 N1Z + 0.5 N1Z + 0.25 N1Z + 0.25 N1Z
dz 3,531.467
- 0.5 N1Z - 0.25 N1Z - 0.25 N1Z - 0.25 N1Z - 0.5 N1Z - 0.25 N1Z )

(2,903.991- 0) (mol%/m3.sec)
N1Z = –15.71(cm2/sec) + (0.822)(-0.5 N1Z) = – 5,214.393 (mol%/m2.sec).
(62 - 0) km

899
Table 4. Defined terms for calculation of molar flux of the methane reactant (N1Z) in chemical reaction number 1.
Reactants Products (*)
4N2Z *
N1Z = , then N2Z = 0.5 N1Z 4N
N1Z =  1Z , then N*1Z = - 0.5 N1Z
2 2
*
N1Z = 4NCZ , then NCZ =0.5 N1Z N1Z =  4N7Z , then N7Z
*
= - 0.25 N1Z
2 1
*
N1Z = 4NHZ , then NHZ = 0.25 N1Z N1Z =  4NCZ , then NCZ
*
= - 0.25 N1Z
1 1
*
N1Z = 4NNZ , then NNZ = 0.25 N1Z N1Z =  4NNZ , then NNZ
*
= - 0.25 N1Z
1 1
*
N1Z =  4NWZ , then N*WZ = - 0.5 N1Z
2
– *
N1Z =  4NHZ , then NHZ
*
= - 0.25 N1Z
1

Then, the diffusion coefficient of the methane reactant is finally revealed as 5,214.393 mol%/m2.sec. The minus
sign of this diffusion of molar flux (N1Z) refers to the direction of methane from the gas well to the gas plant. The
concentration and condition for calculating the molar flux in the inlet slug catcher are shown in Table 5-7.
According to the gas processing plant, flow rate of petroleum fluid per day is 7,000,000 scf or 1,982,179.261
m3.The diffusion coefficient of the chemical reaction (Reaction No.1) is 15.71 cm2/sec. Distance from the gas well
to the production plant is 62 km; the distance from the inlet slug catcher to the high pressure flash vessel is 0.05
km; and the distance from the high pressure flash vessel to the condensate tank is 0.10 km, respectively.
Computational results of molar fluxes of the inlet slug catcher, high pressure flash vessel and condensate tank
based on the chemical reaction (Reaction No.1) are presented in Table 8.

Table 5. Defined terms for calculation of molar flux.

Chemical Species CH4 C2H6 CO2 H2 N2 CH4 C7H16 CO2 N2 H2O H2 Sum of Each Species
Reactants
Methane N1Z 0 0.5 0.5 0.3 0.3 -0.5 -0.3 -0.3 -0.3 -0.5 -0.3 -0.5
Ethane N2Z 2 0 1 0.5 0.5 -1 -0.5 -0.5 -0.5 -1 -0.5 0
carbon dioxide NCZ 2 1 0 0.5 0.5 -1 -0.5 -0.5 -0.5 -1 -0.5 0
Hydrogen gas NHZ 4 2 2 0 1 -2 -1 -1 -1 -2 -1 1
Nitrogen gas NNZ 4 2 2 1 0 -2 -1 -1 -1 -2 -1 1
Products
Methane N1Z -2 -1 -1 -0.5 -0.5 0 0.5 0.5 0.5 1 0.5 -2
Heptane N2Z -4 -2 -2 -1 -1 2 0 1 1 2 1 -3
Carbon dioxide NCZ -4 -2 -2 -1 -1 2 1 0 1 2 1 -3
Nitrogen NHZ -4 -2 -2 -1 -1 2 1 1 0 2 1 -3
Water NWZ -2 -1 -1 -0.5 -0.5 1 0.5 0.5 0.5 0 0.5 -2
Hydrogen NHZ -4 -2 -2 -1 -1 2 1 1 1 2 0 -3

Calculation Outlines for Chemical Reaction 2

For the chemical reaction number 2, mass (or chemical species) transport can also be calculated by Eq.(4) and
the diffusion coefficient (DAB) is calculated by Eq.(5). The subscript numbers refer to each hydrocarbon type and
also their components indicating on the chemical reaction (Reaction No.2), D1-3-C-H-N-1*-7*-C*-N*-W*. Substitute the
diffusion coefficient (v) of each species from Table 3 into Eq.(8). Then, D1-3-C-H-N-1*-7*-C*-N*-W* = DXY.

900
Table 6. Criteria for molar flux calculation (Reaction No. 1) of the inlet slug catcher.
Diffusion Flux Diffusion Flux
Reactants Products
(mol%/ft )3 (in 100%) (mol%/m3) (mol%/ft3) (in 100%) (mol%/m3)
Methane 96.72 82.236 2904.12 Methane 96.72 76.128 2688.432
Ethane 0.893 0.760 26.825 Heptane 20.168 15.874 560.590
Carbon dioxide 10 8.502 300.261 Carbon dioxide 0.503 0.396 13.973
Hydrogen gas 8 6.802 240.208 Nitrogen 1.659 1.305 46.103
Nitrogen gas 2 1.7 60.052 Hydrogen 8 6.297 222.368
Total 117.613 100 3531.467 Total 127.049 100 3531.467

Table 7. Results of molar flux calculation (Reaction No. 1) of the inlet slug catcher.
Concentration Diffusion Mole Molar
Chemical Species Parameter of Bulk Motion Contribution Molar Flux
Gradient Flux Fraction Flux
Reactants (mol%/m3.km) (mol%/m2.sec) (XA) Constant Any Species (Nz) Specific Species (Niz) (mole%/m2.sec)
CH4 (g) N1Z 46.841 7358.666 0.822 -0.5 -0.411 1.411 5214.557
C2H6 (g) N2Z 0.433 67.972 0.008 0 0 1 67.972
CO2 (g) NCZ 4.843 760.822 0.085 0 0 1 760.822
H2 (g) NHZ 3.874 608.657 0.068 1 0.068 0.932 653.079
N2 (g) NNZ 0.969 152.164 0.017 1 0.017 0.983 154.797
Products
CH4 (g) N1Z 43.362 6812.141 0.761 -2 -1.523 2.523 2700.489
C7H16 (l) N2Z 9.042 1420.464 0.159 -3 -0.476 1.476 962.227
CO2 (aq) NCZ 0.225 35.406 0.004 -3 -0.012 1.012 34.991
N2 (aq) NNZ 0.744 116.818 0.013 -3 -0.039 1.039 112.415
H2 (aq) NHZ 3.587 563.453 0.063 -3 -0.189 1.189 473.926

Table 8. Results of molar flux calculation (Reaction No. 1) of the inlet slug catcher, high pressure flash vessel and condensate tank.
Chemical Inlet Slug Catcher High Pressure Flash Vessel Condensate Tank
Chemical Species
Reaction No.1 Molar Fluxes (mol%/m .sec)
2

Reactant CH4 (g) N1Z 5,214.56 6,466,050.47 3,233,025.23

Reactant C2H6 (g) N2Z 67.972 84,285.03 42,142.51
Reactant CO2 (g) NCZ 760.822 943,418.72 471,709.36
Reactant H2 (g) NHZ 653.079 809,818.38 404,909.19
Reactant N2 (g) NNZ 154.797 191,947.79 95,973.90
Total Reactants 6,851.23 8,495.520.382 4,247,760.19
Product CH4 (g) N*1Z 2,700.49 3,387,113.32 1,688,807.91
Product C7H16 (l) N*7Z 962.227 1,295,731.22 635,374.58
Product CO2 (aq) N*CZ 34.991 43,397.94 21,697.80
Product N2 (aq) N*NZ 112.415 139,493.79 69,734.78
Product H2 (aq) N*HZ 473.926 589,432.74 294,500.21
Product H2O (aq) N*WZ - - -
Total Products 4,284.05 5,455,169.01 2,710,115.27
Total of Chemical Reaction 2,567.18 3,040,351.38 1,537,644.92

901
D1-3-C-H-N-1*-7*-C*-N*-W* = DXY
1
 1 1 1 1 1 1 1 1 1 1 2
10-3 (273 K)1.75  + + + + + + + + + 
 3(16) 44 3(44) 4(2) 28 16 100 44 28 4(18)  = 9.20 cm2/sec.
 2
 1 1 1 1 1 1 1 1 1 1 

 2
 
(1 atm)  0.18 13 + 1.21 x 10-16 3 +  0.11C3 + 1.60 H3 +  0.19 N3 +  0.18 1*3 + 3.67 x10-16 
3 +
7*
0.11C*3 + 0.19N*3 + 0.28W*
3 

 

Thus, diffusion coefficient (DAB) of this chemical reaction equals to 9.20 cm2/sec. This value will be used for Eq.(4).
dN1z
The term of N1Z for this reaction is a molar flux of methane reactant (CH4 (g). Define the = 0, then
dz

dC1 C1
N1Z = - DAB + *
(N3Z + NCZ + NHZ + NNZ + N1Z *
+ N7Z *
+ NCZ *
+ NNZ + N*WZ ) . (10)
dz CT

From the chemical reaction (Reaction No.2), the molar flux of N1Z from Eq.(10) is rearranged in terms of reactants
and products. For example, methane is expressed as the N 1Z in this reaction. The diffusion coefficient of the
methane reactant is finally revealed as 5,981.509 mol%/m2.sec. Computational results of molar fluxes of the inlet
slug catcher, high pressure flash vessel and condensate tank of this chemical reaction are presented in Table 9.

Table 9. Results of molar flux calculation (Reaction No.2) of the inlet slug catcher, high pressure flash vessel and condensate tank.
Chemical Inlet Slug Catcher High Pressure Flash Vessel Condensate Tank
Chemical Species
Reaction No.2 Molar Fluxes (mol%/m2.sec)
Reactant CH4 (g) N1Z 5,981.88 7,417,527.97 3,908,956.59
Reactant C3H8 (g) N3Z 40.733 50,509.54 26,793.23
Reactant CO2 (g) NCZ 458.811 568,926.12 335,446.69
Reactant H2 (g) NHZ 356.438 441,983.56 259,915.95
Reactant N2 (g) NNZ 93.9 116,435.82 68,797.98
Total Reactants 6,931.76 8,595,383.00 4,599,910.44
Product CH4 (g) N*1Z 841.013 1,059,932.18 527,854.69
Product C7H16 (l) N*7Z 480.629 639,201.05 314,364.31
Product CO2 (aq) N*CZ 21.67 26,881.99 13,439.63
Product N2 (aq) N*NZ 68.253 84,744.22 42,358.56
Product H2O (aq) N*WZ - - -
Total Products 1,411.57 1,810,759.44 898,017.19
Total of Chemical Reaction 5,520.20 6,784,623.56 3,701,893.25

Petroleum Dynamics for Chemical Reaction 1

Assumption for this mathematical concept for hydrocarbon dynamics is defined for one direction from the gas well to
the gas plant. Both chemical reactions consist of dominantly methane gas and other hydrocarbons, such as ethane
and i-propane, as the reactants. The diffusion coefficient of chemical reaction (No.1) is 15.71 cm2/sec. Molar fluxes of
the inlet slug catcher, high pressure flash vessel and condensate tank for the reactants and products are plotted as a
function of the sampling distances (Fig. 4). Starting from the inlet slug catcher, the quantities of all molar fluxes are
gradually increasing in the high pressure flash vessel because of phase change from gas to liquid and decreasing to
the condensate tank due to the completion of liquid phase. These effects on molar fluxes depend on chemical
species under thermodynamic equilibrium and controlling system in the gas processing plant. The relationship of
molar fluxes as a function of sampling distance are especially expressed in systematic trend for reactants (Fig. 4a),
but complicated for products (Fig. 4b). Overall for the reactants, the molar fluxes of methane are found most which
relevant to concentrations defined at the beginning. The intervals of each molar flux are obviously evident, except
carbon dioxide and nitrogen gas. Obviously distinction of molar fluxes occurs in the products at the high pressure
flash vessel and condensate tank. Relationship of molar fluxes as a function of the sampling distance presents the
different phase at the high pressure flash vessel. Methane and heptanes are found most amounts. By contrast, the
different molar fluxes in high pressure flash vessel compared to the inlet slug catcher and condensate tank are not
systematic tendency because miscible phase between gas and liquid. According to the physical properties of

902
condensate in deferent locations, the API gravity, density, specific gravity and vapor pressure are found low to
high. However, pressures in the controlling system decrease rapidly. Molar fluxes of hydrogen gas (very light gas)
and heptanes (pure liquid) thus increase in the condensate tank, as illustrated in Fig. 4b.
10000000
(a)
1000000
Molar Flux (mol%/m2.sec)

100000 Inlet Slug

Catcher High Pressure
Condensate Tank
10000 Flash Vessel

1000

Methane Ethane
100 Carbon Dioxide Hydrogen Gas
Nitrogen Gas
10
61.98 62.00 62.02 62.04 62.06 62.08 62.10 62.12 62.14 62.16

Distance (km)

10000000
(b)
1000000
Molar Flux (mol%/m2.sec)

100000
Inlet Slug High Pressure
10000 Catcher Flash Vessel Condensate Tank

1000
Methane Heptane
100 Carbon Dioxide Nitrogen
Hydrogen
10
61.98 62 62.02 62.04 62.06 62.08 62.1 62.12 62.14 62.16

Distance (km)
Fig. 4. Molar fluxes as a function of distances (sampling location) of (a) reactants and (b) products for the chemical reaction number 1.

Petroleum Dynamics for Chemical Reaction 2

Chemical equilibrium involved in thermodynamics of this chemical reaction indicates the spontaneous reaction. All
chemical species are representative for evaluation the possibility of phase envelope and applicable to the
assumptions depending on theoretical chemical thermodynamics. The diffusion coefficient of this reaction is 9.20
cm2/sec. Relationship of molar fluxes as a function of sampling distances of the reactants and products is entirely
in the systematic trend. Volatile matters are vanished. For the molar fluxes of reactants (Fig.5a), characteristic
trend is similar to the products, as shown in Fig.5b. The difference of the chemical reaction 1 and 2 is the
hydrocarbon mixture, methane and another (ethane or propane). Propane is slightly heavier than ethane. Thus,
the liquid phase prefers to occur. The physical properties will be affected less in the chemical reaction No.2. Gibbs
free energy (G), enthalpy (H) and equilibrium constant (Keq) at temperature of 0, 15, 25 and 200oC are
calculated, as shown in Table 10. This chemical reaction is spontaneous process and relevant to the assumptions
based on conceptual chemical thermodynamics.

For the chemical reaction No.2, the relationship of methane molar fluxes between reactants and products presents
the large interval between the reactants and products than the chemical reaction No.1 because of different
hydrocarbon types. Phase separation easily occurs in the reaction No.2. Unconformity of relationship of methane

903
molar fluxes between reactants and products depending on the physical properties in the chemical reaction No.1
causes closely molar fluxes at the high pressure flash vessel.

Table 10. Gibbs free energy, enthalpy and equilibrium constant of each chemical reaction.
Reaction Gibbs Free Energy (G) Enthalpy (H) Equilibrium Constant (Keq)
Number (kJ/mol) (kJ/mol) 0oC 15oC 25oC 200oC
1 32.555 -217.900 0.93073 0.92838 0.92695 0.91184
2 -17.94 -310.228 1.01537 1.01033 1.00726 0.97514

10000000
(a)

1000000 Condensate Tank

Molar Flux (mol%/m2.sec)

100000
Inlet Slug
Catcher High Pressure
10000
Flash Vessel
1000 Methane
Propane
Carbon Dioxide
100 Hydrogen gas
Nitrogen gas
10
61.98 62.00 62.02 62.04 62.06 62.08 62.10 62.12 62.14 62.16

Distance (km)

10000000
(b)

1000000 Condensate Tank

Molar Flux (mol%/m2.sec)

100000

Inlet Slug
10000 High Pressure
Catcher
Flash Vessel
1000
Methane
Heptane
100 Carbon Dioxide
Nitrogen

10
61.98 62.00 62.02 62.04 62.06 62.08 62.10 62.12 62.14 62.16

Distance (km)

Fig. 5. Molar fluxes as a function of distance (sampling location) of (a) reactants and (b) products for the chemical reaction number 2.

CONCLUSIONS
This fluid within the reservoir will be brought from underground to a gas processing plant. Methane gas is found most
approximately more than 96 mol%. Hydrocarbon transport is defined only from the gas well to the gas production
plant. Criteria for this transport depend only on two chemical reactions in equilibrium condition. Their physical
properties and chemical composition of natural gas and condensate are examined. Pressures in each process are
decreased staring from the inlet slug catcher to condensate tank. Petroleum-field water indicates fresh water.
Hydrocarbon concentrations of condensate are found C7 + to C16. Dynamic migration of petroleum is pursued in the

904
closed conduit, cylindrical pipe, from the gas well to the gas processing plant. The diffusion coefficient of chemical
reaction No.1 and 2 are 15.71 cm2/sec and 9.20 cm2/sec, respectively. Distinction of molar fluxes occurs in the
products of the chemical reaction No.1. The molar fluxes of the high pressure flash vessel are generally the highest
content because of phase change from gas to liquid, and decrease to the condensate due to the completion of liquid
phase. These effects on molar fluxes depend on chemical species based on thermodynamic equilibrium and
controlling system in the gas processing plant. In general, inclusive the reactants, the molar fluxes of methane are
found most which relevant to concentrations defined at the beginning. The intervals of each molar flux are obviously
evident, except carbon dioxide and nitrogen gas. Hydrocarbon mixture in the chemical reaction No.1 and 2 plays a
critical role for phase changes and affected to the molar fluxes of each chemical species. Moreover, the miscible
phase of methane can cause the variation of their molar fluxes which are affected by temperature controlling during
natural gas and condensate separation.

Acknowledgements
I would like particularly to express my appreciation to Hess (Thailand) Limited to provide all samples and information
for this research. This project could not successfully complete without the kindness of the Chemical Process and
Engineering Program, Department of Mechanical and Process Engineering, The Sirindhorn International Thai-
German Graduate School of Engineering (TGGS) at King Mongkut’s University of Technology North Bangkok
(KMUTNB) to support the GC and several facilities for analyzing all samples.

REFERENCES
Buaphai, P. 2011. Experimental study on clathrate hydrate phase equilibria, B.Sc.’s Thesis in Geotechnology,
Faculty of Technology, Khon Kaen University, Thailand.

Fan, L., B. W. Harris, A. (Jamal) Jamaluddin, J. Kamath, R. Mott, G. A. Pope, A. Shandrygin and C. H. Whitson.
2006. Understanding gas-condensate reservoirs. Oilfield Review 2005/2006, 14-27.

Fuller, E. N., P. D. Schettler and J. C. Giddings. 1966. Prediction of binary gas diffusivities, Retrieved November
15, 2012, from http://www.engin.umich.edu/~cre/course/lectures/eleven/exam4.htm.

Kangsadan, T. 2012. Fundamental concepts of mass transport. Interfacial Transport Phenomena. The Sirindhorn
International Thai–German Graduate School (TGGS) of Engineering King Mongkut’s University of Technology
North Bangkok (Unpublished).

Kumchawong, W. 2009. Phase equilibria of condensate and natural gas corresponding to petroleum-field water in
Sin Phu Horm gas field, B.Sc.’s Thesis in Geotechnology, Faculty of Technology, Khonkaen University, Khon Kaen,
Thailand.

Leahy-Dios, A., L. Zhuo and, A. Firoozabadi. 2008. New thermal diffusion coefficient measurements for
hydrocarbon binary mixtures: Viscosity and composition dependency. The Journal of Physical Chemistry 112:
6,442–6,447.

PTT Exploration and Production Public Company Limited. 2006. Phu Horm Gas Field comes to stream adding
energy security to the northeastern part of Thailand. Retrieved November 1, 2012, from http://www.pttep.com/en/.

Sattayarak, N. 2005. Petroleum potential of the Northeast, Thailand. Proceedings of International Conference on
Geology, Geotechnology and Mineral Resources of Indochaina (GEOINDO 2005), 28-30 November 2005, Khon
Kaen, Thailand. 21-30.

Witherspoon, P. A. and D. N. Saraf. 1965. Diffusion of methane, ethane, propane and n-butane in water from 25o –
43o. The Journal of Physical Chemistry 69: 3,752-3,755.

http://www.thermopedia.com/content/696/#DIFFUSION_COEFFICIENT_TABLE1.

905
A STATISTICAL ANALYSIS OF LEAN MISFIRES IN A GASOLINE ENGINE AND THE EFFECT OF
HYDROGEN ADDITION
a b* b
A. K. Sen , M. A.Ceviz ,and E. Güner
a
Richard G. Lugar Center for Renewable Energy and Department of Mathematical Sciences, Indiana University, 402 N.
Blackford Street, Indianapolis, IN 46202, USA
b
Department of Mechanical Engineering, Faculty of Engineering, University of Atatürk,
Erzurum 25240, Turkey
asen@math.iupui.edu, aceviz@atauni.edu.tr, erdoganguner@atauni.edu.tr

ABSTRACT
A gasoline engine may misfire when operating under lean-burn conditions.This study investigates the effect
of adding hydrogen to the gasoline-air mixture on the misfire phenomenon. The effect of hydrogen addition is
characterized by performing a statisticalanalysis of the indicated mean effective pressure (IMEP) time series
over 1000 engine cycles.Using the coefficient of variation (COV) of the IMEP time series, it is shown that
hydrogen addition can suppress the misfires. In addition, it is demonstrated that the IMEP time series
exhibiting misfire cycles are best modeled by a non-Gaussian probability distribution such as thethree-
parameter generalized logistic distribution, whereas those with no or very little misfire can be modeled by a
normal (Gaussian) distribution.

Keywords: Gasoline engine; Misfire; Hydrogen addition; Indicated mean effective pressure; Statistical analysis

*
Corresponding author (email: aceviz@atauni.edu.tr)

INTRODUCTION
It has long been recognized that the process variables such as pressure in an internal combustion engine
undergo cycle-to-cycle variations (CCV). These variations may become severe under lean-burn conditions, and
for highly dilute mixtures with exhaust gas recirculation(Young, 1981),(Heywood, 1988),(Wanger et al., 1998).A
small amount of CCV (slow burns) leads to unwantedengine vibrations and noise, whereas large CCV
(incomplete burns) result in increased hydrocarbon emissions(Scholl and Russ, 1999). Incomplete burns are
otherwise referred to as misfires. Lean misfires may occur due to an imbalance in air-fuel ratio (too much air but
too little fuel). Since an engine needs a richer mixture for a smooth idle, lean misfires may be more noticeable
when the vehicle is idling. An undesirable consequence of a misfire is that the unburned fuel or flammable
hydrocarbons are delivered to the exhaust manifold where they may ignite unpredictably. The purpose of this
paper is to examine the misfire phenomenon in a gasoline engine and analyze the effect of adding hydrogen to
the gasoline-air mixture on the misfires.

Several researchers have investigated the effect of hydrogen addition on engine performance and exhaust
emissions of gasoline engines under lean conditions. In their experiments, (Apostolescu and Chiriac,
1996)examined the influence of adding small amounts of hydrogen to hydrocarbon-air mixtures. They observed
that hydrogen addition can lead to an increase in brake thermal efficiencyand a reduction in HC emissions. The
study carried out by (D’Andrea et al., 2004) revealed that hydrogen addition improves the work output and
decreases burn duration under lean conditions. However, when operating close to stoichiometry, little difference
in engine performance was seen with hydrogen addition. (Jiet al., 2009,2010) performed a series of experiments
to investigate the effect of hydrogen addition in gasoline engines. In their study, (Ji and Wang, 2009) examined
the effect of hydrogen addition to gasoline at idle condition and observed that hydrogen addition improves the
thermal efficiency and combustion stability. On the other hand, the HC and CO emissions first decreased with
hydrogen addition, but when the hydrogen volume fraction exceeded 14.4%, these emissions began to increase
at idle and stoichiometric conditions. Among others, (Verhelst et al., 2009) presented the variations in thermal
efficiency of a spark ignition engine fueled by gasoline-hydrogen blends, at several engine speeds and loads.
They found that hydrogen offered the possibility of a more flexible load control strategy. At low loads, important
in everyday driving, the efficiency gain from hydrogen addition was significantly high,achieving up to 60%
increase in thermal efficiency.

906
There have been several studies investigating the cycle-to-cycle variations (CCV) in gasoline engines(Ceviz et
al., 2012),(D’Andrea et al., 2004),(Fujikawa et al., 2003),(Galloni, 2009),(Ishii et al., 1997),(Ozdor et al.,
1994),(Wang et al., 2010),(Zervas, 2004). Some of these studies(D’Andrea et al., 2004),(Wang et al.,
2010),(Ceviz et al., 2012)also examined the effect of adding hydrogen to the gasoline-air mixture on the CCV.
However, to our knowledge, the effect of hydrogen addition on misfires in gasoline engines has not received
much attention. In this paper, we examine the lean misfire phenomenon in a gasoline engine, and analyze the
effect of hydrogen addition on the misfire cycles. We characterize the effect of hydrogen addition by performing
a statistical analysis of the indicated mean effective pressure (IMEP) time series over 1000 engine cycles. Using
the coefficient of variation (COV) of the IMEP time series, we show that hydrogen addition can suppress the
misfires. In addition, we demonstrate that the IMEP time series exhibiting misfire cycles are best modeled by a
non-Gaussian probability distribution such as the three-parameter generalized logistic distribution, whereas
those with no or very little misfire can be modeled by a normal (Gaussian) distribution.

EXPERIMENTAL PROCEDURE
The experiments were performed on a FORD MVH-418 spark ignition engine with electronically controlled fuel
injectors. A schematic layout of the experimental set up and the engine specifications can be found in anearlier
paper (Ceviz et al., 2011). The engine was coupled to a hydraulic dynamometer. A lean combustible mixture with
air-fuel ratio of 1.26 was used, and hydrogen was added to thegasoline-air mixture in the amounts 2.14%, 5.28%
and 7.74% by volume. The engine was operated at 2000 rpm.The engine speed was measured using a 60-tooth
sprocket and a magnetic pickup. The in-cylinder pressure measurements were made with a KISTLER
6117BFD17 quartz pressure transducer connected to a charge amplifier. The crank angle was measured by an
optical shaft encoder. The optical shaft encoder was rigidly mounted on the front of the engine and connected to
the crankshaft with a flexible coupler. The in-cylinder pressure data from the charge amplifier and crank angle
data from the encoder were recorded by National Instruments Corporation’s high speed M-series, 6250 model,
1.25 MS/s DAQ Card in an Intel 4 CPU 2.00 GHz personal computer. The software package NI-DAQmx from the
National Instruments Corporation was used for data collection. The pressure data were recorded over 1000
consecutive engine cycles. From the pressure measurements, the IMEP values were calculated as follows. The
IMEP is defined as the average pressure in the cylinder over one engine cycle, and is given by: IMEP = Wc / Vd ,
where Vd is the engine displacement volume, and Wc is the amount of work done per cycle: Wc   PdV . Here P
represents the actual (i.e., measured) pressure inside the engine cylinder.

STATISTICALMETHODOLOGY
As mentioned in the introduction, we investigated the misfire phenomenon by performing a statistical analysis of
the IMEP time series. Consider a time series {xi }, i = 1, 2, 3, …,N. For this time series, the various statistical
parameters, namely, mean (μ), standard deviation (σ), coefficient of variation (COV), skewness (S) and kurtosis
(K) are defined by:
1 N
1 N 

N
 xi ,  2 
i 1

N i 1
( xi   ) 2 , COV  x 100,

(1a)

N N
1 1
N
 ( xi   )3 N
 (x  )
i
4

S  i 1
,K  i 1
(1b)
 3
4
Note that the COV is usually expressed in percent form. The COV is a useful statistic for comparing the degree
of variation between two time series even when their mean values are quite different from each other. The
skewness (S) is a measure of asymmetry in the distribution of data in the time series. A distribution with zero
skewness, such as a normal distribution, is symmetric, i.e., it looks the same to the left and right of the center

907
point or the mean value. A distribution with a negative skewness has a longer tail to the left indicating the
presence of small values in the time series. On the other hand, a distribution with a positive skewness has a
longer right tail. The kurtosis (K) is a statistical measure that indicates whether the data distribution is peaked or
flat relative to a normal distribution. A normal distribution has a kurtosis of 3. Distributions with high kurtosis (K>
3) tend to be peakedaround the mean, decrease rapidly, and have heavy tails, whereas those with low kurtosis
(K< 3) tend to have a flat top around the mean. Note that since a normal distribution is symmetric about the
mean, the probability of seeing a normally-distributed value that is far (i.e., more than a few standard deviations)
from the mean is very low. In other words, a normal distribution does not exhibit tail behavior. We will use the
skewness and kurtosis values of the IMEP time series to describe the misfires that may occur in a spark ignition
engine under lean-burn conditions.

RESULTS AND DISCUSSION

The various IMEP time series analyzed here are depicted in Fig. 1. The statistical parameters of these time
series are listed in Table 1. Consider first Fig. 1(a) which shows the IMEP time series for the gasoline engine
(with no hydrogen added). In this figure, the misfire cycles are identified as those with low IMEP values. The
histogram of this time series is shown in Fig. 2(a), along with the values of skewness = -2.61 and kurtosis = 15.8
from Table 1. Note that the histogram is skewed to the left, i.e., it has a long left tail representing the low IMEP
values. Since a normal or Gaussian distribution has zero skewness and kurtosis equal to 3, clearly, the above
time series cannot be adequately characterized by a normal distribution. The negative skewness is consistent
with the long left tail in the histogram, and the high value of kurtosis indicates that the distribution of this IMEP
time series is more peaked than a normal distribution.We have fitted severalthree-parameter non-Gaussian
distributions to this time series. The distribution with the best fit is found to be the Generalized Logistic (GLO)
distribution(Johnson et al., 1995). A brief description of the GLO distribution is given in the appendix.The GLO
distribution has been used to model the occurrence of extreme events in many applications including analysis of
bioassays(El-Saidi and George, 1990), financial time series analysis (Gettinby et al., 2004),(Tolika and Gettinby,
2009) and flood frequency analysis (Kjeldson and Jones, 2004),(Ali and Davar, 2008).

For the IMEP time series of the gasoline engine shown in Fig. 1(a), the GLO distribution given by Eq. (A1)
provides the best fit with the following parameter values: μ = 391.34, σ = 7.045 and k = -0.278.The goodness of
fit was assessed using the Kolmogorov-Smirnov test statistic (Sheskin, 2007). The probability density function
(pdf) of this distribution is drawn in Fig. 2(a) overlapping the histogram. For comparison, the fitted normal
distribution is also shown in this figure.

Next we consider the IMEP time series of the engine fueled by gasoline-hydrogen-air mixtures. Figs. 1(b, c, d)
depict the IMEP time series for the mixtures containing 2.14%, 5.28%, and 7.74% hydrogen fractions,
respectively. Note that in Figs. 1(c, d), the IMEP values are more or less symmetrically distributed about their
mean. In other words, there are very few misfire cycles present. The histograms of the three time series are
shown in Figs. 2(b, c, d), and their statistical parameters are listed in Table 1. Consider the IMEP time series
shown in Fig. 1(b) which applies to the gasoline-hydrogen blend containing 2.14% hydrogen. The skewness and
kurtosis of thistime series are -1.54 and 11.2, respectively. Clearly, this time series also cannot be modeled by a
normal distribution. In fact, the three-parameter GLO distribution with μ = 379.23, σ = 4.024 and k = -0.129is
found to provide the best fit to this time series. As in Fig. 2(a), the fitted GLO distribution and the normal
distribution areshown in this figure overlapping the histogram. The histogram of the IMEP time series for the
gasoline-hydrogen blend with 5.28% hydrogen is presented in Fig. 2(c). From the shape of this histogram and
the skewness and kurtosis values, it is apparent that the distribution of this time series is close to a normal
distribution. We have fitted a normal distribution and a GLO distribution to this dataset, and drawn them
overlapping the histogram. Finally, consider the gasoline-hydrogen blend with 7.74% hydrogen. Fig. 2(d) shows
the histogram of theIMEP time series for this case. Because the skewness and kurtosis values of this time series
are very close to those of a normal distribution, we have fitted a normal distribution to this time series.The fitted
normal distribution is shown overlapping the histogram in Fig. 2(d). For comparison, the fitted GLO distribution is
also shown in this figure.

908
It is appropriate to note from Table 1 that the COV is highest for the IMEP time series of the gasoline engine
(with no hydrogen added). The COV values continually decrease as the amount of hydrogen added increases,
indicating that the cyclic variations of IMEP are reduced with hydrogen addition. While the cyclic variations may,
in general, be adequately described on the basis of the COV of the IMEP time series, the misfire cycles canbe
more appropriately identified on the basis of their skewness and kurtosis.

CONCLUSIONS
We have examined the lean misfire phenomenon in a gasoline engine. Usingthe coefficient of variation (COV) of
the IMEP time series, it is shown that by adding a small amount of hydrogen to the gasoline-air mixture, the
misfires can be suppressed. We have also demonstrated that the IMEP time series associated with misfires can
be modeled by a non-Gaussian probability distribution such as the generalized logistic (GLO) distribution,
whereas those with almost no misfires can be modeled by a normal (Gaussian) distribution. The large amount of
data collected over 1000 consecutive engine cycles enablesus to clearly observe the many misfire cycles that
can occur in a spark ignition engine under lean-burn conditions.

REFERENCES
Ali, S.and Davar, M. 2008. Comprehensive evaluation of regional flood frequency analysis by L- and LH
moments: Development of LH moments parameters for the generalized Pareto andgeneralized logistic
distributions.Stochastic Environmental Research and Risk Assessment23:137-52

Apostolescu, N. andChiriac, R. 1996.A study of combustion of hydrogen-enriched gasoline in asparkignition

engine.SAE Paper No:960603.

Ceviz, M. A., Çavuşoğlu, B., Kaya, F. and Öner, İ. V., Determination of cycle number for real in-cylinder
pressure cycle analysis in internal combustion engines.Energy 36, 2465-72, 2011.

Ceviz, M. A., Sen, A. K., Kuleri, A. K. and Öner İ. V. 2012.Engine performance, exhaust emissions, and cyclic
variations in a lean-burn gasoline engine fueled by gasoline-hydrogen blends.Applied Thermal Engineering36:
314-24.

D’Andrea, T., Henshaw, P.F. and Ting, D. S. K. 2004.The addition of hydrogen to a gasoline-fueledSI
engine.International Journal of Hydrogen Energy 29:1541–52.

El-Saidi, M. A. and George, E. O. 1990.A generalized logistic model for quantal response bioassay.Biometrical
Journal 32: 943-54.

Fujikawa, T., Nomura, Y., Hatton, Y., Kobayashi, T. and Kanda, M. 2003.Analysis of cycle-by-cyclevariations in a
direct-injection spark-ignition engine using a laser-induced fluorescencetechnique.International Journal of
Engine Research 4:143-53.

Galloni, E. 2009.Analyses about parameters that affect cyclic variation in a spark ignition engine.Applied
Thermal Engineering29: 1131-37.

Gettinby, G. D., Sinclair, C. D., Power, D. M. and Brown, R. A. 2004. An analysis of the extreme share returns in
the UK from 1975 to 2000.Journal of Business, Finance and Accounting 31: 607-46.

Heywood, J. B. 1988. Internal Combustion Engine Fundamentals. New York:McGraw-Hill Inc.

Ishii, K., Sasaki, T., Urata, Y., Yoshida, K. andOhno, T. 1997.Investigation of cyclic variation ofheatreleaseunder
lean burn operation in a spark ignition engine.SAE Paper No:972830.

Ji, C. and Wang, S. 2009.Effect of hydrogen addition on combustion and emissions performance of a spark
ignition gasoline engine at lean condition.International Journal of Hydrogen Energy34:7823-34.

909
Johnson, N. L., Kotz, S. andBalakrishnan, N. 1995.Continuous Univariate Distributions.Volume 2.New
York:Wiley

Kjeldsen, T. R. and Jones, D. A. 2004. Sampling variance of flood quantiles from the generalizedlogistic
distribution estimated using the method of L-moments.Hydrology and Earth Systems Science 8:183-90.

Ozdor, N., Dulger, M. and Sher E. 1994. Cyclic variability in spark ignition engines: a literatureSurvey.SAE Paper
No: 940987.

Scholl, D. and Russ, S. 1999.Air-fuel ratio dependence of random and deterministic cyclic variability in a spark-
ignited engine.SAE Paper No: 1999-01-3513.

Sheskin, D. 2007.Handbook of Parametric and Nonparametric Statistical Procedures,4th edition.Boca Raton,

Florida:CRC Press,.

Tolika, K. andGettinby, G. D. 2009.Modeling the distribution of extreme share returns in Singapore.Journal

ofEmpirical Finance16:252-63.

Verhlest, S., Maesschalck, P., Rombaut, H. and Sierens, R. 2009.Efficiency comparison betweenhydrogen and
gasoline, on a bi-fuel hydrogen/gasoline engine.International Journal of Hydrogen Energy 34: 2504-10.

Wang, S., Ji, C., Zhang, M. and Zhang, B. 2010. Reducing the idle speed of a spark-ignited gasoline engine
with hydrogen addition.International Journal of Hydrogen Energy 35: 10580-88.

Young, M. B. 1981.Cyclic Dispersion in Homogeneous Charge Spark Ignition Engine.SAE Paper No: 810020.

Zervas, E. 2004.Correlations between cycle-to-cycle variations and combustion parameters of a spark ignition
engine.Applied Thermal Engineering 24: 2075-81.

910
APPENDIX

The generalized logistic distribution

The three-parameter generalized logistic (GLO) distribution is a generalization of the two-parameter logistic
distribution. The probability density function (pdf) of the GLO distribution is given by(Johnson et al., 1995):
11 / k
 x
1  k  
f ( x)  2
k 0, (A1)
  x  
1 / k

 1  1  k  
    

 x 
exp   
  
f ( x)  2
k 0. (A2)
  x   
 1  exp   
   

Here the symbols μ, σ and k represent the location parameter, scale parameter and shape parameter,
respectively. A positive value of k corresponds to a positively skewed distribution, i.e., with a longer right tail,
whereas a negative value of k represents a negative skewed distribution with a longer left tail. When k = 0, the
GLO distribution becomes the logistic distribution which is symmetric.

Table 1.Statistical parameters of the IMEP time series for different hydrogen fractionsSD = standard deviation, COV =
coefficient of variation
H2 (%) Mean SD COV Skewness Kurtosis
0.00 387.8 16.59 4.28 -2.61 15.8
2.14 378.4 8.14 2.15 -1.54 11.2
5.28 401.3 4.96 1.25 -0.26 3.56
7.74 396.9 4.84 1.22 -0.02 2.77

911
Fig. 1. Indicated mean effective pressure (IMEP) time series of the gasoline engine with (a) no hydrogen added, (b) 2.14%
hydrogen added, (c) 5.28% hydrogen added, (d) 7.74% hydrogen added.

912
Fig. 2. Histograms of the various IMEP time series shown in Fig. 1, and the probability density functions [f(x)] of the fitted
distributions. Here x represents the IMEP variable.

913
 A NEW APPROACH FOR TURBOMACHINERY MODELLING BY USING ANFIS STRUCTURE

I. Yazar, E. Kiyak and F. Caliskan

Eskisehir Osmangazi University, Department of Mechatronics, Industrial Zone, Eskisehir, TURKEY

iyazar@ogu.edu.tr
Anadolu University, Faculty of Aerospace Sciences, 26470, Eskisehir, TURKEY
ekiyak@anadolu.edu.tr
Istanbul Technical University, Department of Control Engineering, Istanbul, TURKEY
caliskanf@itu.edu.tr

ABSTRACT
Aviation is the one of the continually developing sector in transportation area. People prefer to travel with more
comfortable, more economic, time saver and safety vehicles during their travel. According to this selection, the
number of passengers in civil aviation is increasing day by day. Due to this acceleration, air-lines expand their fleet
and increase the number of their daily flights. Eventually, manufacturing velocity accelerates as well. Today,
aircraft manufacturers focus on trying to design safer, more efficient, more environment-friendly, more economic
aircraft. No doubt, it is not an easy process. Aircraft has a complex structure and its design duration takes a very
long time.
Aircraft is composed of two main parts: structural part and system part. Both of them have a critical design process.
After the design level, manufacturing and test levels are followed respectively. These levels are valid for all parts in
aircraft. It means that every component has a critical mission in aircraft. One of the critical section of the aircraft is the
engine part. In aircraft, different kinds of gas turbine engines are used. A typical example of a gas turbine engine is
formed in five parts. These are respectively inlet, compressor, combustion chamber, turbine, exhaust parts. Basic
operation principle of the gas turbine engine is to convert chemical energy into mechanical energy. In compressor
part, the air that is taken from inlet is compressed and pressurized. In the combustion chamber, the pressurized air is
burnt under the high temperature. Then the gas form is taken to the turbine part and it expands to the ambient
pressure crossing through the turbine. Finally it is exhausted in the exhaust part.
In the design process of gas turbines, all parts have different design issues. Especially, compressor and turbine
design models have critical points that must be considered due to their geometrical structure. When preparing
models for these parts, component maps are used. Especially, for the transient conditions, maps represent the
component performance very well. The maps are difficult to directly use in simulations so, different techniques are
used to read data from maps. In our study, a technique based on the “Adaptive Neuro-Fuzzy Inference System
(ANFIS)” is tested and proposed on MATLAB/Simulink.

Keywords: Compressor Mapping, Adaptive Neuro-Fuzzy Inference System (ANFIS), Gas Turbines, Energy,
Performance Maps

INTRODUCTION
Gas turbines are widely used in aviation sector. Different types of aircraft engines are classified in terms of range,
efficiency, and thrust demands. These demands identify the design cycle of the gas turbine at the same time. Gas
turbines are complex structures. To monitor and control the operation of these complex structures and predict the
abnormal conditions in steady-state and transient operations, reliable mathematical models are needed (Filho et al.
2012). Gas turbine non-linear mathematical models are made up of aerothermodynamics laws, energy-mass
conservation laws, and empirical formulations and so on (Kulikov and Thompson, 2005). These models are prepared
for design process before manufacturing. An accurate mathematical model helps the design of the control system
process, so deriving a realistic mathematical model is very important. Each component of gas turbine can be
modeled by using equations and empirical equations. The critical ones are the compressors and turbines. These
components are composed of multiple individual stages and in modeling, they are assumed to be a stacked single
stage structure. By this way, components dynamic behavior can be analyzed with only inlet and final outlet conditions
(Yarlagadda, 2010). For this examination, compressor and turbine maps are used. These maps are formed from
actual rig test of the engines and then the values are plotted on the map (Yarlagadda, 2010). An example of
compressor map can be seen in Figure 1.

914
Fig.1. Compressor Map with one Beta Line (Sexton, 2001)

Compressor maps can be determined in several ways in terms of the accuracy, complexity and processing time
(Britton, 2008):
 Streamline Curvature Methods
 Map Scaling Techniques
 Stage Stacking Methods
 Row-by-Row Analysis
 Two-Zone Models
 One-Dimensional Loss Models

Map scaling is one of the most used technique in this area. Map Scaling techniques use different digitizing methods
for providing data to the simulations. The following methods are the commonly used examples for the digitizing
process (Polat, 2009):
 Beta Line Method
 Curve Fitting Techniques
 Neural Network Techniques

In the literature, many studies have been done for making the map data comprehensible for computer simulations
(Ghorbanian and Gholamrezaei, 2008). Beta Line Method is used by Kurzke in describing data from maps (Kurzke,
1996). Curve fitting techniques try to form an equation that shows the characteristic outputs of the compressors and
turbines. (Sieros et al. 1997), (Moraal and Kolmanovsky, 1999), (Orkisz and Stawarz, 2000) and (Ailer et al. 2001)
uses functions to represent the characteristics of the components. (Kong et al. 2006) and (Kong and Ki 2007) use
genetic algorithm method to generate digital data from maps. In the last years neural network is a popular concept for
preparing digitalized data from maps. (Bao et al. 2006), (Yu et al. 2007) and (Ghorbanian and Gholamrezaei 2006,
2006, 2007, 2009) are based their studies on studying and comparing different types of Artificial Neural Network
structures for digitizing map data.

In this study, the compressor and turbine components characteristic performance output parameters are calculated
via Adapted Neuro-Fuzzy Inference System (ANFIS). By using test data, corrected mass flow rate and isentropic
efficiency values of the components are estimated for a given input data (RPM, Pressure ratio) between idle to
maximum thrust levels. As a result of training calculations, accurate output data is found at the end of the simulations
in a short time. Furthermore, the transient conditions are simulated as well and the results are satisfactory. Fuzzy
connected neural network structure is suitable for modeling output performance parameters of these components.

COMPRESSOR and TURBINE PERFORMANCE MAPS

Fluid machines that transfer energy between a rotor and a fluid are called turbo machines. Compressors and turbines
are categorized in this type of machines (Turie, 2011). In the literature there are three types of compressors that are
used in gas turbines (FAA, 2009). These are respectively centrifugal flow, axial flow, centrifugal-axial flow
compressors. In centrifugal flow compressor, compression is formed by accelerating the air outward perpendicular to
the longitudinal axis of the compressor. In axial flow compressor, compression is formed by a series of rotating and
stationary airfoils moving the air parallel to the longitudinal axis. The last type, centrifugal-axial flow compressor, uses
both compression methods for achieving the desired compression (FAA, 2009). In a gas turbine, compressor and

915
turbine parts work correspondingly. In turbine part, the energy that is extracted from the expansion of the fluid is
transferred into compressor part mechanically. In modeling of turbo machines, forming the accurate gas turbine
model is important. Especially in compressor and turbine parts, it is possible to find mutual dimensionless parameters
characterizing the flow (Fözö et al., 2010). Maps are used for finding these characteristic performance parameters.
Maps are designed among corrected mass flow parameter, pressure ratio, and corrected speed lines. An example
can be seen in Fig.1. Characteristics are defined in corrected parameters for eliminating the dependence of the
performance characteristics from temperature and pressure. Corrected mass flow rate ( mcorr ) and isentropic
efficiency (  ) can be determined as a function of pressure ratio (  ) and corrected rotational speed in equation 2
and 3 (Yarlagadda, 2010). To define corrected rotational speed equation 1 is used. It is the proportion of actual data
over design data. Final value of mass flow rate is defined as in equation 4.
N
( )
N corr   (1)
N
( ) design


mcorr  f1 ( , Ncorr ) (2)

  f 2 ( , Ncorr ) (3)


m  m corr ( ) (4)

As mentioned earlier, there are many ways to calculate f1 and f2 functions in the literature. f1 and f 2 are
functions that digitize the map data into suitable simulation form. Instead of using 2D Look Up table, curve fitting, beta
lines or individual neural network algorithms, in our work, Adapted Neuro-Fuzzy Inference System is applied to our
simulation model. Map data for 2 types of compressors and 2 types of turbines are used in four different simulation
models. Initially, with real test data, the system is trained and then by using the trained system, intermediate values
and real values are tested on the system.

ADAPTIVE NEURO-FUZZY INFERENCE SYSTEM (ANFIS)

System modeling techniques based on mathematical formulations are sometimes not suitable for nonlinear systems.
In this type of systems, input-output mapping method usage can be convenient (Jang, 1993). This work is an
application example of input-output mapping method in MATLAB/ANFIS structure. ANFIS (Adaptive Neuro-Fuzzy
Inference System) is a combination of Neural Network training structure and Fuzzy Logic clustering structure. It can
be directly used on command window of MATLAB as ANFIS or GUI structure on MATLAB/Simulink.

ANFIS GUI structure consists of two parts: generating FIS and training FIS. In generating FIS structure, MATLAB
presents two types of methods: Grid Partitioning and Sub Clustering. Furthermore, a preformed FIS structure can be
used directly by loading from a file or workspace. In this study, Grid Partitioning method is adopted for generating FIS
structure. Grid Partitioning is a method that generates a single-output Sugeno-type FIS structure by grid partitioning
on the data (Mathworks, 2013).

Training FIS is the important part of the ANFIS structure. There exist two types of training methods in ANFIS:
backpropagation training and backpropagation-least squares combination Hybrid training methods. In our work hybrid
training method is used for training the FIS structure.

The ANFIS architecture used in our study as follows (Jang, 1993):

For two inputs x, y and one output z under two fuzzy if-then rules of Takagi and Sugeno’s type (Takagi and Sugeno,
1983):
Rule 1: If x is A1 and y is B1 , then f1  p1 x  q1 y  r1
Rule 2: If x is A2 and y is B2 , then f 2  p 2 x  q2 y  r2

Layer 1: Every layer in this node has a membership function as the following. This function is the output function of
the node as well.

Oi1   Ai ( x)

916
Fig.2. An example of ANFIS Model Structure

where x is the input, i is the node number, and Ai is the linguistic variable of this node function. The node function in
our work is selected as a Gaussian function. Constants in Gaussian function are labeled as premise parameters in
ANFIS.

Layer 2: In this layer, incoming signals (membership functions) are multiplied. Each node output represents the firing
strength of a rule.

wi   Ai ( x)   Bi ( y), i  1,2.

Then normalized firing strengths are calculated by taking the ratio of related node firing strength to the sum of all
rule’s firing strengths:

wi
wi  , i  1,2.
w1  w2

Layer 3: Output function of this layer is described as the following. The parameters pi , qi , ri  are called
consequent parameters in ANFIS.

Oi3  wi f i  wi ( pi x  qi y  ri )

Layer 4: The last layer is the overall output of the system. It is calculated by summing of all signals from the previous
layer.

w f i i
Overall output= O   wi f i
4
 i
=
w
i
i i
i

917
HYBRID LEARNING ALGORITHM
The algorithm of the Hybrid Learning is composed of two different procedures. These are Forward Pass and
Backward Pass respectively (Jang, 1993):

Table 1. Hybrid Learning Algorithm Table (Jang, 1993)

Forward Pass Backward Pass
Premise Parameters Fixed Gradient Descent
Consequent Parameters Least Squares Estimate Fixed
Signals Node Outputs Error Rates

In the forward pass, input and related functional signals go forward to layer 3 and least squares estimation method
tries to estimate consequent parameters in this period. In the backward pass, the premise parameters are updated by
using gradient descent method corresponding to the sum of all nodes error rates from output to the input (Jang,
1993). ANFIS tries to accurate the described error value in each epoch, so the selected epoch number is important
as well.

SIMULATIONS, RESULTS AND DISCUSSIONS

In the simulations, different compressor and different turbine models that are created on MATLAB/Simulink
are used. In training section, two different types of compressor and two different types of turbine data are
used. The same training data, other intermediate values and transient conditions are checked on the model
for the stability. For each engine component, different membership function test error values searched via
performance parameters. According to the error values, a stable membership function number value is fixed.
For generating Fuzzy Inference System (FIS), Grid Partition Method is used due to the smallest number of
inputs. (Neshat et al. 2011),(Wei at al. 2007). In the FIS training section, instead of only back propagation
gradient descent method, a combination of back propagation gradient descent method and least squares
estimation methods, Hybrid Optimization Method is preferred. Hybrid Optimization Method has faster and
nearer results than the individual back propagation gradient descent optimization method (Jang,
1993),(Jang,1991). Simulink model view for compressor type 1 and an example of membership function view
are shown in Fig. 3. and Fig.4. respectively.

Fig.3. Simulink Model View

Fig.4. An Example of Membership Function View of PR

The average error values, epoch numbers and number of membership function values can be seen in the
following tables and ANFIS figures. In figure 5, for compressor 1, corrected mass flow rate (mf=4) and
isentropic efficiency (mf=5) training solutions are seen. In table 2, the best value of average test error for
corrected mass flow is found with 4 membership functions and in table 3, the best value of average test error

918
for isentropic efficiency is found with 5 membership functions. After these membership function values,
unstability of average test error starts in the increasing direction of membership functions and so does the
simulation time.

Fig.5. Corrected Mass Flow (MF=4) and Isentropic Efficiency (MF=5) Training Data Diagrams for Compressor 1

Table 2 Error- MF- Epoch Number Values of Corrected Mass Flow for Compressor 1
COMPRESSOR 1-CORRECTED MASS FLOW
Epoch Number Average Test Error Number of Membership Function (MF)
5 0.0027257 2
5 0.0016331 3
5 0.0010621 4
5 0.0013662 5
5 0.0085135 6
5 0.0019336 7
5 0.021626 8
5 0.093629 9
5 0.013356 10

Table 3. Error- MF- Epoch Number Values of Isentropic Efficiency for Compressor 1
COMPRESSOR 1-ISENTROPIC EFFICIENCY
Epoch Number Average Test Error Number of Membership Function (MF)
5 0.016904 2
5 0.0023263 3
5 0.0025651 4
5 0.00076046 5
5 0.35618 6
5 0.0012207 7
5 0.045845 8
5 0.0078918 9
5 0.016111 10

In figure 6, for compressor 2, corrected mass flow rate (mf=4) and isentropic efficiency (mf=5) training
solutions are seen. In table 4, the best value of average test error for corrected mass flow is found with 4
membership functions and in table 5, the best value of average test error for isentropic efficiency is found
with 5 membership functions. After these membership function values, unstability of average test error starts
in the increasing direction of membership functions and so does the simulation time.

919
Fig.6. Corrected Mass Flow (MF=4) and Isentropic Efficiency (MF=5) Training Data Diagrams for Compressor 2

Table 4. Error- MF- Epoch Number Values of Corrected Mass Flow for Compressor 2
COMPRESSOR 2-CORRECTED MASS FLOW
Epoch Number Average Test Error Number of Membership Function (MF)
5 0.0031701 2
5 0.0018702 3
5 0.0014095 4
5 0.0030851 5
5 0.055742 6
5 0.0051453 7
5 0.077074 8
5 0.0080933 9
5 0.0095825 10

Table 5. Error- MF- Epoch Number Values of Isentropic Efficiency for Compressor 2
COMPRESSOR 2-ISENTROPIC EFFICIENCY
Epoch Number Average Test Error Number of Membership Function (MF)
5 0.028147 2
5 0.0036708 3
5 0.0039948 4
5 0 0010388 5
5 0.056028 6
5 0.006311 7
5 0.099895 8
5 0.018059 9
5 0.0055647 10

In figure 7, for turbine 1, corrected mass flow rate (mf=11) and isentropic efficiency (mf=11) training solutions
are seen. In table 6, the best value of average test error for corrected mass flow is found with 11
membership functions and in table 7, the best value of average test error for isentropic efficiency is found
with 11 membership functions. After these membership function values, unstability of average test error
starts in the increasing direction of membership functions and so does the simulation time.

920
Fig.7. Corrected Mass Flow (MF=11) and Isentropic Efficiency (MF=11) Training Data Diagrams for Turbine 1

Table 6. Error- MF- Epoch Number Values of Corrected Mass Flow for Turbine 1
TURBINE 1-CORRECTED MASS FLOW
Epoch Number Average Test Error Number of Membership Function(MF)
5 0.004068 2
5 0.0015129 3
5 0.00037277 4
5 0.00025415 5
5 0.00035519 6
5 0.000060146 7
5 0.000015917 8
5 4.2807E-07 9
5 4.2586E-07 10
5 4.2266E-07 11
5 4.5725E-07 12
5 5.1955E-07 13
5 5.217E-07 14

Table 7. Error- MF- Epoch Number Values of Isentropic Efficiency for Turbine 1
TURBINE 1-ISENTROPIC EFFICIENCY
Epoch Number Average Test Error Number of Membership Function (MF)
5 0.0058168 2
5 0.0027477 3
5 0.0018257 4
5 0.0015368 5
5 0.0013878 6
5 0.00040202 7
5 0.00017977 8
5 8.8838E-07 9
5 7.9984E-07 10
5 7.9876E-07 11
5 8.5982E-07 12
5 9.6349E-07 13
5 9.6971E-07 14

921
In figure 8, for turbine 2, corrected mass flow rate (mf=11) and isentropic efficiency (mf=12) training solutions
are seen. In table 8, the best value of average test error for corrected mass flow is found with 11
membership functions and in table 9, the best value of average test error for isentropic efficiency is found
with 12 membership functions. After these membership function values, unstability of average test error
starts in the increasing direction of membership functions and so does the simulation time.

Fig.8.Corrected Mass Flow (MF=11) and Isentropic Efficiency (MF=12) Training Diagrams Data for Turbine 2

Table 8. Error- MF- Epoch Number Values of Corrected Mass Flow for Turbine 2
TURBINE 2-CORRECTED MASS FLOW
Epoch Number Average Test Error Number of Membership Function (MF)
5 0.0055707 2
5 0.0022691 3
5 0.0016757 4
5 0.00082476 5
5 0.00070482 6
5 0.000399 7
5 0.000041398 8
5 0.000082751 9
5 4.618E-07 10
5 3.8769E-07 11
5 4.7972E-07 12
5 4.9337E-07 13
5 4.891E-07 14

922
Table 9. Error- MF- Epoch Number Values of Isentropic Efficiency for Turbine 2
TURBINE 2-ISENTROPIC EFFICIENCY
Epoch Number Average Test Error Number of Membership Function
5 0.0074187 2
5 0.0040202 3
5 0.0028297 4
5 0.0021117 5
5 0.0016568 6
5 0.00094596 7
5 0.00015175 8
5 0.00016766 9
5 4.1732E-06 10
5 1.2372E-06 11
5 8.691E-07 12
5 9.8896E-07 13
5 1.0333E-06 14

CONCLUSION
In this work, an approach for modeling performance parameters of gas turbine engine compressor and turbine is
proposed. The proposed approach is an effective and easy-use way for modeling the performance parameters of
compressors and turbines. The method is tested on two different compressors and two different turbines and trained
Adaptive Neuro-Fuzzy structure estimates good output values for both on-design and off-design conditions. By
altering the number of membership functions, different error values are formed and it shows that the models can be
run accurately with a few number of membership functions. Additionally, this decreases the simulation time. If the
selected number of membership function is over the best membership function value, unstability increases in average
test error and so does the simulation time. By comparing with the previous studies declarations, this approach has
advantages about both accuracy and simulation time. The individual usage of neural network studies sometimes fail
about error value although it’s short simulation time. Curve fitting and beta line methods don’t work accurately. In
addition, as a difference from other studies it is applied to the turbine component as well. The proposed compact
fuzzy-neural network method is functional for all. In the upcoming works, it is planned to study about engine start up
modeling. It is very hard to model the startup operation in an engine. In the next step, this operation regime is tried to
adapt this examination.

Acknowledgement
We would like to thank to the Tusas Engine Industries for their valuable contributions to our study.

NOMENCLATURE
m corr corrected mass flow rate, dimensionless
 pressure ratio, dimensionless
N corr revolution, dimensionless
 isentropic efficiency

m mass flow rate

 compressor input pressure under sea level condition, dimensionless
 compressor input temperature under sea level condition, dimensionless

923
REFERENCES

Ailer P., Santa I., Szederkenyi G. and Hangos K. M. 2001. Non-linear model-building of a low-power gas turbine.
Periodica Polytechnica Ser. Transportation Engineering 29: 117-135.

Bao C, Ouyang M and Yi B. 2006. Modeling and optimization of the air system inpolymer exchange membrane
fuel cell system. Journal of Power Sources 156(2): 232–243.
Britton, I. A. W. 2008. Development of a performance estimation tool for gas turbine engine centrifugal
compressors. Master of Applied Science Thesis in Mechanical Engineering, Ottawa - Carleton Institute for
Mechanical and Aerospace Engineering.

Federal Aviation Administration (FAA), 2009. Pilot’s handbook of Aeronautical Knowledge, Skyhorse Publishing.

Filho F. A. C. B., Goes L. C. S., Boaventura A., Bosa W. and Fernandes G., 2012. Dynamic modelling nonlinear
and control system for a turboshaft, 12th Pan-American Congress of Applied Mechanics, Port of Spain, Trinidad.

Fözö L., Andoga R., Madarasz L. 2010. Mathematical Model of a Small Turbojet Engine MPM-20. Computational
Intelligence and Informatics: Principles and Practice 313: 313-322.

Ghorbanian K and Gholamrezaei M. 2006. Neural network modeling of axial flow compressor off-design
performance.10th Fluid Dynamic Conference Yazd, Iran.

Ghorbanian K and Gholamrezaei M. 2006. Neural network modeling of axial flow compressor performance map.
45th AIAA Aerospace Science Meeting and Exhibit Reno, USA.

Ghorbanian K and Gholamrezaei M. 2007. Axial compressor performance map prediction using artificial neural
network. ASME Turbo Expo, GT2007-27165, Montreal, Canada.

Ghorbanian K and Gholamrezaei M. 2009. An artificial neural network approach to compressor performance
prediction. Journal of Applied Energy 86: 1210-1221.

Jang J. R., 1991.Fuzzy Modeling using generalized Neural Networks and Kalman Filter algorithm, AAAI-91
Proceedings, 762-767.

Jang J. R., 1993. ANFIS: Adaptive-Network-Based Fuzzy Inference System, IEEE Transactions on Systems, Man,
And Cybernetics 23(3): 665-685.

Kong CD, Kho S and Ki JY. 2006. Component map generation of a gas turbine using genetic algorithms. Journal
of Engineering for Gas Turbines and Power 128(1): 92–96.

Kong CD and Ki JY. 2007. Components map generation of gas turbine engine using genetic algorithms and
engine performance deck data. Journal of Engineering Gas Turbines and Power 129(2): 312–317.

Kulikov G. G. and Thompson H. A. 2005. Dynamic Modelling of Gas Turbine: Identification, Simulation, Condition
Monitoring and Optimal Control. USA: Springer.

Kurzke, J. 1996. How to get component maps for an aircraft gas-turbine’s performance calculations. ASME paper
96-GT-164.

Moraal P. And Kolmanovsky I. 1999. Turbocharger modeling for automotive control application. SAE Technical
Paper Series 108: 1324–1338.

Neshat M., Adeli A.,Masoumi A., Sargolzae M. 2011. A Comparative Study on ANFIS and Fuzzy Expert System
Models for Concrete Mix Design. IJCSI International Journal of Computer Science Issues 8(3): 196-210.

Orkisz M. and Stawarz S. 2000. Modeling of turbine engine axial-flow compressor and turbine characteristics.
Journal of Propulsion and Power 16(2): 336–339.

924
Polat, C. 2009. An electronic control unit design for a miniature jet engine. Master of Science Thesis in Mechanical
Engineering Graduate School of Natural and Applied Sciences of Middle East Technical University.

Sexton W. R. 2001. A method to control turbofan engine starting by varying compressor surge valve bleed. Master
of Science Thesis in Mechanical Engineering Graduate Faculty of Virginia Polytechnic Institute and State
University.

Sieros G, Stamatis A and Mathioudakis K. 1997. Jet engine component maps for performance modeling and
diagnosis. Journal of Propulsion and Power 13 (5): 665–674.

Takagi T. and Sugeno M. 1983. Derivation of fuzzy control rules from human operator’s control actions.
Proceedings of IFAC Symposium on Fuzzy Information, Knowledge Representation and Decision Analysis 55-60.
http://www.mathworks.com/help/fuzzy/anfis-and-the-anfis-editor-gui.html#bq97_i_

Turie S. E., 2011. Gas Turbine Plant Modeling for Dynamic Simulation. Master of Science Thesis in Industrial
Engineering and Management KTH School of Industrial Engineering and Management.

Yarlagadda S. 2010. Performance analysis of j85 turbojet engine matching thrust with reduced inlet pressure to
the compressor. Master of Science Thesis in Mechanical Engineering Graduate Faculty of University of Toledo.

Yu Y, Chen L, Sun F and Wu C. 2007. Neural network based analysis and prediction of a compressor’s
characteristic performance map. Journal of Applied Energy 84(1):48–55.

Wei M., Bai B., Sung A. H. , Liu Q., Wang J., Cather M. E., 2007. Predicting injection profiles using ANFIS,
Elsevier, Information Sciences Journal, 4445-4461.

925
 OVERVIEW OF IONIC LIQUIDS USED AS WORKING FLUIDS IN ABSORPTION
CYCLES

M. khamooshi, K. Parham, U. Atikol, P. Agboola

Eastern Mediterranean University

Mechanical Engineering Department Faculty of Engineering Eastern Mediterranean University Gazimagusa
TRNC (Via Mersin 10) Turkey
email: mehrdadkhamooshi@yahoo.com , kiyan.parham@emu.edu.tr, ugur.atikol@emu.edu.tr,
phillips.agboola@cc.emu.edu.tr.

Abstract: The cycle performance of refrigeration cycles not only depends on their configuration,
but also on thermodynamic properties of working pairs regularly composed of refrigerant and
absorbent. The commonly used working pairs in absorption cycles are aqueous solutions of either
lithium bromide–water or ammonia–water. However, corrosion, crystallization, high working
pressure and toxicity are their major disadvantages in industrial applications. Therefore, seeking
more advantageous working pairs with good thermal stability, minimum corrosion and without
crystallization has become the research focus in the past two decades.
Ionic liquids (ILs) are room-temperature melting salts that can remain in the liquid state at near or
below room temperature. ILs have attracted considerable attentions due to their unique
properties, such as negligible vapor pressure, non-flammability, thermal stability, good solubility,
low melting points and staying in the liquid state over a wide temperature range from room
temperature to about 300°C.
The above-mentioned highly favorable properties of ILs motivated us for carrying out the present
research and reviewing the available ILS found in the literature as the working fluids of absorption
cycles. Absorption cycles contain absorption heat pumps, absorption chillers and absorption
transformers.

Keywords: Absorption cycle, Coefficient of performance, Corrosion, Crystallization, Refrigerant,

Ionic Liquid.

1. Introduction
With fast economic growth and constantly increasing energy consumption, the human kind is
about to face with a growing degradation of the environment if business continues as usual. For
this reason, the utilization of low-grade energy has become one of the most attractive solutions to
heating and cooling problems encountered in industrial and residential applications. It is possible
to recover the low-grade heat wasted in many industries to use it in some processes in order to
increase the energy efficiency.

Absorption cycles which include absorption heat pump (AHP), absorption chiller (AC) and
absorption heat transformer(AHT) can use waste heat economically resulting in decreasing the
consumption of primary energy and reducing the negative impact on the environment. The basic
principle of an absorption cycle is shown in Fig.1. Since the invention of absorption cycles, the
properties of working fluids have been a challenging issue, as performance of an absorption cycle
critically depends on thermodynamic properties of working pairs composed of refrigerant and
absorbent. Thus, searching for more beneficial working pairs with excellent thermal stability, no
corrosion and no crystallization has become the subject of research in recent years (1).The basic
components of the absorption cycles are the evaporator, the condenser, the generator, and the
absorber, while the evaporator and generator are supplied waste heat at the same temperature
and the increased heat is delivered from absorber. The operating system of the basic absorption
cycles is explained as follows:
Refrigerant vapor is produced at state 4 in the evaporator, by low/medium grade heat source. The
refrigerant vapor dissolves and reacts with the strong refrigerant-absorbent solution that enters
the absorber from states 10 and weak solution return back to generator at state 5. In the

926
generator some refrigerant vapor is removed from the weak solution to be sent to the condenser
and consequently the strong solution from the generator is returned to the absorber. After
condensing the vaporized refrigerant in the condenser it is pumped to a higher-pressure level as
it enters the evaporator. The waste heat delivered to the evaporator causes its vaporization.
Again the absorber absorbs the refrigerant vapor at a higher temperature. Therefore, the
absorption cycles has the capability of raising the temperature of the solution above the
temperature of the waste heat(2).

The requirements of working fluids of absorption cycles areas follows:(3)

1. The difference in boiling point between the pure refrigerant and the mixture at the same
pressure should be as large as possible,

2. Refrigerant should have high heat of vaporization and high concentration within the absorbent
in order to maintain low circulation rate between the generator and the absorber per unit of
cooling capacity,

3. Transport properties that influence heat and mass transfer, e.g., viscosity, thermal conductivity,
and diffusion coefficient should be favorable,

4. Both refrigerant and absorbent should be non-corrosive, environmental friendly, and low-cost

Fig.1. The basic absorption cycle

Ionic liquids (IL) are a class of low-temperature molten salts, which are constituted by an organic
cation and an inorganic anion. During recent years, ILs have been used as organic green solvent
in catalysis, separation process, electrochemistry and many other industries for their unique
physical and chemical properties, such as negligible vapor pressure (4), negligible flammability

927
and thermal stability(5), low melting temperature and liquid state over a wide temperature range
and good solubility(6). In particular, the low volatility of the ILs enables easy separation of the
volatile working fluid from the ILs by thermal stratification with the minimum harmful impacts on
environment (7). ILs can be a substitute for some of the most used toxic working fluids (such as
ammonia/water) in absorption cycles. Since many of the Ionic Liquids have melting points below
the lowest solution temperature in the absorption system (~300)(8, 9), they also wipe out the
crystallization and metal-compatibility problems of water/LiBr system.

Although a large number of ILs as the working fluids in absorption cycles, are given in the
literature, there is still need of a complete review with comparison of all ILs for achieving higher
performance of the absorption cycles. It seems to be necessary to have a review in this class of
working pairs. So in the present worked a review of specific properties of Ionic Liquids as working
fluids for absorption cycles has been carried out.

2.A review of ILs as working fluids

The thermodynamic performance of an absorption chiller operating with water + 1-ethyl-3-

methylimidazoliumdimethylphosphate[EMIM][DMP] and H2O+LiBr mixtures has been compared
under the same operating conditions condition by Zhang and Hu (10). The coefficient of
performance of the absorption chiller for working fluid H2O+[EMIM][DMP] were lower than that of
H2O+LiBr by 7%, but still higher than 0.7.Also for the same condensation and absorption
temperatures the generation temperature of refrigeration cycle for H2O+[EMIM][DMP] is
somewhat lower than that for H2O+LiBr, allowing the operation of the absorption chiller at lower
temperatures of heat supply. These features indicate that the binary ionic liquid solution,
H2O+[EMIM][DMP] has potential to be a new working pair for the absorption chiller driven by low-
grade waste heat or hot water generated by common solar thermal collectors. The performance
of absorption refrigeration cycle with methanol and 1,3-methylimidazolium
dimethylphosphate(CH3OH+[MMIM][DMP])is investigated and a COP of 0.82 was found with
condensing, absorption and evaporation temperatures of 40°C, 30°C and 5°C respectively. The
requirements of operating pressure, condensing pressure, and vacuum for the system using
CH3OH+ [MMIM][DMP] are lower, which makes the operation and the maintaining of the system
more trouble-free(11, 12).Performance benchmarks and system-level simulations for
refrigerant/[BMIM][PF6] pairs, are carried out by Kim et al. (13). A non-random two liquid model is
built and used to predict the solubility of the mixture as well as the mixtures properties such as
enthalpy and entropy. The evaporator and condenser saturation temperatures are 25°C and
50°C, respectively, with the operating temperature set at 85°C. As R32, R134a and R152a are
compared with each other it is observed that R32 produces the highest performance with a
maximum COP of 0.55 while R134a and R152a returned similar COPs of 0.4 at a generator outlet
temperature of 80°C.

Kim et al. (14)in one of their other studies have obtained the thermo physical properties of H2O
+[EMIM][BF4] with saturation temperature at the evaporator and condenser being 25°C and
50°C,respectively. A power dissipation of 100 W is estimated while COP value of the system
reaches 0.91. The suitable compatibility of water with [EMIM][BF4] and the superior properties of
water as a heat transfer fluid, such as large latent heat of evaporation, followed by extremely
small refrigerant (water) flow rate, resulted in its high performance. Higher viscosity ILs causes an
increased pressure drop in the compression loop, which would result in larger pumping power or
larger pipes and system volume. The viscosity increases with cation mass: EMIM< BMIM< HMIM.
The viscosity is more dependent on the anion with the following order: Tf2N< BF4< PF6. The
viscosity of [EMIM][Tf2N] is only 31.3 MPa s at 294 K, which is 10 times smaller than that of
[HMIM][PF6] (14).The COP of the [MMIM]DMP/methanol absorption refrigeration is lower than
LiBr+H2O in absorption refrigeration in the same temperature conditions, while higher than
H2O+NH3 in absorption refrigeration under most temperature conditions. [MMIM] DMP+methanol
has excellent potential to be applied as the working pair of absorption refrigeration (15).

928
Ionic liquids with longer cation alkyl chain length cause a larger solubility but lower dependence of
the solubility on temperature (16, 17). Thus, Ionic liquids with shorter alkyl chains in cation are
preferred ([EMIM]> [BMIM]> [HMIM]) due to more sensitive dependence of the solubility on
temperature. R143a+[BMIM][PF6] and R134a+[HMIM][PF6] are remarkably less sensitive to
temperature and show relatively low COPs (14).The working pairH2O+EMISE was considered as
a potential working pair (18, 19). The vapor pressure, heat capacity and density of the
H2O+EMISE system were measured and correlated which verified the availability of being used in
absorption cycles. Dong et al.(20)recommendH2O+([DMIM]DMP) as an alternative to H2O+LiBr
by comparing the coefficient of performance (COP). To predict the cycle performance, a single-
effect absorption refrigeration cycle was simulated based on the models obtained from the
studies of vapor pressure and heat capacity of the H2O+([DMIM] DMP) system. The simulation
results show that the cycle performance of H2O+([DMIM] DMP) is close to that of conventional
working pair H2O+LiBr. However, for the cycle using alternative working pair of the H2O+ ([DMIM]
DMP) system, the operating temperature range has been extended and the disadvantages of
crystallization and corrosion caused by H2O+LiBr can be relieved. Ionic Liquids have a large
capacity of dissolving CO2, so that the [BMIM][PF6]can be used as a refrigerant in an absorption
cycle. The COP of a system utilizing CO2/bmimPF6 is much lower than that of a traditional
system using the NH3+H2O pair. However, it should be noted that serious questions remain
regarding the accuracy of the ideal-gas heat capacity coefficients for the CO2+bmimPF6 mixture
(21).

The permanent ion–dipole interaction between the ionic liquid and TFE evokes a considerable
negative deviation from Raoult’s law, which is normal in absorption pairs (22). The Br anion
shows a stronger interface with TFE than the BF4 anion. This means that the required
temperature for regeneration will be smaller, the vapor pressure lines will be steeper (23). Novel
working fluids for an absorption heat pump have been proposed by investigating ionic liquids and
fluoroalcohol. In this study [BMIM][Br]+TFE and [BMIM][BF4]+TFE have been considered as
potential working pairs for the first time. The [BMIM][Br]+TFE system was found to be more
favorable than the [BMIM][BF4] + TFE from the results of vapor pressure. The excess volume,
apparent molar volume, partial molar volume and apparent molar expansibility of [MMIM] Cl
aqueous solutions were investigated. Based on the results it was found out that the water content
has significant impact on the volumetric properties, and the temperature dependence of the
density is significantly less than the water content. It was proposed here that [MMIM] Cl has the
potential to be used as a novel absorbent species of an absorption cycle working fluid (24).

High circulation ratio increases the energy requirements of heating and pumping processes.
Circulation ratios in H2O+LiBr and NH3+H2O systems usually are smaller, with typical values
around f = 10 (25, 26).ILs circulation ratios compared to conventional absorption refrigerators are
high. The lowest and therefore the best circulation ratios observed ranged from 20 to 25 (27). The
estimated energetic efficiency of the cycle with ionic liquids as absorbents with COP of 0.21 is
lower than conventional pairs (NH3+H2O and LiBr+H2O) in absorption refrigerators. However,
CO2+[BMPYRR][Tf2 N] with COP of 0.55 is compared with NH3+H2O pairs operating in
equivalent conditions. This is due to the necessity of operating the cycle with a relatively high
circulation ratio (24 in the case of CO2+[BMPYRR/Tf2N] pairs compared to 10 in conventional
NH3+H2O systems), which increases the energy necessities of heating and pumping processes.
The content of the ILs has a direct relation with the deviation from the Raoult’s law; means that
these binary solutions have a negative deviation from Raoult’s law. Solutions such as water,
ethanol and methanol containing [EMIM][DMP] have strong absorbing ability for coolant, which is
a very important property of a working pair for absorption heat pump or absorption refrigeration
(1).

The processes of [EMIM][DMP] mixed with water, alcohol and methanol are exothermic under the
temperature of the 298.15 K and pressure of 1atm. They are ranked in the order of water,
methanol and alcohol according to the magnitude of mixing heat. Because of electron negativity

929
of phosphorus contained in the [EMIM][DMP], hydrogen bonds are formed between [EMIM][DMP]
and water, alcohol and methanol, respectively. This property is especially important for the
working pair of the absorption heat pump or refrigeration. They are ranked in the order of water,
methanol and alcohol according to the magnitude of mixing heat (1).The three binary solutions,
[MMIM][DMP] + water/ethanol/ methanol systems have exhibited some important characteristics
which are needed for an absorption heat pump working pair. Their basic thermodynamic
properties including vapor pressure, heat capacity, excess enthalpy, viscosity and density were
measured(18).The viscosity of pure [MMIM][DMP] was very high, but it would decrease sharply
when either [MMIM][DMP] was heated or the water/ethanol/methanol was added. As the
absorption and generation temperatures of an absorption heat pump or absorption heat
transformer are usually high, the viscosity of the [MMIM][DMP]+water/ethanol/methanol pairs is
expected to be low. Consequently, the high viscosity of the ionic liquid would not limit its use as a
heat pump absorbent. Zhao et al.(28)investigated the Vapor pressure data for nine binary
systems at varying temperature and IL-content were measured using a quasi-static method. The
effect of ILs on the vapor pressure lowering of solvent follows the order [MMIM][DMP]>
[EMIM][DEP]> [BMIM][DBP] for water, and [BMIM][DBP]>[EMIM][DEP]>[MMIM][DMP] for organic
solvent methanol and ethanol. This suggests that the “ionic” is specifically dominant in water and
“molecule” is prevailing in organic solvents.

Research on the solubility of ammonia in four ILs which contain; [EMIM][Ac], [EMIM][SCN],
[EMIM]- [EtOSO3] and [DMEA][Ac] has been studied by Yokozeki and Shiflett(29). The very high
solubility behavior has been clearly demonstrated in terms of the thermodynamic excess
functions based on the present EOS. Discussion shows that there is opportunity for the
absorption cycle application using ammonia-IL systems replacing the traditional ammonia–water
system. Also, the flow ratio and COP of series of ammonia-IL and NH3+water has been reported
with the same cycle condition as temperatures of (generator)/(condenser)/(absorber)/(evaporator)
= 100/40/30/10 °C, these properties which are shown in Fig 2,3 are based on results taken from
Yokozeki and Shiflett studies (29-31).

930
 COP
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0

Fig 2.COP of series of NH3+IL and NH3+water(29-31)

Flow ratio
30

25

20

15

10

Fig 3. Flow ratio of series of NH3+IL and NH3+water(29-31)

Among ammonia–ILs pairs studied, [DMEA][Ac] has shown the best result. The performance
(COP) of the present ammonia–ILs systems is somewhat lower than that of the ammonia–water
system. The functional capability of replacing LiBr with ILs as an absorbent for H2O in absorption
heat pumps and chillers has been demonstrated using theoretical absorption cycle analysis in
some experimental works (10, 32-35). Yokozeki and Shiflett(32) used twelve ILs with H2O in a

931
simple cycle configuration analysis, using coefficient of performance and flow ratio as comparing
parameters, which explains the efficiency and compactness of the system. In the analysis
[MMIM][(CH3)2PO4] and [EMIM][(CH3)2PO4] indicate the best results, with the highest COP and
lowest flow rate values. The COP result was about 85-88% that oftheH2O+LiBr system. The
results of cycle simulations demonstrate that IL+H2O systems could be competitive with H2O+LiBr
system particularly for optimized ILs.

The possible application of water and room temperature Ionic Liquids (RTIL) mixtures for the
absorption cooling cycle has been investigated using the present EOS and a simple absorption
cycle model. It was found that a H2O+RTIL system, when used with an optimized RTIL, could
compete with an existing H2O+LiBr system (32). Also the flow ratio and COP of series of H2O+IL
and NH3+water have been reported with the same cycle condition as T (generator)/T
(condenser)/T (absorber)/T(evaporator) = 100/40/30/10 °C, these properties which are shown in
Fig 4 are based on results taken from Yokozeki and Shiflett study (32)

COP
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0

Fig 4. COP of series of H2O +IL and H2O +LiBr(32)

932
 Flow ratio
25

20

15

10

Fig 5. Flow ratio of series of H2O +IL and H2O +LiBr(32)

3. Discussion

A summary was made among different Ionic liquids in order to choose suitable working pairs.
Among the most of ionic liquids, water was chosen as refrigerant due to its superior properties as
a heat transfer fluid; such as large latent heat of evaporation followed by extremely small flow
ratio resulted in its high performance. Selecting working fluids for absorption cycles should be
suitable for initial operating conditions and limitations, such as crystallization, corrosion problems,
and environmental limits and also material requirement should be considered.

4. Conclusions

In this paper a number of researches about working fluids of absorption cycles, which contain
absorption heat pumps, absorption chillers and absorption heat transformers including Ionic
liquids were reviewed. Environmental issues and saving energy concerns have always been a
major global problem. Absorption cycles can play an important role due to their capability of
reducing CO2 discharge and to reuse large amount of industrial waste heat. Since the
performance of absorption cycles mainly depends on the working fluids, the research aspect of
improving working fluids is quite essential and in progress. Using Ionic liquids as the working
fluids of absorption cycles can lead us to benefit from factors such as less crystallization, less
corrosion, low toxicity and non-flammability in comparison with conventional working fluids
including (NH3+water and LiBr+water).

933
 Reference

1. REN, J., ZHAO, Z. C. & ZHANG, X. D. (2011) Vapor pressures, excess enthalpies,
and specific heat capacities of the binary working pairs containing the ionic
liquid 1-ethyl-3-methylimidazolium dimethylphosphate, Journal of Chemical
Thermodynamics, 43, 576-583.
2. HORUZ, I. & KURT, B. (2010) Absorption heat transformers and an industrial
application, Renewable Energy, 35, 2175-2181.
3. HOLMBERG, P. & BERNTSSON, T. (1990) ALTERNATIVE WORKING FLUIDS IN
HEAT TRANSFORMERS, Ashrae Transactions, Vol 96, Pt 1, 1990: Technical and
Symposium Papers Presented at the 1990 Winter Meeting, 96, 1582-1589.
4. SUN, J., FU, L. & ZHANG, S. G. (2012) A review of working fluids of absorption
cycles, Renewable & Sustainable Energy Reviews, 16, 1899-1906.
5. ROGERS, R. D. & SEDDON, K. R. (2003) Ionic liquids - Solvents of the future?,
Science, 302, 792-793.
6. SEDDON, K. R. (2003) Ionic liquids - A taste of the future, Nature Materials, 2,
363-365.
7. KIM, S., KIM, Y. J., JOSHI, Y. K., FEDOROV, A. G. & KOHL, P., A. (2012 ) Absorption
Heat Pump/Refrigeration System Utilizing Ionic Liquid and
Hydrofluorocarbon Refrigerants, Journal of Electronic Packaging 134, 9.
8. MARSH, K. N., BOXALL, J. A. & LICHTENTHALER, R. (2004) Room temperature ionic
liquids and their mixtures - a review, Fluid Phase Equilibria, 219, 93-98.
9. VAN VALKENBURG, M. E., VAUGHN, R. L., WILLIAMS, M. & WILKES, J. S. (2005)
Thermochemistry of ionic liquid heat-transfer fluids, Thermochimica Acta,
425, 181-188.
10. ZHANG, X. D. & HU, D. P. (2011) Performance simulation of the absorption
chiller using water and ionic liquid 1-ethyl-3-methylimidazolium
dimethylphosphate as the working pair, Applied Thermal Engineering, 31,
3316-3321.
11. LIANG, S. Q., ZHAO, J., WANG, L. & HUAI, X. L. (2010) Absorption refrigeration
cycle utilizing a new working pair of ionic liquid type, Journal of Engineering
Thermophysics, 4.
12. LIANG, S. Q., CHEN, W., CHENG, K., GUO, Y. X. & GUI, X. H. (2011) The Latent
Application of Ionic Liquids in Absorption Refrigeration, in: Handy, P. S. (Ed.)
Applications of Ionic Liquids in Science and Technology.
13. KIM, Y. J., KIM, S., JOSHI, Y. K., FEDOROV, A. G. & KOHL, P., A. (2011) Waste-Heat
Driven Miniature Absorption Refrigeration System Using Ionic-Liquid as a
Working Fluid, Paper presented at the ASME 2011 5th International
Conference on Energy Sustainability Washington, DC, USA
14. KIM, Y. J., KIM, S., JOSHI, Y. K., FEDOROV, A. G. & KOHL, P. A. (2012)
Thermodynamic analysis of an absorption refrigeration system with ionic-
liquid/refrigerant mixture as a working fluid, Energy, 44, 1005-1016.
15. CHEN, W., LIANG, S. Q., GUO, Y. X. et al. (2012) Thermodynamic performances of
mmim DMP/Methanol absorption refrigeration, Journal of Thermal Science,
21, 557-563.

934
16. KERLE, D., LUDWIG, R., GEIGER, A. & PASCHEK, D. (2009) Temperature
Dependence of the Solubility of Carbon Dioxide in Imidazolium-Based Ionic
Liquids, Journal of Physical Chemistry B, 113, 12727-12735.
17. REN, W. & SCURTO, A. M. (2009) Phase equilibria of imidazolium ionic liquids
and the refrigerant gas, 1,1,1,2-tetrafluoroethane (R-134a), Fluid Phase
Equilibria, 286, 1-7.
18. HE, Z. B., ZHAO, Z. C., ZHANG, X. D. & FENG, H. (2010) Thermodynamic properties
of new heat pump working pairs: 1,3-Dimethylimidazolium
dimethylphosphate and water, ethanol and methanol, Fluid Phase Equilibria,
298, 83-91.
19. ZUO, G. L., ZHAO, Z. C., YAN, S. H. & ZHANG, X. D. (2010) Thermodynamic
properties of a new working pair: 1-Ethyl-3-methylimidazolium ethylsulfate
and water, Chemical Engineering Journal, 156, 613-617.
20. DONG, L., ZHENG, D. X., NIE, N. & LI, Y. (2012) Performance prediction of
absorption refrigeration cycle based on the measurements of vapor pressure
and heat capacity of H2O + DMIM DMP system, Applied Energy, 98, 326-332.
21. CAI, W. H., SEN, M., PAOLUCCI, S. & ASME (2008) Dynamic modeling of an
absorption refrigeration system using ionic liquids.
22. ISHIKAWA, M., KAYANUMA, H. & ISSHIKI, N. (1999) Absorption heat pump using
new organic working fluids, Paper presented at the ISHPC'99, Kyoto, Japan.
23. KIM, K. S., SHIN, B. K., LEE, H. & ZIEGLER, F. (2004) Refractive index and heat
capacity of 1-butyl-3-methylimidazolium bromide and 1-butyl-3-
methylimidazolium tetrafluoroborate, and vapor pressure of binary systems
for 1-butyl-3-methylimidazolium bromide plus trifluoroethanol and 1-butyl-
3-methylimidazolium tetrafluoroborate plus trifluoroethanol, Fluid Phase
Equilibria, 218, 215-220.
24. DONG, L., ZHENG, D. X., WEI, Z. & WU, X. H. (2009) Synthesis of 1,3-
Dimethylimidazolium Chloride and Volumetric Property Investigations of Its
Aqueous Solution, International Journal of Thermophysics, 30, 1480-1490.
25. KARAMANGIL, M. I., COSKUN, S., KAYNAKLI, O. & YAMANKARADENIZ, N. (2010) A
simulation study of performance evaluation of single-stage absorption
refrigeration system using conventional working fluids and alternatives,
Renewable & Sustainable Energy Reviews, 14, 1969-1978.
26. KAYNAKLI, O. & YAMANKARADENIZ, R. (2007) Thermodynamic analysis of
absorption refrigeration system based on entropy generation, Current
Science, 92, 472-479.
27. MARTIN, A. & BERMEJO, M. D. (2010) Thermodynamic analysis of absorption
refrigeration cycles using ionic liquid plus supercritical CO2 pairs, Journal of
Supercritical Fluids, 55, 852-859.
28. ZHAO, J., JIANG, X. C., LI, C. X. & WANG, Z. H. (2006) Vapor pressure measurement
for binary and ternary systems containing a phosphoric ionic liquid, Fluid
Phase Equilibria, 247, 190-198.
29. YOKOZEKI, A. & SHIFLETT, M. B. (2007) Vapor-liquid equilibria of ammonia plus
ionic liquid mixtures, Applied Energy, 84, 1258-1273.

935
30. YOKOZEKI, A. (2005) Theoretical performances of various refrigerant-
absorbent pairs in a vapor-absorption refrigeration cycle by the use of
equations of state, Applied Energy, 80, 383-399.
31. YOKOZEKI, A. & SHIFLETT, M. B. (2007) Ammonia solubilities in room-
temperature ionic liquids, Industrial & Engineering Chemistry Research, 46,
1605-1610.
32. YOKOZEKI, A. & SHILLETT, M. B. (2010) Water Solubility in Ionic Liquids and
Application to Absorption Cycles, Industrial & Engineering Chemistry
Research, 49, 9496-9503.
33. ROMICH, C. H., MERKEL, N., SCHABER, K. et al. (2011) A COMPARISON BETWEEN
LITHIUMBROMIDE - WATER AND IONIC LIQUID - WATER AS WORKING
SOLUTION FOR ABSORPTION REFRIGERATION CYCLES 23rd Iir International
Congress of Refrigeration, pp. 941-948.
34. KUHN, A., BUCHIN, O., SEILER, M., SCHWAB, P. & ZIEGLER, F. (2009) Ionic liquids-a
promising solution for solar absorption chillers 3rd Solar air-conditioning
conference (Palermo, Sicily, Italy).
35. RADPIELER, M. & SCWEIGLER, C. (2011) Experimental investigation of ionic
liquid EMIM+EtSO4 as solvent in a single effect cycle with adiabatic
absorption and desorption ISHPC (Padova, Italy).

936
 FULL SCALE EXPERIMENTAL STUDIES OF A PASSIVE COOLING ROOF
IN HOT ARID AREAS
H .Bencheikh¹
¹ Laboratoire de génie civil Amar Telidji University Laghouat Algeria
BO 7072 oasis Nord Laghouat 03000 Algeria
email :h.bencheikh@mail.lagh-univ.dz

ABSTRACT
A full scale Evapo-reflective roof for arid climates has been developed. The proposed roof design is composed of a
concrete ceiling over which lies a bed of rocks in a water pool. Over this bed is an air gap separated from the
external environment by an aluminum plate. The upper surface of this plate is painted with a white titanium-based
pigment to increase reflection of a radiation to a maximum during the day. At night, the temperature of the
aluminum sheet falls below the temperature of the rock bed mixed with water. Water vapor inside the roof
condenses and falls by gravity. This heat pipe effect carries heat outwards and cold inwards. Heat exchange is
improved by radiation between two humid internal surfaces. The efficiency of this cooling system is studied using
finite difference method. Numerical calculations performed for different external temperatures and solar radiation
show that the cooling produced by such a system is significant. As a result of this, the mean air temperature in the
room may be kept a few degrees above the minimum nocturnal outdoor temperature throughout the day. However,
the maximum indoor air temperature was observed at sunset. This could further be lowered by allowing ventilation
of the building in the evening.

A full scale experimental study of passive cooling roof was carried out for a typical summer day of June for
Laghouat in Algeria. The proposed roof design is composed of a concrete ceiling over which lies a bed of rocks in
a water pool. Over this bed is an air gap separated from the external environment by an aluminum plate. The
upper surface of this plate is painted with a white titanium-based pigment to increase the radiation reflection
process during daytime. Several passive modifications have been introduced to the roof in order to reduce indoor
air temperature in hot climates. An experimental investigation, employing passive procedure, has been carried out
to study the possibility of reducing air temperature in buildings. The results show that the air temperature can
decrease with a range from 6 to 10°K. This decrease can further be lowered by 2 to 3°C if night natural ventilation
of buildings is allowed.
Keywords: Evaporative cooling, Evapo-reflective roof, hot dry climate, Night ventilation, dynamic model.

INTRODUCTION
In hot climates such as southern Algeria, excessive heat is the major problem that causes human thermal
discomfort. Space cooling is therefore the most desirable factor for the inhabitants. Various examples of
dwellings responsive to climatic constraints were found in vernacular architecture throughout the world.
Compact cellular layout with minimum external surface exposure to the sun, whitewash surfaces to reduce
absorptivity, blind external facades, courtyards, vegetation to provide humidity and shade, heavy buildings
with high thermal capacity materials are common passive features in most hot arid climates such M'Zab
settlements in southern Algeria, Egypt and Iran [1-4]. Wind towers for cooling ventilation are well known in
Iranian and Middle East architecture, which along with cooling of the air by water evaporation kept the building
comfortable in hot periods [5]. Underground Buildings have the advantage of large thermal capacity storage of
the earth. They are used in Matmata in Tunisia and Cappadocia in central Turkey [2].

In recent years several investigations were performed and showed that there can be multiple solutions to
the excessive heat problem. Popular is cooling ventilation using a solar chimney [2, 6, 7]. The results showed
that cooling ventilation using a solar chimney can reduce internal temperatures of buildings. Shading devices
(overhangs and verandas) to reduce summer solar radiation are also investigated and useful depths of
these shading elements for various orientations in continental climates were defined [8].

Space cooling can also be achieved by improving of the performance of roofs. This is because the roofs
are the most exposed surfaces to direct solar radiation and can cause excessive heat gain in hot periods.
Some efforts were made by investigators to improve the roof thermal performance. The use of low emissive
material in the attic of a roof reduced the underside ceiling surface temperature which lowered the room
air temperature [9]. Evaporative cooling approach for passive cooling of buildings in hot arid climates has
also become an attractive subject of investigation for many researchers. The relative advantage of evaporative

937
cooling in relation to many other approaches (cavity wall, insulation, white wash and large exposure
orientations, vegetable pergola shading, roof with removable canvas, water film , soil humid grass and roof
with white pots cover) were demonstrated by [10, 11].

The reduction of heat gain through the roofs using evaporative cooling systems was extensively
investigated on open roof pond [18, 19], on water spraying over the roof, moving water layer over the roof, thin
water film and roofs with wetted gunny bags [12-17]. Chandra and Chandra [12] have developed a periodic
heat transfer model to study the effects of evaporative cooling using water spray and variable ventilation on
the temperature control of a non-air-conditioned building. The influence of evaporative cooling over the roof as
compared to bare roof case and intermittent ventilation as compared to the continuous or no-ventilation case
have been assessed for controlling the indoor air temperature. It was found that the effectiveness of the
evaporative cooling can be improved by conscious choice of the rate and duration controls the inside air
temperature significantly. It was concluded that a combination of evaporative cooling and variable ventilation
can make the internal environment of a building more comfortable. Chandra et al. [13] presented a
theoretical assessment of three roof cooling systems for a non air conditioned building, and showed that
the maximum cooling is achieved by water spray over the roof. But the roof pond system with stationary
water is more effective in stabilizing the fluctuations of indoor temperature.

The present study suggests an improved roof design by combining the advantages of previous described
cooling techniques (water ponds, low emissive surfaces) and inserted rocks of high thermal capacity
materials. The resulting design can be more advantageous and effective than other systems for reducing heat
during daytime and storing coolness at night. High thermal capacity materials (rocks bed) will delay the entry
of daytime heat into the building by such a period that it reached the interior during the evening when it is
least bothersome or often welcome. The roof is composed of a concrete ceiling and a flat aluminum plate
separated with air space partially filled with rocks inserted in small quantity of water. The system is closed to
prevent water vapor skipping outside. A schematic diagram of the model design is shown in figure 1.

Fig. 1. (a) Room with cooling roof system Fig. 2. (b) Room without cooling roof system.

MATHEMATICAL MODEL
The basic configuration of the model considered here are shown in figure 1. It is a cubic room with 3m high and 3m
wide. South wall is provided with a window and the North one is provided with a door. Physical properties of materials
used for the roof are presented in table 1.

The purpose of the present mathematical model is to determine the inside air temperature at each time step as a function of
outside air temperature, solar radiation and heat due to ventilation. The solution is based on the inside heat balance at each
time step, the method of lagging with zone capacitance uses information from previous time steps to predict system response
and update the zone temperature at the current time. 15 minutes is used as a time step (the shorter the time step the smaller the
error). The simulation was done for the described model for two highest temperature summer days. The model situated in
Algeria (Laghouat, latitude +33.46°, longitude +2.56° and elevation 767 m). The simulated day was the 26 of July, the
maximum and the minimum temperature were respectively 42.7°, 24.5°.

938
 Table1. Material properties
Material Density Specific Conductivity
Kg/m3 heat (w/m°k)
(j/kg °k)
Concrete slab 2400 1080 1.8
Rocks 2600 800 2.3
Water 1000 4175 0.613
Aluminum 1.22 1008 0.026
Concrete slab 2750 936 204

Inside air heat balance equation

The heat balance for the inside room air may be formulated as follows:
dTai
Cai  Qint + Qci +Q ( 1)
dt v
dTai
Cai  : The heat stored in the inside air Cai = mai .cai
dt
in
Qint =  Qint i (w)
i 1
in
Qci =  hi Ai (Tsi  Tai ) (w) (2)
i 1
Qv = mae .cae .(Tae  Tai ) (w) (3)
dTai
The derivative term can be expressed by finite differences approximations as follow;
dt
dTai t t
= (Tai  Tai ) 1
t
(4)
dt t
From the above equations the inside air temperature;

t t
C 
 Qi   ai Tai   Ai hi Tsi  mae caeTae 
t

Tait   t  (5)
Cai
   Ai hi mae cae 
t
Surface temperatures
To calculate the internal surface temperatures Tsit , , at each time step t, as function of outside conditions, finite
deference equations based on heat balance at each node were used, which allows for temperature determination
at any point of interest. The first step is to select these points, by subdividing the medium into a number of small
regions represented by reference points called nodes. In our case, we considered the heat flow in one direction in
plan elements (walls, roof and floor) composed of deferent materials, so each layer of these material is divided into
small regions and represented by nodes. Clarke suggested that three nodes per homogeneous element and a 1 h
time step, in building applications are consistent with acceptable accuracy [20]. The temperature for each single
node at time t is evaluated using heat balance equations.
The heat exchange between internal slab nodes is modeled using Fourier’s one dimensional heat conduction
equation [21]

(6)
dT
 d d ²T
dt c dx ²

939
This equation can be solved numerically [21] by dividing the element into layers of thickness dx called nodes,
making a heat balance for each node. The boundary condition for the inside surface nodes in contact with room air
may be given by:
(7)

The boundary condition for the outside surface nodes in contact with outside air may be formulated using the
following equation:

(8)

The upper roof surface exchanges heat with the outside air by convection and by radiation to the sky. According to
[23], a horizontal surface with emissivity and absolute temperature Tr, produces a net radiative cooling rate Qr,
where;

(9)
Where

The surface dew point temperature was computed as a function of the ambient temperature (Tae) and the relative
humidity (RH), using the expression by Murray [24]:
(10)

The heat exchange by convection for outside horizontal surface is given by:
(11)

The heat exchange between the lower aluminum surface and the upper rock bed surface is by radiation,
convection and evaporation. Following equations reported in [25], we may write the following.
Heat exchange by radiation is given by:
(12)

Heat exchange by convection is given by:

(13)

Heat exchange by evaporation and condensation is given by:

(14)

Where L is the latent heat of evaporation at an average temperature, which is equal to 2350 kJ/kg at temperature
T in °C. For the temperature range

The following polynomial gives acceptable results [25]:

(15)

In modeling the floor elements the earth temperature at 60 cm of depth below the floor is considered constant and
equal to the daily average temperature of the region [2]. In the above equations the number of the unknowns is
greater than the number of equations; these equations were solved by proposing the initial inside air temperature
at start time t. This initial temperature Tai (t) will not be correct and it is necessary to simulate the model with the
same daily repetition of air temperature and solar radiation until the temperature of each node returns to the same
value at the same time in each day simulation. At this point the building is in thermal harmony with the
environment.

EXPERIMENTAL MEASUREMENTS
The experimental set-up consisted of two identical test cells (A) and (B), a cubic room with 3m high and 3m wide.
South wall is provided with a window and the North one is provided with a door. The experimental cell (B) was the
basic reference unit. The roof was constructed of simple aluminum sheet painted white .

940
TEMPERATURE MEASUREMENTS
Air temperatures outside the room were measured using weather stations installed near the laboratory, far from the
test cell by 150m. The temperature at different positions under the roof level has been measured by copper
constant thermocouples connected to digital thermometer. Thermocouples fixed under the roof surface the end of
the thermocouples were enveloped in thin aluminum paper to reflect the radiation from the surrounding interior
surfaces. The readings of all thermocouples have been averaged to give the average temperature.

RESULTS AND DISCUSSION

Figures 3 shows the calculated and measured inside and outside air temperature in test cell (A) with cooling
system, without nocturnal natural ventilation. When the maximum outside temperature was 39 °C the maximum
measured and calculated temperatures were respectively 36°C and 35°C, the minimum ones were 27°C and
29°C. The maximum inside air temperature accrued at 19.00PM when the outside one was 32°C, at this time
natural ventilation was associate. Figure 4 shows the calculated and measured inside and outside air temperature
in cell (A) with cooling system, with nocturnal natural ventilation. The ventilation lower the inside air temperature by
3 to 4°C.

Figure 5 and 6 show the calculated and measured inside and outside air temperature in test cell (B) without
cooling system, with and without nocturnal natural ventilation. The calculated and measured inside air temperature
in figure 5 present no differences whoever in figure 6 the two graphs present small differences during ventilation
period due to the uncertainty in wind speed which variable from hour to another.

Fig.7. shows a comparison of room air temperatures with cooling roof system and with bare roof without room
night natural ventilation. It can be seen from this figure that the evaporative reflective roof can reduce the internal
room air temperatures during the day up to 10 °C in comparison to the air temperatures for a bare roof over the
room. Fig.7. is the comparison of room air temperatures with cooling roof system and with bare roof when room
night natural ventilation is allowed. The ventilation was allowed from 8 pm till 9 am, a period when the outside air
temperature is relatively low. This can significantly improve cooling of room air temperatures, as shown in Fig. 8.

Figure.3: Calculated and measured temperature in test cell (A) Figure.4 : Calculated and measured temperature in test cell (A)
without night natural ventilation with night natural ventilation

941
Figure. 5: Calculated and measured temperature in test cell (B) Figure.6: Calculated and measured temperature in test cell (B
without night natural ventilation with night natural ventilation

Figure.7: Comparison of room air temperatures, test cell (A) and (B) Figure.8: Comparison of room air temperatures, test cell (A) and (B
without night natural ventilation. with night natural ventilation

CONCLUSION
In comparison between calculated and measured inside air temperatures in cell (B) without cooling system, without
nocturnal natural ventilation, the two temperatures have almost the same values , however the calculated and
measured ones in the same cell with nocturnal natural ventilation, present a small difference between calculated and
measured temperatures during ventilation period as shown in figure 8, that due to wind speed variations during
night time, which was in calculations usually considered constant value.

Measured and calculated temperature in cell (A) with cooling system, with and without nocturnal natural ventilation,
presents a small differences in two periods time, from 6.00 Am till 1500Pm and form midnight till 4.00 Am which
correspond to the evaporations and condensations periods. The differences were due to that the quantities of
water vapor and condensate water were not exactly well known.

Under hot arid conditions a full scale test cell for an evaporative reflective roof used to improve space cooling in
buildings has been tested. The experimental results examined the effectiveness of such a roof cooling system in

942
comparison to a bare roof. The results showed that cooling inside buildings can be improved by the application of
such a cooling design. It was also seen that combining evaporative reflective roof with night ventilation increases
such cooling more significantly.

NOMENCLATURE
Cai specific heat of inside air (J / kg°K)
Cae specific heat of outside air (J/ kg° K)
E surface emissivity
I total solar radiation (W/ m²)
Ij long wave radiation (W / m²)
hci inside convection heat transfer coefficient (W /m² ° K)
hr radiation heat transfer coefficient (W /m² ° K)
²°
hc,wr,al convection heat transfer coefficient between the rock bed and aluminum (W /m K)
Pvs saturated vapor pressure (kPa)
v
Qint heat exchange by convection in the inside
Qci heat exchange by convection between aid and other surfaces
Qv heat exchange by ventilations and infiltration
Tal aluminum outside surface temperature (° C)
Tao solar temperature
v
Twr rock bed upper surface temperature (° C)
Tdp surface dew point temperature in (° C)
σ Stefan–Boltzman constant

REFERENCES
[1] Bouchair A. Dupagne A. Building traditions of Mzab facing the challenges of re-shaping of its built form
and society. Building and Environment, 38;1:1345–1364.
[2] Bouchair A. Solar induced ventilation in the Algerian and similar climates. PhD thesis. UK:
University of Leeds; 1989.
[3] Golany G. Housing in arid lands: design and applications. London: The Architectural Press; 1980.
[4] Evans M. Housing, climate and comfort. London: The Architectural Press; 1980.
[5] Bahadori MN. Passive cooling systems in Iranian architecture. Scientific American 1978;238:144–54.
[6] Bouchair A. Solar chimney for promoting cooling ventilation in southern Algeria. Building Services
Engineering Research and Technology 1994.
[7] Bouchair A, Fitzgerald D, Tinker JA. Passive solar induced ventilation. Alternative energy sources VIII.
Proceedings of the Miami International Conference on Alternative Energy Sources (1987: Miami
Beach, Florida). Solar energy fundamental and applications, vol. 1. New York, USA: Hemisphere
Publishing Corporation; 1989.
[8] Pour-Deihimi S. Control of sunshine in buildings by fixed shading in continental climates. PhD the- sis.
Leeds university; 1984.
[9] R. Lambert. Heat transfer through roofs of low cost Brazilian houses. PhD thesis. UK: University of
Leeds; 1988.
[10] Verma R, Bansal NK, Garg HP. The comparative performance of different approaches to passive
cooling. Building and Environment 1986;21(N2):65–9.
[11] Nayak JK, Srivastava A, Singh U, Sodha MS. The relative performance of different approaches to
passive cooling of roofs. Building and Environment 1982;17(2):145–61.
[12] Chandra S, Chandra S. Temperature control in a building with evaporative cooling and variable
ventilation. Solar energy 1983;30(4):381–7.
[13] Chandra S, Kaushik SC, Bansal PK. Thermal performance of a non-air-conditioned building for
passive solar air-conditioning: evaluation of roof cooling systems. Energy and buildings 1985;8:51–69.
[14] Nahar NM, Sharma P, Purohit MM. Performance of different passive techniques for cooling of
buildings in arid regions. Building and Environment 2003;38:109–16.

943
[15] Runsheng T, Etzion Y, Erell E. Experimental studies on a novel roof pond configuration for the cooling
of buildings. Building and Environment 2003;28:1513–22.
[16] Sodha MS, Kharty AK, Malik MAS. Reduction of heat flux through a roof by water film. Solar Energy
1978;20:189.
[17] Sodha MS, Srivastava A, Kumar A, Tiwari GN. Heating and cooling of a building by flow of
water over the roof. Applied Energy 1980;7(1):229.
[18] Sodha MS, Kumar A, Singh U, Srivastava A, Tiwari GN. Experimental validation of thermal
model of open roof pond. Building and Environment 1981;16(2):93.

[19] Sodha MS, Singh U, Tiwari GN. Periodic theory of an open roof pond.
Applied Energy1980;7(4):307.
[20] Clarke AJ. Energy simulation in building design. Bristol: Adam Hilger; 1985.
[21] Kreith F. Principles of heat transfer. New York: Index Educational Publishers; 1985.
[22] CIBSE guide. Vol. A. London: Chartered Institution of Building Services Engineers; 1986.
[23] Berdahl P, Fromberg R. The thermal radiance of clear skies. Solar Energy 1982;29(4):299–314.
[24] Murray FW. The computation of saturated vapor pressure. Journal of Applied Meteorology
1967;6:203–4.
[25] Sfeir AA, Guarracino G. Ingénerie des systeme solaires applications à l’habitat. Technique et
documentation; 1981.

944
 EXERGETIC ANALYSIS OF A VERTICAL GROUND-SOURCE HEAT PUMP SYSTEM
WITH WALL HEATING/COOLING
1 2 2 3
Ugur Akbulut , Ozgen Acikgoz , Olcay Kincay , T. Hikmet Karakoc
1
Recep Tayyip Erdogan University, Faculty of Engineering, Rize, Turkey
2
Yildiz Technical University, Mechanical Engineering Department, Istanbul, Turkey
3
Anadolu University School of Civil Aviation, Yunusemre Kampusü, Eskisehir, Turkey

ABSTRACT
The present study deals with an exergetic analysis and assesment of a Vertical Ground Source Heat Pump
System (VGSHP) combined with a Wall Heating System (WHS) in a building. This study is an experimental
investigation of a real building’s heating system. System is located at Yildiz Renewable Energy House (YREH)
in Yildiz Technical University and fulfills the heating demand of YREH and a living room of the neighbor
dormitory. In order to validate an exergetic model, system is divided into three subsystems such as: (i) the
ground coupling circuit, (ii) the refrigerant circuit and (iii) the WHS circuit. The schematic diagram of the
constructed experimental system is given in Fig 1. Exergetic model is obtained by applying mass, energy and
2
exergy equations for each system component. YREH has 4 rooms each one’s floor area is 8 m and neighbor
2
dormitory has a 50 m living room. In this study 3 rooms of YREH and living room have been heated in heating
season. Heating season was assumed to be between 1 January and 31 March. As average results on heating
season 6.509 kW heat energy was extracted from ground and 5.799 kW was used in the WHS. In this process
electrical energy consumption of system components are; compressor 1.711 kW, ground heat exchanger pump
0.092 kW, accumulator tank circulation pump 0.114 kW and WHS circulation pump 0.108 kW. For heating
season, calculated overall system efficiency was 67.36% while GSHP unit’s efficiency was 85%. In addition,
overall system COP was 2.76 while GSHP unit’s COP 4.13. Total exergy destruction was found 1.759 kW and
largest exergy destruction has occurred in the compressor as 0.714 kW. The exergy efficiency values for the
individual components of the system have been found ranging from 58.3% to 98.4% according to P/F concept.
It is expected that the model would be beneficial for evaluating low exergy heating systems which use ground
source as renewable energy.

Keywords: energy analysis, renewable energy, vertical ground source heat pump system, wall heating system

INTRODUCTION
Ground-source heat pump (GSHP) systems make use of renewable energy stored in the ground for building
heating and cooling. They are suitable for a wide variety of building types, provide high levels of comfort, and
are particularly appropriate for energy saving and environmentally attractive (Sanner et al., 2003; Lund, 2003;
Esen et al., 2006; Ozgener et al., 2007a; Kavanaugh et al., 1992; Agustos et al., 2008; Akbulut et al., 2012).
Many theoretical and experimental works on GSHP systems have been accomplished, since late 1940s. GSHP
systems are inherently more effective than air source heat pumps because the ground maintains relatively
stable source and sink temperatures. This situation provides a better COP for the GSHP system. For the
ground heat exchangers, VGSHP systems are usually preferred over HGSHP systems. They are more
efficient and less ground area is required ( Grandall, 1946; Ingersol et al., 1948; Hepbasli et al., 2002 and
2007; Li et al., 2006; Esen et al., 2007; Lund et al., 2005; Urchueguia et al., 2008).

WHC systems have low operation temperature benefit. This creates substantial energy savings for heating and
cooling compared with conventional systems. The energy savings from a WHC system may reach more than
30%, as demonstrated in some theoretical and experimental case studies. Main advantages of WHC systems
are enabling better thermal comfort, providing better indoor air quality with low air velocities and homogenous
heat distribution, being available to use waste heat and low-enthalpy renewable energy resources, and having
low initial investment, maintenance, and operating costs (ASHRAE, 2004). Researchers have mostly
determined the convection heat transfer coefficient of heated and cooled walls. They have solved the natural
convection problems in enclosures. In these studies, heat distribution from floors, walls, and ceilings were
investigated for heating and cooling situations. All studies considered in the literature about WHC systems
emphasized that this system was comfortable, economical, and very suitable to use with renewable energy
systems. Because of this reason, a real VGHP system, combined with a WHC system, has been analyzed
experimentally in our previous research. Energetic aspects of the system were introduced. However, exergy
analysis of the system hasn’t been conducted yet (Yoru et al., 2010; Kincay et al., 2010; Akbulut et al., 2011;
Başkal et al., 2012).

945
Exergy analysis is widely used for to scale a process’s thermodynamic ideality. This helps designing efficient
and cost effective systems that also meet environmental conditions. In addition to the energy analysis, exergy
analysis must be used to identify the components where inefficiencies occur. Improvements should be done to
these components to minimize the irreversibility and optimize the system (Moran, 1989; Kotas, 1995; Bejan et
al., 1996; Rosen, 1999; Dincer et al., 2001). Recent years, exergy analyses of space heating/cooling in
buildings and GSHP systems have been developed (Bejan et al., 2002; Yildiz et al., 2009; Kincay et al., 2004;
Hepbasli et al., 2004; Utlu et al., 2007; Ozgener et al., 2007b; Bi et al., 2009; Madani et al., 2010). These
studies focused on the building heating mode, and few on the building cooling mode. This paper presents an
exergy analysis of a real VGSHP system, combined with a WHC system. Real-time data has been obtained
and used to represent destroyed exergy and exergy efficiency for both heating and cooling modes.

SYSTEM DESCRIPTION
The investigated VGSHP system with the WHC is shown in Fig. 1. This system consists of a ground heat
exchanger loop, heat pump unit, accumulation tank, and a WHC loop in the building. The vertical ground heat
exchanger consists of two boreholes, each containing a U-tube pipe. The depth of the boreholes is 65 m, and
the diameter is 0.15 m. For generating the ground-loop heat exchanger, 240 m DN40 polyethylene composite
pipe was used in a vertical direction and 45 m in horizontal. There are three circulation pumps and an isolated
500 lt accumulation tank in the system. In the WHC system, 840 m DN16 and 40m DN20 polyethylene pipe
was used. Systems were set up for holding the test room temperature at 20°C on the heating and 24°C on the
cooling mode. For this reason, 31°C and 18°C V-GSHP operating temperatures were chosen especially to
maintain comfort and prevent condensation on wall-mounted coils. Heating season was analyzed using data
collected between Jan. 1 and March 31, 2010. In this period, the GSHP unit was online a total of 863 hours for
a mean of 30%. Furthermore, cooling season was analyzed using data collected between July 1 and
September 30. In this cooling period the GSHP unit was online a total 120 hours for a mean of 5%.

Fig. 1. Schematic of investigated VGSHP and WHCS on heating mode.

Schematic diagrams of the constructed experimental system on heating and cooling modes are illustrated in
Fig. 1 and Fig. 2. This system mainly consists of three separate circuits as follows: (i) the ground coupling
circuit, (ii) the refrigerant circuit, and (iii) the WHC system circuit. In the heating session, three rooms of
YREH and the living room were heated; only two rooms of YREH were cooled in the cooling session.
Conversion from the heating cycle to the cooling cycle is implemented by means of a four-way valve. The
working fluid is R-410A. This system is installed at Yildiz Renewable Energy House (YREH) at Yildiz
Technical University (latitude 41° N, longitude 29° E), Istanbul Turkey.

946
Fig. 2. Schematic of investigated VGSHP and WHC system on cooling mode.

EXERGETIC MODELING
The general exergy balance can be expressed below Eq. (1), as the total exergy input is equal to total
exergy output:

(1)

where is the heat transfer rate crossing the boundary at temperature at location k, the work rate,
the flow exergy which can be calculated by Eq. (2).

(2)
Here denotes the enthalpy, the entropy, and the subscript zero indicates properties at the restricted dead
state of and .

The exergy rate is calculated by Eq. (3).

(3)

To obtain exergy destruction, the entropy generation is calculated first and used in the Eq. (4) which is
called as Stadola Law.

(4)

The exergy destructions in the system components are calculated as follows, respectively:

 Compressor and pumps

(5)

 Heat exchangers ( evaporator, condenser and accumulator tank)

(6)

947
  Expansion valve
(7)

 Ground heat exchanger in heating session

(8)

 Ground heat exchanger in cooling session

(9)

 WHCS in heating session

(10)

 WHCS in cooling session

(11)

System components’ exergy efficiencies are calculated on product/fuel basis, by using Eq.12

(12)

RESULTS AND DISCUSSION

During the calculations for the system energy losses from the ground heat exchanger, the expansion valve
and isolated pipes were neglected. The dead state temperature, which is the dry-bulb temperature and the
dead state pressure were taken 0.01°C and 101.325 kPa, respectively. After modeling the system and
analysis of given data in Tables 1 and 2, exergetic results for the vertical ground-source heat- pump system
with wall heating and cooling were obtained by using thermodynamic equations given in the exergetic
modeling section. Results are shown by Tables 3 and 4. In the year 2010, heating season was assumed to
be between Jan. 1 and March 31, and cooling season between July 1 and Sept 30. Data was recorded for
each second and converted to a MySQL database. Collected data was arranged and transferred to other
software for calculations and analysis. Refrigerant (R410A) properties were taken from Solkane 6.0, and
water properties were taken from EES (Limited Academic version) software. Microsoft Excel pages were
formed for calculations. After the calculations and analysis, for the heating and cooling seasons, average
values of mass flow, temperature, pressure, enthalpy, and exergy rate are given in Tables 1 and 2,
respectively. Also, the exergy rate of fuel, exergy rate of product, destroyed exergy, energy efficiency and
exergy efficiency are shown in Tables 3 and 4.

Table 1. Measured data and calculated values of system in the heating session
No Name of element [kg/s] T[°C] P[kPa] h[kJ/kg] s[kJ/kgK] [kW]
0 Water (Dead state) - 0.01 0.1032
101.325 4.279x10‾6 0
0′ Refrigerant (Dead state) - 0.01 439.18
101.325 2.0982 0
1 Evaporator outlet/compressor inlet 0.035 7.97 795 429.62 1.84070 2.127
2 Compressor outlet/ condenser inlet 0.035 63.16 2021 465.84 1.86900 3.124
3 Condenser outlet 0.035 28.37 2021 245.49 1.15500 2.238
4 Evaporator inlet 0.035 -0.14 795 245.49 1.16670 2.127
5 Ground heat exchanger water outlet 0.44 7.48 257 31.53 0.11350 0.186
6 Ground heat exchanger water pump inlet 0.44 3.89 250 16.46 0.05945 0.052
7 Ground heat exchanger water inlet 0.44 3.94 300 16.67 0.06021 0.053
8 Water circulating pump inlet 0.57 29.65 250 124.3 0.43170 3.577
9 Water circulating pump outlet 0.57 29.70 300 124.5 0.43240 3.582
10 Heat pump outlet 0.57 32.66 293 136.9 0.47328 4.285
11 Accumulator tank outlet 0.27 30.60 250 128.3 0.44470 1.815
12 Heating water inlet 0.27 30.70 300 128.7 0.44610 1.820
13 Accumulator tank inlet 0.27 24.77 250 103.9 0.36378 1.196

948
Table 2. Measured data and calculated values of system in the cooling session
No Name of element [kg/s] T[°C] P[kPa] h[kJ/kg] s[kJ/kgK] [kW]
0 Water (Dead state) - 0.01
101.325 0.1032 4.279x10‾6 0
0′ Refrigerant (Dead state) - 0.01 439.18
101.325 2.0982 0
1 Compressor outlet/ condenser inlet 0.035 44.71 1668 450.21 1.81590 2.439
2 Evaporator outlet/compressor inlet 0.035 17.71 1113 431.75 1.83890 2.865
3 Evaporator inlet 0.035 10.87 1113 233.27 1.11490 2.194
4 Condenser outlet 0.035 21.08 1668 233.27 1.11680 2.176
5 Ground heat exchanger water outlet 0.44 20.80 257 87.27 0.30760 1.383
6 Ground heat exchanger water pump inlet 0.44 24.59 250 103.10 0.36110 1.918
7 Ground heat exchanger water inlet 0.44 24.63 300 103.30 0.36180 1.922
8 Water circulating pump inlet 0.57 17.66 250 74.14 0.26265 1.307
9 Water circulating pump outlet 0.57 17.70 300 74.31 0.26325 1.310
10 Heat pump outlet 0.57 14.53 293 61.05 0.21730 0.906
11 Accumulator tank outlet 0.27 17.91 250 75.19 0.26620 0.641
12 Heating water inlet 0.27 18.00 300 75.56 0.26750 0.645
13 Accumulator tank inlet 0.27 24.26 250 101.70 0.35650 1.138

Table 3. Exergetic analysis results for representative unit in the heating session
No Name of element [kW] [kW] [kW] η [%] ε[%]
I Overall system 2.254 0.496 1.759 67.36 22.0
II GSHP unit 1.840 0.698 1.142 85.00 37.9
III Pump1 0.092 0.001 0.091 70.75 1.1
IV Ground heat exchanger 0.230 0.133 0.097 100.00 57.9
V Evaporator 0.134 0.001 0.133 97.63 0.5
VI Compressor 3.838 3.124 0.714 74.10 58.3
VII Expansion valve 2.238 2.127 0.112 100.00 95.0
VIII Condenser 0.886 0.703 0.183 91.65 79.3
IX Pump 2 0.114 0.005 0.109 87.69 4.4
X Pump 3 0.108 0.005 0.103 83.08 4.4
XI Accumulator tank 0.708 0.620 0.088 91.73 87.6
XII WHCS 0.625 0.496 0.129 86.61 79.4

For the heating season, the exergy efficiency peak values for the expansion valve and accumulator tank
were found 95% and 87.6% respectively. During the calculations energy loss from the expansion valve was
neglected. Thereby, energy and exergy efficiency values for this component are suitable for these working
conditions. Also, energy and exergy efficiency values for the accumulator tank seem to be normal and prove
that the tank is well isolated. On the contrary, the exergy efficiency values are lowest for the evaporator and
three circulation pumps. We should consider the energy efficiency values in order to decide whether these
components are inefficient and should be replaced. These components energy efficiency values are 97.63%,
70.75%, 87.69% and 83.01% respectively. These values are suitable for a plate heat exchanger and small
capacity circulation pumps. Thus, evaporator and pump1 have too small exergy efficiency values since their
working temperatures which are too close to the dead state. Also, all three circulating pumps are not
changing the temperature of water much. Whole exergy is destroyed on the pumping process. However, our
overall system’s energy and exergy efficiency values are 67.36% and 22.00% respectively and definitely we
should develop some system components. Main indicator for this consideration is exergy destruction rate.
That’s why compressor, WHCS and the ground heat exchanger must be developed respectively.
Improvements on the compressor will positively affect the performance of the condenser if the gas
temperature of compressor outlet can be reduced.

949
Table 4. Exergetic analysis results for representative unit in the cooling session
η [%] ε[%]
No Name of element [kW] [kW]
[kW]
I Overall system 1.892 0.567 1.325 74.85 29.9
II GSHP unit 1.470 0.535 0.935 80.76 36.4
III Pump1 0.130 0.004 0.126 67.69 3.0
IV Ground heat exchanger 0.539 0.436 0.103 100.00 80.9
V Evaporator 0.404 0.263 0.141 91.91 65.1
VI Compressor 1.066 0.426 0.640 60.60 40.0
VII Expansion valve 2.194 2.176 0.018 100.00 99.2
VIII Condenser 0.671 0.535 0.136 91.73 79.7
IX Pump 2 0.130 0.003 0.127 74.54 2.7
X Pump 3 0.130 0.004 0.126 76.85 3.1
XI Accumulator tank 0.498 0.400 0.098 95.93 80.4
XII WHCS 0.567 0.494 0.073 90.29 87.2

For the cooling season, the exergy efficiency peak values for the expansion valve, ground heat exchanger
and accumulator tank were found 99.2%, 80.9%, and 80.4% respectively. The expansion valve and
accumulator tank have similar results in the heating and cooling seasons because of the reasons explained
above. But the ground heat exchanger works very different. Much difference on the exergy efficiency values
exist. Because the ground heat exchanger water inlet temperatures are too different, 3.94 °C and 24.63 °C
respectively. There is a higher heat transfer coefficient due to water molecules activity at 24.63 °C
temperature and also 3.94 °C is very close to dead state. However, in both cases the exergy destruction
rates are similar, 0.097 kW and 0.103 kW respectively. The three circulating pumps have similar results
working in both heating and cooling sessions. Therefore, the compressor, the WHCS and the ground heat
exchanger must be developed to make the system more efficient in both sessions.

CONCLUSIONS
We have presented exergetic aspects of a vertical ground-source heat pump system combined with wall
heating/cooling in general. We analyzed the overall system components such as compressor, condenser,
evaporator, expansion valve, ground heat exchanger, accumulator tank and WHCS. We used energy and
exergy analysis results of experimental data from our real system which works for heating and cooling
purposes. Combining this study with our previous work which presents energetic aspects of the same system
(Akbulut et al., 2012), we can point out some concluding remarks from this study as follows:

(a) In the literature, main indicator to choose system’s inefficient components is exergy destruction rate.
Nonetheless, exergy efficiency and exergy efficiency values must be considered too attentively if most of the
components’ exergy destruction rates are very similar like in our cases.
(b) Technical and economical availabilities and environmental aspects must be considered much before
developing the system. That’s why exergy economic and exergy environmental analysis of this kind of
systems are needed.
(c) For the heating season, the exergy efficiency peak values for the expansion valve and accumulator tank
are 95% and 87.6% respectively. For the cooling season, the exergy efficiency peak values for the expansion
valve, ground heat exchanger and accumulator tank are 99.2% 80.9 and 80.4%.
(d) Calculated exergy efficiencies of the circulation pumps seem low at first glance. Because their operating
temperatures are too near to dead state. Pumps are not changing the temperature of water much and all of
the exergy is destroyed on pumping process. Their energy efficiency values are within normal limits.
(e) As a result of exergy analysis of this system, the order of development priority is the compressor, the
WHCS and the ground heat exchanger, from much to less.
(f) Heat extraction rate of 50 W/m and heat rejection rate of 54 W/m of bore depth for the heating and cooling
periods, remain within the range reported in the literature. But ground heat exchanger still has a
development capacity.
(g) For heating season, calculated overall system efficiency was 67.36%, while the GSHP unit’s efficiency
was 85%. Adding to that, the overall system COP was 2.76, while the GSHP unit’s COP was 4.13. As for the
cooling season, calculated system efficiency was 74.90%, while the GSHP unit’s efficiency was 80.76%.
Furthermore, the overall system COP was 4.38 while the GSHP unit’s COP was 4.78.
(i) The circulator wattage for the three closed loop of the system can be categorized as efficient and
excellent. However frequency converter circulation pumps can be used for energy saving.

950
(h) The system can be improved by using a variable speed compressor. According to the literature, increase
in the compressor speed lowers the heat pump COP. Optimum compressor speed is around 50 Hz (Madani
et al., 2010).

ACKNOWLEDGEMENT
The authors gratefully acknowledge the financial support from the Scientific Research Projects
Administration Unit of Yildiz Technical University (YTU-BAPK / 27-06-01-03, 2007).

NOMENCLATURE
COP coefficient of performance, dimensionless
exergy rate, kW
h specific enthalpy, kJ/kg
mass flow rate, kg/s
P pressure, kPa
heat transfer rate, kW
T temperature, K
rate of work or power, kW
s specific entropy, kJ/kgK
entropy rate, kJ/K

Greek Letters
exergy (second law) efficiency, dimensionless
energy (first law) efficiency, dimensionless
specific exergy, kJ/kg

Subscripts
0 reference (dead) state
C cooling
c/p compressor or pump
dest destroyed
expv expansion valve
F fuel
gr ground
H heating
HE heat exchanger
in input
k k’th element
P product
ref refrigerant
out output
whsc wall heating cooling system

951
REFERENCES
Sanner B., Karytsas C., Mendrinos D. and Rybach L. 2003. Current status of ground source heat pumps and
underground thermal energy storage in Europe. Geothermics 32(4–6):579–88.

Lund J.W. 2003.The USA geothermal country update. Geothermics 32(4–6):409–18.

Esen H., Inalli M. and Esen M. 2006. Technoeconomic appraisal of a ground source heat pump system for a
heating season in eastern Turkey. Energy Conversion Management 47(9–10):1281–97.

Ozgener L., Hepbasli A. and Dincer I. 2007. A key review on performance improvement aspects of geothermal
district heating systems and applications. Renew Sustain Energy Rev 11(8):1675–97.

Kavanaugh S.P. 1992. Field test of vertical ground-coupled heat pump in Alabama. ASHRAE Trans
98(2):607–16.

Kavanaugh S. P. 1998. Development of design tools for ground-source heat pump piping. ASHRAE Trans
104(1B):932–7.

Agustos H., Acikgoz O., Akbulut U. and Kincay O. 2008. Dikey tip toprak kaynaklı ısı pompası kullanımında
güneş enerji desteğinin araştırılması, TMMOB Makine Mühendisleri Odası Tesisat Mühendisliği Dergisi
106:47-53.

Akbulut U., Acikgoz O., Kincay O and Karakoc H.T. 2012. Performance assessment of a vertical ground-
source heat pump system with wall heating and cooling, Energy Education Science and Technology Part A:
Energy Science and Research Special Issue: 423-428.

Grandall A.C. 1946. House heating with earth heat pump. Electrical World 126(19): 94–5.

Ingersoll L.R. and Plass H.J. 1948. Theory of the ground pipe heat source for the heat pump. Heating, Piping &
Air Conditioning 20/7: 119-122.

Hepbasli A. 2005.Thermodynamic analysis of a ground-source heat pump system for districtheating. Int J
Energy Res 7:671–87.

Hepbasli A. 2002. Performance evaluation of a vertical ground source heat pump system in Izmir, Turkey. Int J
Energy Res 26:121-139.

Li X., Chen Z. and Zhao J. 2006. Simulation and experiment on the thermal performance of U-vertical ground
coupled heat exchanger. Appl Therm Eng 26(14–15):1564–71.

Esen H., Inalli M. and Esen M. 2007. Numerical and experimental analysis of a horizontal ground-coupled heat
pump system. Building Environment 42(3):1126–34.

Lund J.W., Freeston D.H. and Boyd T.L. 2005. Direct application of geothermal energy: 2005 worldwide review,
Geothermics 34:691–727.

Urchueguia J.F., Zacares M., Corberan J.M., Montero A., Martos J. and Witte H. 2008. Comparison between
the energy performance of a ground coupled water to water heat pump system and an air to water heat pump
system for heating and cooling in typical conditions of the European Mediterranean coast. Energy Conversion
Management 49:2917- 2923.

ASHRAE. 2004. ASHRAE Handbook-HVAC Systems and Equipment, Atlanta: American Society of Heating,
Refrigeration and Air-Conditioning Engineers Inc.

Yoru Y., Akbulut U., Kıncay, O. and Karakoc T.H. 2010. Dynamic Exergy Analysis of a Real Vertical Ground
Source Heat Pump System by Using Fast Artificial Neural Network Library (FANN).The Fifth International
Ege Energy Symposium and Exhibition, IEESE-5. June 27-30, Pamukkale University, Denizli, Türkiye.

952
Kincay O., Akbulut, U., Yoru, Y., Acikgoz, O. and Karakoc, H. 2010. Investigation of Usage of Ground
Source Heat Pump System on Wall Heating. AIVC International Conference, Low Energy and Sustainable
Ventilation Technologies for Green Buildings, October 26-28, Seoul, Korea.

Akbulut, U., Yoru, Y. and Kıncay, O. 2011. Annual Investigation of Ground Source Heat Pump System
Performance on a Wall Heating System in Istanbul. The 6th International Green Energy Conference, IGEC-
VI, Eskişehir, June 5-9, Anadolu University, Türkiye.

Başkal, A., Akbulut, U. and Kıncay, O. 2012. Toprak Kaynaklı Isı Pompası Destekli Duvardan Isıtma Sistemi.
TMMOB Makina Mühendisleri Odası, Tesisat Mühendisliği Dergisi 128: 43-50. (in Turkish)

Moran MJ. 1989. Availability analysis—a guide to efficient energy use. New York: ASME Press.

Kotas TJ. 1995. The exergy method of thermal plant analysis. Melbourne FL: Krieger.

Bejan A., Tsatsaronis G. and Moran M. 1996.Thermal design & optimization. New York: Wiley.

Rosen M.A. 1999. Second law analysis: approaches and implications. Int J Energy Res 23(5): 415–29.

Dincer I. and Cengel Y.A. 2001. Energy, entropy and exergy concepts and their roles in thermal engineering.
Entropy 3(3):116–49.

Bejan A. 2002. Fundamentals of exergy analysis, entropy generation minimization, and the generation of flow
architecture. Int J Energy Res 26(7):545–65.

Yildiz A. and Güngör A. 2009. Energy and exergy analyses of space heating in buildings. Appl Energy 86: 1939
–1948

Kıncay, O., Akbulut, U. and Temir, G. 2004. İstanbul’da Bulunan Bir İş Merkezindeki Örnek Bir Katın Soğutma
Sezonunda Ekserji Analizi, Tesisat Dergisi, 108: 68-73 (in Turkish)

Hepbasli A. and Akdemir O. 2004. Energy and exergy analysis of a ground source (geothermal) heat pump
system. Energy Conv & Management 45:737–753.

Utlu Z. and Hepbasli A. 2007. Parametrical investigation of the effect of dead (reference) state on energy and
exergy utilization efficiencies of residential–commercial sectors: A review and an application. Renewable and
Sustainable Energy Reviews 11: 603–634.

Ozgener O. and Hepbasli A. 2007. A parametrical study on the energetic and exergetic assessment of a solar-
assisted vertical ground-source heat pump system used for heating a greenhouse. Build Environ 42(1):11–24.

Yuehong Bi, Xinhong Wang, Yun Liu, Hua Zhang and Lingen Chen. 2009. Comprehensive exergy analysis of a
ground-source heat pump system for both building heating and cooling modes. Appl Energy 86:2560-2565.

Madani H., Ahmadi N., Claesson J. and Lundqvist P. 2010. Experimental Analysis of a V ariable Capacity Heat
Pump System Focusing on the Compressor and Inverter Loss Behavior. International Refrigeration and Air
Conditioning Conference, Perdue University.

953
 THE RELATION BETWEEN THE EXERGETIC PERFORMANCE OF HEAT PUMP
SYSTEM AND EXERGETIC PERFORMANCE OF ITS COMPONENTS
1 2 3 2
U. Cakir , K. Comakli A. C. Kus and O. Comakli
1
Engineering Faculty, Mechanical Eng. Dep., Bayburt University
Bayburt / TÜRKİYE
2
Engineering Faculty, Mechanical Eng. Dep., Atatürk University
Erzurum / TÜRKİYE
Engineering Faculty, Mechanical Eng. Dep., Iğdır University
3

Iğdır / TÜRKİYE
Email: ucakir@bayburt.edu.tr; kcomakli@atauni.edu.tr; ocomakli@atauni.edu.tr

ABSTRACT
Energetic and exergetic performance of heat pumps are affected by many parameters and factors. For
instance physical properties and thermal specifications of the heat source and heat sink are the most
impressive parameters for heat pumps’ performance. For that reason generally heat pumps are entitled
according to type of their heat source and sink like air to air or ground source heat pump. Additionally type
of the refrigerant is a very important factor for heat pumps. On the other hand just like a team work,
performance of every component of the system plays a big role on the energetic and exergetic performance
of heat pump. In this study the relation between the exergetic performance of the heat pump and its
components’ is investigated and presented experimentally. In addition the effects of the components’
working condition on the system performance is exposed and discussed.

INTRODUCTION
Energy savings and efficiency became most important issues for energy policies for most of the
governments over the world. Especially in European countries, using energy more efficiently and
decreasing the CO2 emissions are in the main targets of governments. In the light of these events most of
the countries develop new building regulations or change existing regulations which determines the
properties of energy using systems of buildings. For instance European Energy Performance of buildings
Directive (EPBD) can be thought [1, 2 and 3]. One of the systems that offer us an efficient energy using
possibility is heat pump system. Heat pump system can be used in a lot of different applications or sectors
like heating, cooling, residential or industrial buildings. For that reason improving the performance of heat
pump systems is very important to save the energy and to decrease the cost of energy. Energetic and
exergetic performance of heat pumps are affected by many parameters and factors. For example physical
properties and thermal specifications of the materials these are used by heat pump as heat source or heat
sink are the most impressive parameters for heat pumps’ performance. Additionally the type of the
refrigerant is a very important factor for heat pumps. On the other hand just like a team work, performance
of every component of the system plays a big role on the energetic and exergetic performance of heat
pump. There are some studies which were made on the relation between the performances of heat pump
components and heat pump system or on the effects of the performance of any heat pump component on
the performance of heat pump system.

Shuxue and Guoyuan made a research on air source heat pump coupled with economized vapor injection
scroll compressor and ejector. They presented the design method of the heat pump system with ejector
(EVIe) and summarized the process for designing the heat pump systems with ejector. They reached that
the reasonable value for the entrainment ratio of the ejector is between 0.1 and 0.2. One prototype heat
pump was designed under the condition of the evaporation temperature of _20 _C, and an experimental
setup was established to test the prototype. The measured results demonstrated that the heating EER of
the heat pump system with ejector could reach about 4% higher than that of the system without ejector
when the heating capacity remained nearly constant [4]. Another study was made by Schiffmann and Favrat
on experimental investigation of a direct driven radial compressor for domestic heat pumps which presents
the derivation of the specifications, the choice for an appropriate refrigerant fluid and the design of a proof
of concept prototype with the various tradeoffs between the impeller characteristics to follow the seasonal
heat demand, the bearing and rotor dynamics for a stable operation [5]. A heat pump system with a flash-
tank coupled with a scroll compressor is compared with a system with a sub-cooler by Buo-Yan and Hui-
Xia. They measured the heat pump performance experimentally. The heating capacity of the prototype
decreased as the evaporation temperature decreased, but the decrease was much slower than that of a
conventional air-source heat pump. The power input varied slightly with the evaporation temperature. The
heat pump system with a flash tank is more efficient than the system with a sub-cooler at low ambient
temperatures, so it will be very useful as a small capacity air-source heat pump [6]. Adhikari et. al. made a

954
study about that four different kind of HVAC systems based on heat pump technology at three site locations
(Milan, Rome and Palermo) were implemented and simulated in it. In particular, the study is focused on
comparative performance of single speed and variable speed compressor equipment based on heat pump
technology. The results demonstrated that variable speed equipment can achieve considerable energy
savings especially in the air source heat pump system applications [7]. In the paper which was made by
Schiffmann and Favrat design, experimental investigation and multi-objective optimization of a small-scale
radial compressor for heat pump applications are studied. The paper shows a possible way to deal with
that difficulty, by coupling an appropriate modeling tool to a multi objective optimizer. The optimizer
manages to fit the compressor design into the possible specifications field while keeping the high efficiency
on a wide operational range [8]. Ko et. al. Made a study on the selection of volume ratio of two-stage rotary
compressor and its effects on air-to-water heat pump with flash tank cycle. In this study two-stage rotary
compressor was developed for vapor injection cycle, the effect of volume ratio of a two-stage rotary
compressor on AWHP was investigated and experimental investigation was made for vapor injection air-to-
water heat pump. In addition to these; the volume ratio of two cylinder of a two-stage rotary compressor has
significant effect on the performance of the AWHP so that it was experimentally investigated. Based on this
result, a two-stage rotary compressor designed with an optimized volume ratio was manufactured and
incorporated into the AWHP system [9]. An analytical investigation study was made on investigation of high
temperature heat pump with screw compressor and screw expander for power recovery by Tamura et al..
An analytical investigation of a high temperature heat pump system was developed for estimating the
thermal cycle and assessing the thermal fluids for their high-temperature delivery (up to 180oC) capacities
without decomposition or the use of lubricant mechanisms. Then a screw type compressor was applied in
the above conditions. Furthermore a screw expander was also used as a replacement for the throttle valve.
Numerical results of systems were presented from cases where steam-water (for higher temperatures) and
ammonia (for lower temperatures) were used [10]. Defrost cycle performance is investigated by Payne and
O'Neal for an air-source heat pump with a scroll and a reciprocating compressor. In this study A 10.6 kW
nominal cooling capacity air-source heat pump was tested according to ANSI/ASHRAE Standard 116-1983
for the frost accumulation and defrost cycle. These tests required indoor conditions of 21.1 °C (70 °F) dry-
bulb, 15 °C (60 °F) maximum wet-bulb, with outdoor conditions of 1.7 °C (35 °F) dry-bulb, 0.5~C (30°F) wet-
bulb. The unit was tested with the original scroll compressor and a reciprocating compressor that yielded
similar heating performance. Heating capacity for the scroll system peaked at 8.4 kW (2.38 tons), while the
reciprocating system heating capacity peaked at 8.5 kW (2.42 tons) during the frosting period. Power
demand for the scroll system peaked at 2.9 kW, and the reciprocating system power demand peaked at 3.1
kW. During the frosting period, the reciprocating system power demand averaged 7% higher than the scroll
system power demand. The reciprocating system completed a defrost 5% faster than the scroll system [11].
The design optimization of a heat exchanger was discussed, using a thermo economic approach in the
study which was made by Dentice d’Accadiaa, and Vanolib. The investigation was referred to the tube-in-
tube condenser of a conventional vapor compression heat pump, with a two phase refrigerant flowing in the
inner tube and a single-phase fluid flowing in the annulus. A cost function to be minimized is introduced,
defined as the sum of two contributions: the amortization cost of the condenser, related to the heat
exchange area and the operating cost of the electric-driven heat pump in which the heat exchanger will
work, depending on the overall exergy destruction rate in the system. In the paper, this latter contribution is
related to the local irreversibility rate in the condenser, using the so-called structural approach. The optimal
trade-off between amortization and operating cost was therefore investigated, by minimizing the above-
mentioned cost function. A numerical example was discussed, in which, for a commercial heat exchanger,
the design improvements needed to obtain a cost-optimal configuration are investigated [12]. Granryd made
a study on analytical expressions for optimum flow rates in evaporators and condensers of heat pumping
systems. Considering a base case of a refrigerating system where the design data for the evaporator and
condenser had been known, simple analytical relations were deduced for optimum flow rates that would
result in highest overall COP of the system when energy demand for the compressor as well as fans (or
pumps) were included. It is reported in the study that the expressions could be used for design purposes
but also for checking flow velocities used in an existing plant. The relations may also be incorporated in
algorithms for optimal operation of systems with variable speed compressors. Finally a simple general
relation was given for the “pump power ratio” for forced convection heat exchangers defined as the ratio
between the necessary power for pumps or fans and the rate of heat transferred [13]. The defrosting of a
heat-pump evaporator with warm air was investigated experimentally under controlled conditions in an air-
conditioned wind tunnel by Haukås. The work was initiated to explore the possibility of defrosting an
evaporator under conditions where a heat pump continues to perform heating duty. A four-row straight-fin
heat exchanger was used, supported from below by a load transducer to measure its weight variation
during both frost accumulation and removal. The rate of melting was determined as a function of air
temperature, relative humidity and velocity. The removal from the evaporator surfaces of water produced by
defrosting was found to be dependent on air velocity [14]. In a study the performance of a novel dry-
expansion evaporator developed for a wastewater source heat pump was experimentally tested and the test

955
 results were reported. An experimental wastewater source heat pump containing the novel evaporator using
waste bath water as heat source for water heating was built up. A one-month long test to examine the effect
of tube bio-fouling on the operating performance of the WWSHP was conducted. The comparison results
demonstrated clearly that the novel evaporator with defouling function was effective in easily removing the
bio-fouling build-up at a low cost [15].

In this study the relation between the exergetic performance of the heat pump and its components’
performances is investigated and presented experimentally. In addition the effects of the components’
working condition on the system performance is exposed and discussed.

Various studies were made on criticizing the factors that affect the performance of heat pumps. Most of
them were made on only the performance of any component of system. For instance some of them deal
with using different kind of compressors or deals with the compressor efficiency with cooling or not. In some
studies, energetic performance of using different kinds of heat exchangers as evaporator or condenser was
studied. In this study firstly exergetic treatments of each component (compressor, condenser, exp. valve
and evaporator) of a classical heat pump system was determined for different running conditions. Then this
analysis was made for heat pump system. In this way the relation between the exergetic performance of the
heat pump and its components’ is investigated and presented experimentally. In addition the effects of the
components’ working condition on the system performance is exposed and discussed.

CASE STUDY AND CALCULATIONS

The schematic diagram of the experimental system can be seen on Figure 1. The system was originally
designed for operating with R22 and it is an air to water heat pump. . The main components of the system
are a scroll compressor, an air cooled evaporator, a water cooled condenser, thermostatic expansion valve,
and the other elements like measurement and control equipment. In addition electrical air and water heaters
are used in order to keep the temperature of air and water passing into the evaporator and condenser at the
desired levels. One electrical fan is used to circulate the air on the evaporator. First, the system was
charged with 15 Bar R22 while all the valves were in opened position under the temperature of outdoor air.
Temperature and pressure values in the key-point of the plant, as shown in Fig. 1, were continuously
monitored in order to check the achievement of steady state conditions. Usually, the start-up time required
about 1 hour. After each experimental running, the raw data which consist of temperatures from the
thermocouples, pressures from the manometers, air flow rates from the anemometer, and refrigerant flow
rate from the flow meters, water flow rates from the rotameters, compressor input current and voltage from
the amper meter were recorded. The tests were performed under laboratory conditions, during the air
temperature was about 20oC-22oC, and the relative humidity was about 40%-60%. The numerical values of
the evaporator air temperature and flow rate are presented in Table 1 for two heat pump tests. All of the
measured values which consist of temperatures from the thermocouples, pressures from the manometers,
compressor input power from the wattmeter, and flow rates of the fluids used on the system from the flow
meters, rotameters and anemometers were used for determining the thermodynamics performances of
components and heat pump to achieve the aim mentioned previously.

Table 1. Levels of parameters

Parameter Fluid Level 1 Level 2 Level 3 Level 4 Level 5 Level 6
Mass flow rate (kg/s) Air 0.41 0.71 0.83 1.1 1.2 --
Water 0.100 0.110 0.125 0.139 0.150 --
Temperature (0C) Air 20 22 26 28 30 32

It is needed to make energy analysis to make a comprehensive exergy analysis. Mass and energy balances
are employed to determine the heat input, and energy efficiencies. From the measured parameters; the
heat delivered by the condenser to the water and the heat extracted by the evaporator from the air and
power input to the compressor are calculated as seen below;
=Q cond m wCw (Tw out − Tw in ) (1)

=Q evap m air Cair (Tair in − Tair out ) (2)

Wcomp elec = 3Cos (φ ) U I W (3)

956
Where; w is water C is specific heat of materials; U, I and Cos (θ) are voltage (V), current (A) and power
factor, respectively. Total exergy of a fluid, specific exergy of water, refrigerant and air are calculated as
seen below [16, 17 and 18]
 = m ex
Ex (4)
 
ex
air
= C ( air
+ ω C vapor T  T
0

) 
T 
0

-1- ln  T

T  (
0
air 0


P

+ 1+1.6078 ω ) R T ln  P  π
0 (5)
  

+R

air 0 


 0  (
T (1+1.6078ω ) ln  1+1.6078ω / (1+1.6078ω )  +1.6078 ω ln  ω

ω 
0 

)
exw =( hw − hw 0 ) − T0 ( sw − sw 0 ) (6)

P T P T
HOT GAS
BY-PAS
Oil Seperator

COMPRESSOR

Drayer
T
Water cooled
Condenser

SOL.
T
P T VALF
RESIVER Drayer

Sight Glass Exp.

Flowmeter
Valf

P Pressure Measuring
Water
storage T Temperature Measuring
Tank
Valve

Elec.
Heater Air cooled evaporator
Evaporator
T T
T

AIR

Fan
Condense (4750 m3/h)
(Speed adjustable)

Figure 1. Experimental system

exr =( hr − hr 0 ) − T0 ( sr − sr 0 ) (7)

Exergy efficiency of the components can be calculated as shown follow;

m w ( exw in − exw out )
ηex cond =
m r ( exr in − exr out )
(8)

m r ( exr in − exr out )

ηex evap =
m air ( exr air − exair out )
(9)

957
Exergy destruction rate of the compressor, condenser and evaporator can be calculated as seen below:
 
dest comp = mr (exr in − exr out ) + Wcomp
Ex  (10)

dest cond = mr (exr in − exr out ) + mw (exw in − exw out )
Ex   (11)

dest evap = mr (exr in − exr out ) + mair (exair in − exair out )
Ex   (12)

RESULTS
This study analyzes the effect of temperature and mass flow rate change on the exergetic performance of
heat pump unit, heat pump system, and components of heat pump unit. Heat pump unit means the basic
heat pump which consists of condenser, compressor, expansion valve and evaporator. Heat pump system
represents the whole system with other auxiliary equipment. Exergy efficiency and exergy destruction is
evaluated and presented in scope of exergetic performance.
Exergy analysis of compressor
Change of compressor exergy efficiency depending on evaporator external fluid temperature is presented in
Figure 2. Additionally change of exergy destruction rate of compressor versus change of air temperature is
shown on Figure 3. As seen on the Figure 1, when mass flow rate of air is minimum (0.41 kg/s), exergy
efficiency of compressor becomes maximum. There is not a stable increase or decrease on exergy
efficiency of compressor when temperature of air increases but according to mean and average values
exergy efficiency of compressor increases when air temperature increases. Specific volume and
superheating rate of refrigerant increase with the increasing of air temperature. Depend on this event,
refrigerant temperature reaches higher values at exit of compressor. When the refrigerant goes on next
cycle, refrigerant arrives to the evaporator with higher temperature and as a result of this it takes less heat
energy from the air and reaches with higher temperature so the work of compressor affected by this event.

mw = 0.15 kg/s - Tw= 20 oC

62
60
ηex (%)

58
56
mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
54
mh=0.71 kg/sn mh=0.41 kg/sn
52
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 2. Change of compressor exergy efficiency versus air temperature

mw = 0.15 kg/s - Tw= 20 oC

1,1

1,0
Exdest (kW)

0,9
mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
mh=0.71 kg/sn mh=0.41 kg/sn
0,8
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 3. Change of compressor relative irreversibility versus air temperature change

958
Figure 4 and 5 represent the relative irreversibility change of compressor on heat pump unit and heat pump
system according to the air temperature for different mass flow rates of air. As seen on the change of
compressor relative irreversibility is more stable than exergy efficiency. Ranking of the compressor
irreversibility rates of compressor with respect to mass flow rate is as follows; 0.41 kg/s , 0.71kg/s, 0.83
kg/s, 1.1. kg/s and 1.2 kg/s. As seen relative irreversibility of compressor increases when mass flow rate of
air decreases. In addition to this; relative irreversibility of compressor decreases with the increasing of air
temperature.

mw = 0.15 kg/s - Tw= 20 oC

45
40
IRUnit (%)

35
30
25 mh=1.2 kg/sn mh=1.1 kg/sn
20
18 22 20 24 26 28 30 32
Air Temperature (oC)
Fig. 4. Change of relative irreversibility of compressor on heat pump unit according to the change of air
temperature

mw = 0.15 kg/s - Tw= 20 oC

50
45
40
IRSystem (%)

35
30
25 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
mh=0.71 kg/sn mh=0.41 kg/sn
20
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 5. Change of relative irreversibility of compressor on heat pump system according to the change of air
temperature

Exergy analysis of condenser

Figure 6 shows the change of condenser exergy efficiency and Figure 7 presents the exergy destruction
rates depending on the air temperature increase for different mass flow rates of air. As seen on it change of
condenser exergy efficiency is not stable just like compressor’. For that reason it is difficult to comment this
event and to get any conclusion. We may make an explanation by the help of exergy and exergy
destruction. The exergy destruction and exergy efficiency in any component of the heat pump is not caused
only by the inefficiencies of it. That means the exergy destruction of a component depends on the
inefficiencies of other system components. We can clearly see the change of exergy destruction rates of the
systems on the figures which were formed by using the results of tests but the contributions of the
components to the total exergy destruction of the system must be taken into account to comment these
figures.

959
 mw = 0.15 kg/s - Tw= 20 oC
16
14
12
10
ηex (%)

8
6
4 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
mh=0.71 kg/sn mh=0.41 kg/sn
2
18 20 22 24 26 28 30 32
Evaporatör Hava Sıcaklığı (oC)
Fig. 6. Change of condenser exergy efficiency according to the change of air temperature

mw = 0.15 kg/s - Tw= 20 oC

0,80 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
mh=0.71 kg/sn mh=0.41 kg/sn
Exdest (kW)

0,75

0,70
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 7. Change of amount of condenser exergy destruction according to the change of air temperature

Change of relative irreversibility of condenser on heat pump unit and heat pump system is shown by the
help of Figure 8 and Figure 9. According to the mean values; approximately 20%-35% percentage of the
whole irreversibility that exists on the heat pump unit comes out from condenser. Relative irreversibility of
condenser increases with the increasing of air mass flow rate and it decreases with the increasing of air
temperature. Because of that there is no use of pump to circulate the water which is used for cooling the
condenser, relative irreversibility rates of components on heat pump unit and heat pump system are very
closed to each other. There is only a fan is used to circulate the air and the relative irreversibility of fan on
whole heat pump system is between 1-1.5%. It is understood from the figures that nearly 70% of exergy
that enters to the condenser is saved and the other part cannot be saved by the system.

mw = 015 kg/s - Tw= 20 oC

40
35
30
IRUnit (%)

25
20
15 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
mh=0.71 kg/sn mh=0.41 kg/sn
10
18 20 22 24 26 28 30 32
Air Temperature (oC) Figure
Fig. 8. Change of relative irreversibility of condenser on heat pump unit according to the change of air
temperature

960
 mw = 0.15 kg/s - Tw= 20 oC
35

30
IRSystem (%)

25

20

15 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn

mh=0.71 kg/sn mh=0.41 kg/sn
10
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 9. Change of relative irreversibility of condenser on heat pump system according to the change of air
temperature

Exergy analysis of expansion valve

However expansion valve is the smallest component of the heat pump systems, it has a big and important
mission in a refrigeration/heat pump cycle. Exergetic performance of expansion valve has not a big effect
on the performance of heat pump directly but it has a serious effect on heat pump performance indirectly
because of that the thermal performance of an expansion valve is very important for the performances of
evaporator and compressor. When Figure 10 and 11 is analyzed carefully we understand that exergetic
efficiency of expansion valve is very high (between 0.97-0.99) and exergy destruction rates of it is very low
(between 0.025-0.045 kW). As it is understood easily, these values are very low values when they are
compared with the values presented above for compressor and condenser. Figure 12 and 13 are present
show us the change of relative irreversibility of expansion valve on heat pump unit and heat pump system.
They support our conclusion that we reached here.

mw = 0.15 kg/s - Tw= 20 oC

100
99
ηex (%)

98
97
mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
96
mh=0.71 kg/sn mh=0.41 kg/sn
95
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 10. Change of expansion valve exergy efficiency according to the change of air temperature

mw = 0.15 kg/s - Tw= 20 oC

0,05
0,04
Exdesdt (kW)

0,03
0,02
0,01 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
mh=0.71 kg/sn mh=0.41 kg/sn
0,00
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 11. Change of expansion valve exergy destruction according to the change of air temperature

961
 mw = 0.15 kg/s - Tw= 20 oC
2,0

1,5
IRUnıt (%)

1,0

0,5 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn

mh=0.71 kg/sn mh=0.41 kg/sn
0,0
18 20 22 24 26 28 30 32

Air Temperature (oC)

Fig. 12. Change of relative irreversibility of expansion valve on heat pump unit according to the change of
air temperature

mw = 0.15 kg/s - Tw= 20 oC

2,0

1,5
IRSystem (%)

1,0

0,5 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn

mh=0.71 kg/sn mh=0.41 kg/sn
0,0
18 20 22 24 26 28 30 32

Air temperature (oC)

Fig. 13. Change of relative irreversibility of expansion valve on heat pump system according to the change
of air temperature

Exergy analysis of evaporator

One of the most important parameter which has an important effect on the performance of heat pump is the
heat energy which is gained from the heat source (it is air in here). This event is interrelated with thermal
specifications and properties of the material used as heat source. For these reasons temperature and mass
flow rate of evaporator air are selected a main parameter and we get the related evaluations according to
these parameters. Figure 14 and 15 are present the change of exergetic performance of evaporator with
respect to change of air temperature. Unlike the previous results; the results which are obtained for
evaporator are rather stable. This is a normal event because of that increasing of air temperature and air
mass flow rate affects the evaporator performance firstly and directly while other components are affected
indirectly. As seen on the Figure 14 when air temperature and air mass flow rate increase, exergy efficiency
of evaporator decreases smoothly. Both of them have negative effects on the performance of evaporator.
Exergy destruction rate of evaporator increases with the increasing of air temperature and with the
increasing of air mass flow rate.

962
 mw = 0.15 kg/s - Tw= 20 oC
60
50
ηex (%)

40
30
20 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
mh=0.71 kg/sn mh=0.41 kg/sn
10
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 14. Change of evaporator exergy efficiency according to the change of air temperature

1,8 mw = 0.15 kg/s - Tw= 20 oC

1,5
1,2
Exdest (kW)

0,9
0,6
0,3
mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn
0,0
18 20 22 24 26 28 30 32
Air Temperature (oC)
Fig. 15. Change of evaporator exergy destruction according to the change of air temperature

Change of related irreversibility of evaporator on heat pump unit and heat pump system depending on the
change of air temperature and air mass flow rate are visualized by the help of Figure 16 and 17. As seen on
the figures related irreversibility of evaporator increases when the temperature and mass flow rate of air
increases. And the curves are stable and it changes smoothly.

msu = 0,15 kg/sn - Tsu= 20 oC

50

40
IRÜnite (%)

30

20

10 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn

mh=0.71 kg/sn mh=0.41 kg/sn
0
18 20 22 24 26 28 30 32
Air Temperature (oC)

Fig. 16. Change of relative irreversibility of evaporator on heat pump unit according to the change of air
temperature

963
 msu = 0,15 kg/sn - Tsu= 20 oC
50

40
IRSistem (%)

30

20

10 mh=1.2 kg/sn mh=1.1 kg/sn mh=0.83 kg/sn

mh=0.71 kg/sn mh=0.41 kg/sn
0
18 20 22 24 26 28 30 32
Air Temperature (oC)

Fig. 17. Change of relative irreversibility of evaporator on heat pump system according to the change of air
temperature

Exergy analysis of heat pump unit

Figure 18 and Figure 19 show the exergetic performance (exergy efficiency) of heat pump unit according to
temperature increase and mass flow rate increase. As seen on the figure 18 when the temperature of air
increases from 20oC to 32oC (by 12oC) exergy efficiency of the unit increases from 24.31% to 24.92% first
and becomes nearly constant a little. At the end of the second increase, it reaches to the efficiency of 25%.
We can say that temperature increasing of air which is used as heat source has a positive effect on the
performance of heat pump. As it is understood from the Figure 19, increase of air mass flow rate has a
positive effect on the unit performance like temperature.

Figure 20 and 21 show us the change of exergy destruction that occurs in the heat pump unit with respect
to the increase in evaporator air temperature and mass flow rate. As seen on the figures when temperature
or mass flow rate of air increase, rate of heat pump exergy destruction increases parallel. That is not an
abnormal event. Increasing of the temperature or mass flow rate of air causes the heat pump to get more
heat energy and more exergy in, so that amount of exergy that is destructed in the system can increase.
But this situation does not mean that, exergy efficiency os the system will decrease.

25,5

25,3

25,0
ηex (%)

24,8

24,5

24,3

24,0
0 4 2 6 8 10 12
Increasing rate of air temperature (oC)
Fig. 18. Change of exergy efficiency of heat pump unit according to the change of air temperature

964
 25,2

24,9

24,6
ηex (%)

24,3

24,0
0 1 2 3 4 5
Levels of mass flow rate

Fig. 19. Change of exergy efficiency of heat pump unit according to the change of air mass flow rate

3,0

2,8
Exdest (kW)

2,6

2,4

2,2

2,0
0 2 4 6 8 10 12 14
Increasing rate of air temperature (oC)
Fig. 20. Change of heat pump unit exergy destruction according to the change of air temperature

3,4
3,2
3,0
Exdest (kW)

2,8
2,6
2,4
2,2
2,0
0 1 2 3 4 5 6

Levels of mass flow rate Fig. 21. Change of

heat pump unit exergy destruction according to the change of air mass flow rate

CONCLUSION
In this paper, exergy analysis of four main components of heat pump (compressor, condenser, expansion
valve and evaporator) has been made; regarding evaporator fluid temperature and mass flow rate which is
used as heat source. Then the same analysis is made for heat pump unit and heat pump system to expose

965
the relation between the performances of components and heat pump. Exergetic treatment of components
and system primarily depends on the temperature and mass flow rates of the heat source.
Expansion valve is the component which has the maximum exergy efficiency for all tests. Second
component according to the exergy efficiency is compressor. Exergy efficiency of the condenser becomes
lowest for all tests. Evaporator’s exergy efficiency value is between the others’ but closer to the exergy
efficiency of the compressor with an average of 44.25%. This table says us condenser is the first
component to be improved for the exergy efficiency and to overcome the irreversibility.

Exergy efficiency of heat pump unit, compressor and condenser increases while exergy efficiency of the
evaporator and expansion valve decreases when temperature of air increases. That means exergetic
performance of condenser and compressor have more important role on the system performance when we
made a comparison according to their exergy efficiency. But if we made this comparison according to
exergy destruction rates of components, evaporator becomes most important component. This table says
us condenser is the first component to be improved for the exergy efficiency and evaporator is first to
overcome the irreversibility.

REFERENCES
[1] EPBD-Directive 2002/91/EC of the European Parliament and of the Council of 16 December 2002
on the Energy Performance of Buildings. Official Journal of the European Union.
[2] EPBD recast- Directive 2010/31/EU of the European Parliament and of the Council of 19 May 2010
on the energy performance of buildings (recast). Official Journal of the European Union,
18/06/2010.
[3] Energy Savings through Variable Speed Compressor Heat Pump Systems R.S.Adhikari, N.Aste*,
M.Manfren and D.Marini Energy Procedia 14 (2012) 1337 – 1342
[4] Xu Shuxue, Ma Guoyuan Research on air-source heat pump coupled with economized vapor
injection scroll compressor and ejector International Journal of refrigerations 34 (2011)158-595
[5] J. Schiffmanna, D. Favratb. Experimental investigation of a direct driven radial compressor for
domestic heat pumps. International Journal of refrigerations 32 (2009)1918-1928
[6] Ma Guo-Yuan, Zhao Hui-Xia. Experimental study of a heat pump system with, flash-tank coupled
with scroll compressor. Energy and Buildings 40 (2008) 697–701
[7] R.S.Adhikari, N.Aste*, M.Manfren and D.Marini. Energy Savings through Variable Speed
Compressor Heat Pump Systems. Energy Procedia 14 (2012) 1337 – 1342 Conference on
Advances in Energy Engineering (ICAEE). 2nd International Conference on Advances in Energy
Engineering 2011
[8] J. Schiffmann, D. Favrat. Design, experimental investigation and multi-objective optimization of a
small-scale radial compressor for heat pump applications Energy 35 (2010) 436–450
[9] Younghwan Ko. Sangkyoung Park. Simon Jin. Byungsoon Kim. Ji Hwan Jeong. The selection of
volume ratio of two-stage rotary compressor and its effects on air-to-water heat pump with flash
tank cycle Applied Energy 104 (2013) 187–196
[10] Tamura I., Taniguchi H., Sasaki H., Yoshida R., Sekiguchi I., Yokogawa M., An analytical
investigation pf High-temperature heat pump with a screw compressor and screw expander for
power recovery. Energy conversion and management Vol. 38 NO:10-13 pp:1007-1013, 1997
[11] Payne, V.; O'Neal, D. L. Defrost cycle performance for an air-source heat pump with a scroll and a
reciprocating compressor. International Journal of Refrigeration vol. 18 issue 2 February, 1995. p.
107-112
[12] Massimo Dentice. d’Accadiaa. Laura Vanolib. International Journal of Refrigeration 27 (2004) 433–
441 Thermoeconomic optimisation of the condenser in a vapour compression heat pump
[13] E. Granryd* Analytical expressions for optimum flow rates in evaporators and condensers of heat
pumping systems International Journal of refrigeration 33 (2010)1211-1220
[14] H.T Haukås. Design of a plate-type evaporator for heat pumps International Journal of Refrigeration
Vol: 7, Issue:1, Pages 59-63, 1984
[15] Shen Chao a, Jiang Yiqiang a,⇑, Yao Yang a, Deng Shiming. Experimental performance evaluation
of a novel dry-expansion evaporator wit defouling function in a wastewater source heat pump
Applied Energy 95 (2012) 202–209

966
 MODEL BASED PERFORMANCE ANALYSIS OF A CONCENTRATING PARABOLIC
TROUGH COLLECTOR ARRAY
İ. H. Yılmaz, M.S.Söylemez, H. Hayta and R. Yumrutaş

Department of Mechanical Engineering

University of Gaziantep, 27310 Gaziantep, Turkey
iyilmaz@gantep.edu.tr, sait@gantep.edu.tr, hakanhyt@hotmail.com, yumrutas@gantep.edu.tr

ABSTRACT
This study presents a comprehensive thermo-mathematical model analysis of a parabolic trough solar collector
(PTSC) array. It is to be set-up in Gaziantep for an experimental study in order to perform its thermal efficiency
with respect to the different system parameters. The PTSC arrangement consists of three modules connected in
series. A detailed model for the absorbed solar energy falling on the aperture area of the PTSC array, the optical
efficiency changing with incidence angle and the useful energy gained by heat transfer fluid (HTF) for a linear with
a single-axis tracking motion rotating about a horizontal north-south orientation was programmed in Engineering
Equation Solver (EES) based on the actual system parameters. The model comprises steady-state and one-
dimensional heat transfer approach using the thermodynamics and the heat transfer relations to evaluate the
thermal losses of the receiver (heat collector element) by taking into account the effects of collector dimensions,
material properties, and fluid properties. In the performance analysis of the PTSC array, the effects of hourly solar
radiation flux, ambient conditions, collector inlet temperature and mass flow-rate of the working fluid were
investigated. Typical operating conditions on the 1th day of July at 12:00 solar time exhibited that when the HTF
mass flow-rate 0.3 kg/s, inlet temperature 150°C, ambient temperature 39°C, ambient air velocity 1 m/s and having
a 10.26 m2 PTSC aperture area and 940 W/m2 direct beam radiation incident with 13.98°, the estimated collector
array efficiency is about 59.2%. The model predictions are to be confirmed by the operation of PTSC being
installed at Gaziantep (Latitude: N37°02', Longitude: E37°19', Altitude: 881 m).

Keywords: Parabolic trough collector, thermo-mathematical modeling, performance analysis

INTRODUCTION
Concentrating collectors are widely used for many applications to supply thermal energy at temperatures higher
than those nonimaging collectors. Not only this feature but also having a relatively high thermal efficiency put them
forward in usage of compact thermal systems. A typical PTSC consists of a reflector material shaped in parabola,
a linear receiver to collect solar energy and a tracking unit to configure the position of the reflector. The linear
receiver is commonly composed of an absorber tube and its cover envelope. A glass cover tube is usually placed
around the receiver tube to reduce the increasing convective heat loss from the receiver at higher temperatures.
The receiver follows the sun path by a tracking unit on the mode of single or double-axis during its operation while
concentrating direct normal irradiance to the receiver locating at the focal line of the parabolic reflector. The
concentrated radiation reaching the receiver tube heats the fluid that circulates through it. The resulting thermal
energy is transferred continuously to the fluid as useful energy. In order to calculate the energy absorbed by the
working fluid, thermal and optical analyses are needed. These analyses are essential to reach the thermal and
optical losses from the receiver element and correspondingly to the thermal performance of the PTSC. The
analyses being made serve as a thermo-mathematical modeling of the PTSC while projecting the system
characteristics.

Many researchers presented studies of energy models of parabolic trough collectors. When the recent studies
conducted on detailed modeling of PTSC are handled, the most noticeable one was established by Forristall
(2003) who developed both a 1-D and a 2-D heat transfer model of parabolic trough solar receiver implemented in
EES. His model included input variables as collector geometry, optical properties, HTF properties, HTF inlet
temperature, flow rate, solar insolation, wind speed and ambient temperature. Corresponding model outputs were
investigated to determine collector efficiency, outlet HTF temperature, heat gain, and heat and optical losses of
parabolic trough collector. The other study is of Jacobson et al. (2006) who developed a solar parabolic trough
simulation to determine the optimum parameters to sustain a solar thermal power plant system in Thailand.
Another study was applied with a comprehensive model for the tubular receiver of the PTSC to improve the PTSC
design and overall system performance (Qu et al., 2010). The model was verified by the experimental data from
the tests on the PTSC in a high temperature solar cooling and heating system.

Other various groups’ published models on PTCs were followed, Padilla et al., (2011) performed a 1-D numerical
heat transfer analysis, and an optical analysis for the PTC to optimize and understand its performance under
different operating conditions. To validate the numerical results, the model was compared with experimental data

967
obtained from Sandia National Laboratory and other 1-D heat transfer models demonstrating a better agreement
with experimental data compared to the others. Huang et al. (2012) proposed a new analytical model for optical
performance to simulate the performance of a parabolic trough solar collector. Kalogirou (2012) presented a
detailed thermal model of a parabolic trough collector considering all modes of heat transfer. The model was
written in EES and validated with known performance of existing collectors to be installed at the Cyprus University
of Technology.

In this study, a comprehensive thermo-mathematical model analysis of a PTSC array is presented. Solar energy
calculations for determining the optical efficiency and the useful heat gain by the HTF are conducted for obtaining
the thermal performance of the PTSC array. The heat losses from the receiver are evaluated considering steady-
state and one-dimensional heat transfer approach. In the performance analysis of the PTSC array, the effects of
hourly solar radiation flux, ambient conditions, collector inlet temperature and mass flow-rate of the working fluid
are investigated.

PARABOLIC TROUGH SOLAR COLLECTOR

System Description
The PTSC array consists of three modules connected in series shown in Fig. 1. It is supplied from the German
company Smirro and has the technical specifications as indicated in Table 1. Its linear receiver is considered to
have oriented coincidentally with the north-south axis of the local and a HTF which is circulated within it by means
of a frequency-controlled gear pump. The resulting thermal energy is transferred continuously to the HTF whose
typical technical data given in Table 2 (http://www.fuchs-oil.de).

Fig.1. The PTSC array

Table 1. Characteristics of a single PTSC module

Description Value
Parabola length, 3000 mm
Parabola aperture, 1140 mm
Focal distance, 35 mm
Rim angle, 78°
Outside absorber diameter, 35 mm
Inside absorber diameter, 32 mm
Outside cover diameter, 60 mm
Inside cover diameter, 54 mm
Mirror reflectance, 0.90
Absorber material Stainless steel
Cover material DURAN glass
Mirror material Aluminum

Table 2. Technical specifications of Fuchs Renolin therm 320

Temperature Density Specific heat Thermal Kinematic Prandtl
capacity conductivity viscosity number
°C kg/m3 kj/kg·°C W/m·°C m2/s ×10-6 –
0 879 1.864 0.134 535 6543

968
 50 848 2.078 0.131 28.6 385
100 816 2.293 0.127 6.5 96
200 750 2.721 0.120 1.5 26
300 685 3.151 0.113 0.7 13.4
320 672 3.236 0.111 0.6 11.8

Solar Energy Calculations

The emitted radiation from the sun has a directional characteristic that is defined by a set of geometric relations
determining the angle of incidence of the radiation on a surface. Due to the instant change in the position of the
sun, the incidence angle of the collector surface is to be calculated depending on solar time. Thus, the geometric
relations between a plane moving relative to the earth and the beam radiation are to be determined during solar
energy calculations. These calculations are considered as the PTSC array installed at Gaziantep where the local
geographic position is latitude N37°02', longitude E37°19' and the site elevation 881 m above sea level.

Theoretical upper limit of solar radiation available at the earth surfaces is called as the extraterrestrial radiation.
The radiation emitted by the sun and its spatial relationship to the earth causes a nearly fixed intensity of solar
radiation external side of the earth’s atmosphere. This intensity is represented by the solar constant, which is
the energy received per unit time from the sun on a unit area of surface perpendicular to the direction of spreading
of radiation at mean earth sun distance. The value of of 1367 W/m2 is adapted from the literature having a
minimum uncertainty of the order of 1%. Accordingly, the extraterrestrial radiation depending on time of year is
given by

(1)

where is the extraterrestrial radiation incident on a plane normal to the radiation on the nth day of the year.

Solar time does not coincide with local clock time and it is used in all of the sun-angle relationships so it is
necessary to convert standard time to solar time. The general equation for calculating the apparent solar time is

(2)

where is the standard meridian for the local time zone, is the longitude of the location interested. The
values of the equation of time, E as a function of the day of the year can be obtained approximately from the
following equations:

(3)

where

The declination angle, denoted by , varies seasonally due to the tilt of the Earth on its axis of rotation at an angle
of 23.45° with the ecliptic axis. This requires a coordinate transformation i.e. the sun moves around the earth. The
declination is zero at the equinoxes (March 22 and September 22), positive during the northern hemisphere
summer and negative during the northern hemisphere winter. The declination reaches a maximum of 23.45° on
June 22 (summer solstice in the northern hemisphere) and a minimum of -23.45° on December 22 (winter solstice
in the northern hemisphere). The declined function of the day of the year can be defined as

(4)

Hour angle, determines angular displacement of the sun east or west of the local meridian due to rotation of the
earth on its axis at 15 per hour and takes negative values before solar noon and positive values after solar noon.

(5)

Zenith angle is the angle between the beam radiation and normal of earth surface. Thus, for the horizontal
surfaces, the angle of incidence becomes the zenith angle of the sun, . For this situation, the zenith angle is
defined as

(6)

969
The solar radiation received from the sun falling on a horizontal surface depends on the incidence angle. Solar
tracking systems are intended to perform a PTSC while lessening the incidence angle in order to increase the
beam component of the solar radiation falling on its aperture area. In addition that the orientation of the collector
system with respect to local position is significant from the view of following the sun path as close as normal
incidence to be able to harvest much more solar energy. For a plane rotated about a horizontal north-south axis as
shown in Fig. 2 with continuous adjustment from east to west to minimize the angle of incidence estimated by the
relation (Duffie and Beckman, 2005).
(7)

Fig.2. Orientation of the PTSC array

Optical Efficiency Analysis of the Collector Array

Optical efficiency is defined as the ratio of the energy absorbed by the receiver to the energy incident on the
collector’s aperture. The optical efficiency is related with the radiative properties of optical materials, construction
and operation of the collector and the geometric factors due to the position of the collector with respect to the sun.
The optical efficiency of the collector, can be defined with the following terms,

(8)

The optical efficiency of the collector depends upon the radiative and optical properties of the receiving and the
reflecting surfaces which are affected instantaneously by change in incidence angle. Thus the relation between the
incidence angle and the radiative properties of the optical elements are demonstrated herein through the
calculation of the optical efficiency.

Cover transmittance
The cover material is made of DURAN type glass whose transmission property depending on the wavelength and
the wall thickness shown in Fig. 3 (http://www.schott.com). As it seems from the figure, its transmittance is nearly
constant within the solar radiation spectrum. However, the directional surface property of the cover varies with
solar incidence angle. The relationship between the normal specular transmittance and its angular dependency is
described to take into account the cover material used.

Fig. 3. Transmission of DURAN glass for thicknesses 1, 2 and 8 mm

970
Since the receiver is composed of a single glass cover, the solar radiation is refracted from the ambient air with a
refractive index, , to the glass with a refractive index, . The angles and are related to the indices of
refraction by Snell’s law,

(9)

The parallel and perpendicular components of unpolarized radiation for smooth surfaces can be defined using
Fresnel’s derivation,

(10)

The absorption of the solar radiation in the partially transparent cover is determined according to the assumption
that some portion of the incidence radiation absorbed by the cover material.

(11)

where K is called the extinction coefficient that is considered as 4 m-1 for “water white” glass.

For a single cover, the transmittance is defined by averaging the transmittance of the parallel and perpendicular
components of polarization when only reflection losses are considered. As a result, the transmittance of the cover
element is specified from Eq. (12) depending on incidence angle.

(12)

Absorptivity of the receiver

The absorbance property of the absorber tube, which is made of stainless steel, depends also upon the incidence
angle however there is not sufficient study available in the literature for describing the relationship between the
angular absorbance and the normal absorbance. Nevertheless it is estimated from the relation that provides fair
results within 0° and 80° incidence angle (Duffie and Beckman, 2005).

(13)

Intercept factor of the PTSC

The intercept factor is an optical property which is defined as the ratio of the solar radiation intercepted by the
receiver to the radiation reflected by the parabolic mirror that strikes directly to the receiver. It is affected by
random and non-random errors (Guven and Bannerot, 1984). Random errors are truly random in nature owing to
the fact that they can be represented by normal probability distributions with zero mean. These errors originate
from the misalignment of the ideal focus of the reflector depending on the energy distribution of the sun shape due
to purely random tracking errors, , the slope errors of the reflector surface during its manufacture, assembly
and/or operation due to the surface imperfections, , and the scattering effects due to the property of the
optical material used, . However, random errors are modeled statistically, by determining the standard
deviation of the total reflected energy distribution, at normal incidence (Guven and Bannerot, 1984), are defined in
the form of

(14)

Non-random errors are generally defined with their permanent characteristics which can be related directly to
anticipate errors in manufacture/assembly and/or in operation. In general these errors will cause the central ray of
the reflected energy distribution (effective sunshape) to shift from the design direction of the focusing line. The
value of the intercept factor is governed by random and non-random errors as well as the collector geometric
parameters like area concentration ratio (the ratio of the area of aperture to the area of the receiver) and receiver
diameter to yield universal error parameters. Using the universal error parameters, the intercept factor is defined
as (Guven and Bannerot, 1985)

971
 (15)

Where

: dislocation of receiver center from ideal focus (m).

: outside diameter of receiver tube (m).
: misalignment slope error (deg).
: area concentration ratio, .
: universal nonrandom error parameter due to receiver mislocation and reflector profile errors,
.
: universal nonrandom error parameter due to slope error, .
: universal random error parameter, .

For the evaluation of the intercept factor, a computer program was written in Mathematica to solve Eq. (15)
numerically using Trapezoidal rule. The obtained numerical solutions used for different random and nonrandom
collector parameters gave consistently satisfactory results compared to (Guven and Bannerot, 1984). Based on
this program, the intercept factor of the collector was calculated as considering it a carefully-fabricated collector
(Arasu and Sornakumar, 2007) so that the parameters were taken into account regarding a lower limit of 0.0025
radians, which is the standard deviation of the energy distribution of the sun's rays at solar noon on a clear day, for
the random error parameter, , 0.004 radians for and 0.002 radians for . When the maximum
tracking error is considered as 0.0035 rad and the nonrandom slope error is used as 0.0035 rad, the intercept
factor is obtained approximately 1.

End-effect Correction
Another optical property of the PTC is the end-effect which is a measure of the reduction in aperture area due to
the fact that some part of the reflected beam radiation along the aperture end of the collector cannot be received
by the absorber tube therefore the image is formed beyond the end of it. The aperture area lost is given by end-
effect correction is expressed with Eq.(16).

(16)

Thermal Analysis of the Collector Array

Thermal analysis of the collector system is significant to determine its thermal efficiency which means how much
of the solar radiation is converted into useful gain by the working fluid. In this stage, the heat losses from the
receiving element are to be analyzed. A comprehensive analytical analysis using thermodynamics and heat
transfer relations have been made in order to establish a model that is consistently to meet the nearest actual
system behavior. Thus the model study is developed on the base of a number of assumptions that give reasonable
results with holding the physical nature of the system. The main characteristics of the model include the following
assumptions:

 The PTSC performs under steady state conditions in other words; there is no change in the
thermodynamic state properties with respect to time since the surface temperatures remain constant.
 Heat transfer is one-dimensional since any significant temperature gradients exist in radial direction.
There is a negligible temperature change along the length of the PTSC. On account of the fact that
Forristall (2003) compared its one-dimensional model with two-dimensional model and indicated that the
two-dimensional model provides a higher degree of accuracy than the one-dimensional model when
evaluating long receivers with a length greater than 100 m. However the total length of the receiver tube
used in this model study is lower than 10 m. Hence, one-dimensional model is totally reasonable from this
view.
 Heat losses from the retainers for absorber tube are negligible.
 The atmospheric emission is treated as a blackbody’s emission at an effective sky temperature.
 Dust and dirt over the PTSC array are neglected.

During the process of gaining energy from the sun, the receiver tube heats and correspondingly losses its some
energy as heat during the system operation. At the beginning of the thermal analysis, the calculation of the heat
losses from the receiver tube is required to reach the useful heat gain of the working fluid. The cross-section of the
receiver tube and its resistance network is shown in Fig. 4. As it is seen from the figure, the forced heat convection
within the absorber is conducted through the absorber wall then the transfer of heat flow exhibits itself as the sum

972
of the radiation and the natural convection between the absorber and the cover. It is conducted through the cover
material and equals to the sum of convection to the ambient air and radiation to the sky over the outer surface of
the cover material.
The heat transfer from the HTF to the absorber pipe is stated by forced convection due to pumping operation. The
flow regime is defined according to the Reynolds number (Re) which specifies whether it is laminar or turbulent.
The flow in a circular pipe is laminar when the generally accepted value of the critical Re is considered below
2300. The total mass flowrate is specified to reach the mean velocity using the conservation of mass principle and
Re of the HTF for defining the flow regime. Typical technical data of the HTF is given in Table 2.

(17)

Fig. 4. Thermal analysis of the receiver

It is considered that a constant heat flux occurs on the surface of the absorber tube due to the distribution of sun’s
radiation. For fully developed flow in a circular tube subjected to constant surface heat flux, the Nusselt number
(Nu) is constant and its value is obtained from Eq. (18).

(18)

When the Re is higher than 2300, the flow turns to transition. In this case, the Nu relation for a circular smooth
pipe is defined by the Gnielinski equation which covers both the transition and the turbulent regions for fully
developed turbulent flow (Rohsenow et al., 1998).

(19)

where is the friction factor, for smooth pipes.

Thus, the convection heat transfer coefficient of the HTF is obtained by

(20)

Then, the convection heat transfer within the absorber tube is defined using the Newton’s law of cooling,

(21)

The bulk mean fluid temperature, which is used as the arithmetic average of the mean temperatures of the
fluid at the inlet and the exit. During heating process, the mean temperature of the fluid changes so the fluid
properties in internal flow are usually evaluated at this temperature.
Heat transfer through a hollow cylinder using the Fourier’s law of heat conduction can be applied analogously for
the absorber tube.

(22)

973
The heat transfer between the absorber and the cover is equal to the sum of the convection and the radiation.
Convection heat transfer within the annulus is considered to have generated naturally. Heat transfer in enclosure
with an annulus type was studied by Raithby and Hollands for two long concentric horizontal cylinders with a range
of and (Cengel and Ghajar, 2011). The rate of heat transfer through the
annular space by natural convection with a characteristic length, is expressed as

(23)

The recommended relation for effective thermal conductivity is defined by

(24)

where the geometric factor for concentric cylinders is

(25)

Radiation heat transfer within the annulus occurs due to temperature difference between the absorber and the
cover even if the annulus space comprises air. Considering the enclosure consisting of two opaque surfaces at
equilibrium temperatures, the radiation heat transfer is calculated by Eq.(26) and it is applicable to two gray
(independent of wavelength), diffuse (independent of direction) and opaque surfaces. These approximations are
reasonable since the absorber behaves nearly gray in the solar spectrum and is diffuse emitter and its emissivity
remains nearly constant for about the incidence angle lower than 40° such that metals (Touloukian and DeWitt,
1970). The calculations for the incidence angle illustrates that this situation is possible when it is considered with
solar tracking. The cover material behaves like an opaque material at longer-wavelength infrared regions of the
electromagnetic spectrum during heat loss. When these approaches are considered, the radiation heat transfer
within the annulus can be expressed by

(26)

Heat transfer through the cylindrical layer of the cover material is described by

(27)

Heat loss on the outer side of the cover is treated as the sum of the convection and the radiation heat transfer
modes. The convection heat transfer can change depending on windy conditions of ambient air. The calculation of
heat transfer requires the average heat transfer coefficient over the entire surface although variation of the local
heat transfer coefficient along the circumference of a circular cylinder is not uniformly distributed. Supposed that
the cover material is cylindrical in shape and the heat transfer over it is considered as cross-flow of air. The
average Nusselt number in this case is estimated with Zhukauskas’ correlation (Cengel and Ghajar, 2011) when
the flow is presumed external forced convection for flow across cylinders and expressed compactly as

(28)

where and the experimentally determined constants C and m are given in Table 3.

Table 3. Constants for Eq.(28)

Range of Re C m
0.4‒4 0.989 0.330
4‒40 0.911 0.385
40‒4000 0.683 0.466
4000‒40,000 0.193 0.618
40,000‒400,000 0.027 0.805

974
If Re is lower than 0.4, the flow is treated to turn towards free convection. The relation implies this circumstance is
given in Eq.(29). All the fluid properties are evaluated at the film temperature , which is the
average of the free-stream and surface temperatures.

(29)

Then, the convection heat transfer outside of the cover can be denoted by

(30)

The radiation heat transfer is caused by the temperature difference between the cover and the sky. Although the
atmospheric emission character is far from a blackbody’ exhibits, it is convenient to treat it as a blackbody at some
lower fictitious temperature i.e. the effective sky temperature that emits an equivalent amount of radiation energy.
The simple relationship between the local air temperature and the sky temperature is expressed in (Duffie and
Beckman, 2005).

(31)

In this case, the radiation transfer between the glass cover and the sky is given by

(32)

The net heat loss from the receiver tube is caused by the temperature difference arising in radial direction. In order
to determine the temperatures through the receiver, the steady-state energy balance relations are written between
the nodal points of the resistance network. The model algorithm makes it possible to solve the non-linear
equations for calculating the temperatures and the heat transfer coefficients through the receiver. These energy
balance relations are expressed with three sets of equality as shown in Eq. (33).

(33)

During the evaluation of the thermal losses, the estimation of the absorbed solar radiation, the thermal loss
coefficient and the heat removal factor are required. All these parameters are calculated subsequently to find the
temperature of the surfaces until the initial iteration value for the exit temperature of the HTF converges to its
ending value while satisfying the energy balance and heat transfer relations.
The thermal loss coefficient of the receiver tube is defined as below when considering the absorber tube covered
and no temperature gradients around the receiver (Duffie and Beckman, 2005).

(34)

where and is the radiation heat transfer coefficients. The radiation heat transfer coefficient enables to
express radiation conveniently in an analogous manner with convection in terms of a temperature difference. The
linearized form of the radiation coefficient can calculated from Eq.(35) and Eq.(36) for the annulus region and the
outside of cover, respectively.

(35)

(36)

The actual useful energy gain is quantified by the collector heat removal factor, which is used as the
effectiveness of the receiver tube. It is related to lowering of the maximum possible useful energy gain since the
whole receiver surface temperature is not at the fluid inlet temperature.

975
 (37)

where is the absorbed solar energy by the receiver. It is directly related with beam radiation coming from the sun
and the optical efficiency of the collector as shown in Eq.(38).

(38)

The useful energy gain transferred to the HTF under steady-state condition can be written in the form of

(39)

where . Here is named as shading factor which is basically used as a multiplier

for specifying the effective collector aperture area. It is related with what fraction of the reflector is shaded by the
receiver. Since the shaded area has lost its effectiveness on account of not having subjected to the solar heat flux.
After all the parameters affecting the thermal efficiency are defined, the useful energy gain absorbed by the HTF
can be obtained from the energy balance. The exit temperature of the HTF whose sensible energy increases due
to not undergoing in phase change is calculated from Eq.(40).

(40)

The thermal efficiency of the array is established maintaining the system parameters or variables stable under
steady state conditions. Hence, these variables can be regarded as constant during system operation, and the
thermal efficiency of the PTSC array is associated with

(41)

RESULTS AND DISCUSSION

The thermal performance of the PTSC is affected by operating conditions under which it functions. There are many
system parameters which play roles over the PTSC’s performance. The very dominants of them can be
considered as DNI (Direct Normal Irradiance) and operating temperature. The effects of the other factors were
figured out from the model analysis as mass flow-rate of HTF, receiver thermal properties and ambient conditions.
Fig. 5 indicates the variation in the thermal efficiency of the PTSC array with respect to the bulk mean fluid
temperature of the HTF. Typical operating condition on the 1th day of July at 12:00 solar time is dealt with when
the incident angle of the direct beam radiation, 13.98°, the HTF mass flow-rate 0.3 kg/s, ambient temperature
39°C, ambient air velocity 1 m/s; the estimated collector array efficiency is reduced due to increasing thermal
losses. The model predictions are to be confirmed by the operation of PTSC being installed at Gaziantep.

0.70

0.65

0.60

0.55
940 W/m²
ηPTSC 0.50 840 W/m²

0.45 740 W/m²

0.40

0.35

0.30
50 100 150 200 250 300
Tbulk C

976
Fig. 5. Thermal performance of the PTSC array

CONCLUSIONS
In this paper, a comprehensive thermo-mathematical modeling used for the analysis of a parabolic trough solar
collector is presented. The model is performed by developing a program in EES. The model comprises one-
dimensional steady-state heat transfer approach using the thermodynamics and the heat transfer relations to
evaluate the thermal analysis of the receiver by taking into account the actual system parameters. In addition that
it covers the solar energy calculations and optical efficiency of the collector array. The validation of the model is in
accordance with the literature. The consistency of the model is to be identified after installing the collector system
in Gaziantep.
Acknowledgements
The authors extend their appreciation to the Scientific Research Projects Governing Unit at Gaziantep University
for funding the project No. MF.11.13.

NOMENCLATURE
cross-sectional area, m2
apparent sun time, min
concentration ratio
specific heat capacity, J/kg°C
diameter, m
daylight savings time, min
equation of time, min
focal distance, m; friction factor
heat removal factor
solar constant, W/m2
extraterrestrial radiation, W/m2
convection heat transfer coefficient, W/m2°C
beam radiation, W/m2
thermal conductivity, W/m°C
extinction coefficient, m-1
effective thermal conductivity, W/m°C
thickness of glass cover, m
parabola length, m
local meridian, deg
standard meridian, deg
local sun time, min
mass flow-rate, kg/s
number of day; refractive index; number of PTSC
rate of heat transfer, W
useful energy gain, W
perpendicular component of unpolarized radiation
parallel component of unpolarized radiation
absorbed solar energy, W/m2
temperature, °C
ambient temperature, °C
bulk mean fluid temperature, °C
exit temperature, °C
film temperature, °C
inlet temperature, °C
thermal loss coefficient, W/m2°C
mean velocity, m/s
parabola aperture, m

Greek Letters
absorptivity
normal absorptivity
intercept factor
declination angle, deg
emissivity

977
 shading factor
optical efficiency
efficiency of PTSC array
incidence angle, deg
zenith angle, deg
end-effect correction
average specular reflectance of parabolic mirror; density, kg/m3
total reflected-energy standard deviations at normal incidence; Stefan-Boltzmann constant, W/m2K4
standard deviation of the energy distribution of the sun’s rays at normal incidence, rad
standard deviation of the distribution of local slope errors at normal incidence, rad
standard deviation of the variation in diffusivity of the reflective material at normal incidence, rad
tranmissivity
rim angle, deg
latitude, deg
hour angle, deg

Subscripts
medium number
absorber
gap between absorber and cover
ambient air
annulus
cover
coduction
convection
heat transfer fluid
inside of absorber
inside of cover
outside of absorber
outside of cover
radiation
sky

REFERENCES
Arasu A. V. and T. Sornakumar. 2007. Design, manufacture and testing of fiberglass reinforced parabola trough
for parabolic trough solar collectors. Solar Energy 81: 1273–1279.

Cengel Y. A. and A. J. Ghajar. 2011. Heat transfer and mass transfer: Fundamentals & Applications. New York:
McGraw-Hill.

Duffie J. A. and W. A. Beckman. 2005. Solar engineering of thermal processes. New York: John Willey & Sons.

Guven H. M. and R. B. Bannerot. 1984. Optical and thermal analysis of parabolic trough solar collectors for
technically less developed countries–Technical report 1. USAID Project Grant No. 386–0465.

Guven H. M. and R. B. Bannerot. 1985. Derivation of universal error parameters for comprehensive optical
analysis of parabolic troughs. Proceedings of the ASME-ASES Solar Energy Conference, Knoxville, USA, 168–
174.

http://www.fuchs-oil.de. Product Information–Renolin Therm 320 Heat Transfer Fluid.

http://www.schott.com. Technical datasheet of DURAN Glass–Tubing, Capillary and Rod of Borosilicate Glass 3.3.

Huang W., P. Hu, Z. Chen. 2012. Performance simulation of a parabolic trough solar collector. Solar Energy 86:
746–755.

Jacobson E., N. Ketjoy, S. Nathakaranakule and W. Rakwichian. 2006. Solar Parabolic Trough Simulation and
Application for a Hybrid Power Plant in Thailand. ScienceAsia 32: 187–199.

Padilla R. V., G. Demirkaya, D. Y. Goswami, E. Stefanakos, M. M. Rahman. 2011. Heat transfer analysis of
parabolic trough solar receiver. Applied Energy 88: 5097–5110.

978
Qu M., H. Yin and D. H. Archer. 2010. Experimental and Model Based Performance Analysis of a Linear Parabolic
Trough Solar Collector in a High Temperature Solar Cooling and Heating System. Journal of Solar Energy
Engineering 132: 1–12.

R. Forristall. 2003. Heat transfer analysis and modelling of a parabolic trough solar receiver implemented in
engineering equation solver.: NREL/TP–550e34169.

Rohsenow W. M., J. P. Hartnett and Y. I. Cho. 1998. Handbook of Heat Transfer. New York: McGraw-Hill.

S. A. Kalogirou. 2012. A detailed thermal model of a parabolic trough collector receiver. Energy 48: 298–306.

Touloukian Y. S. and D. P. DeWitt. 1970. Thermal Radiative Properties, Metallic Elements and Alloys. New York:
IFI/Plenum.

979
 SELECTION OF KAPLAN AND FRANCIS TURBINE IN HYDRAULIC POWER PLANTS:
A CASE STUDY

A.Y. Demirhan, Z. Yumurtaci

Yildiz Technical University

Mechanical Engineering Department, 34349 Besiktas-Istanbul
zyumur@yildiz.edu.tr

ABSTRACT
“Discharge” and “head” are the main characteristics that settle the turbine type at hydroelectric power plants.
At some specific speed values, the turbine can be either Francis or Kaplan during selection procedure. This
study is concerned with the joint operating regions of Francis and Kaplan turbines. The calculations are
carried out on a hydraulic plant, Berdan, in Turkey. Berdan hydraulic power plants consists two unit vertical
axis Francis turbine of 5.2 MW, total of 10.4 MW. With respect to its discharge and head, turbine types can
be Francis or Kaplan. The studies indicate that, according to the annual working reports, if the turbine type
was Kaplan, electricity generation would be 600.000 kWh farther on 2005 because of the higher efficiency
rates of the Kaplan turbine at partial discharges.

INTRODUCTION
Hydraulic turbines involve in their action a continuous transformation of the potential energy of the fluid into
kinetic energy and a conversion of kinetic energy, or of both kinetic and potential energy, into mechanical
shaft power, which can be used to drive an electrical generator, or other machinery. The power available is
proportional to the product of pressure head and volume flow rate. The general formula for any hydro
system’s power output is:

(1)

Hydraulic turbines are divided according to their hydraulic action into two mail classes; impulse turbines and
reaction turbines.
In reaction turbines water pressure can apply a force on the face of the runner blades, which decreases as it
proceeds through the turbine. Turbines that operate in this way are called reaction turbines. The turbine
casing, with the runner fully immersed in water, must be strong enough to withstand the operating pressure.
In action turbines water pressure is converted into kinetic energy before entering the runner. The kinetic
energy is in the form of a high-speed jet that strikes the buckets, mounted on the periphery of the runner.
Turbines that operate in this way are called impulse turbines. As the water after striking the buckets falls into
the tail water with little remaining energy, the casing can be light and serves the purpose of preventing
splashing.
A significant factor in the comparison of different turbine types is their relative efficiencies both a design point
and at reduced flows. An important point to note is the Pelton, Crossflow and Kaplan turbines retain very
high efficiencies when running below design flow; in contrast the efficiency of the Francis turbine falls away
sharply if run at below half its normal flow, and fixed pitch propeller turbines perform very poorly except
above 80% of full flow.(Celso P. 2004, Başeşme H. 2003, V. J. Zipparro.1993,Pash.O 2002)
In this study annual energy production of Berdan HPP was calculated for two alternative turbine settlements
with Kaplan type turbines and compared with current turbine settlement.

Table 1.Hydraulic Turbine Types

Pelton Turbine
Impulse Turbines Turgo Turbine
Cross-Flow Turbine
Francis Turbine
Diagonal Flow Turbine
Reaction Turbines
Propeller (Kaplan) Turbine
Tubular (Bulb) Turbine

980
OPTIMAL HYDRAULIC TURBINE SELECTION
He turbine type is selected on the basis of the effective head and turbine discharge, while considering such
factors as river flow, operation of the reservoir and regulating pondage(head fluctuation and flow fluctuation).
The type, geometry and dimensions of the turbine will be fundamentally conditioned by the following criteria:
 Net head
 Range of discharges through the turbine
 Rotational speed
 Cavitation problems
 Cost
The first criterion to take into account in the turbines selection is the net head. Table 2.specifies for each
turbine type its range of operating heads. The table shows some overlapping, so that for a certain head
several types of turbines can be used. When two or more turbine types are possible, they are determined by
comprehensive studying their cost, efficiency and maintainability.(Celso P. 2004)

Table2. Hydraulic turbines head ranges

Turbine Type Head Range (m)
Kaplan ve Propeller 2<Hn<40
Francis 25<Hn<350
Pelton 50<Hn<1.300
Crossflow 5<Hn<200
Turgo 50<Hn<250

As a turbine can only accept discharges between the nominal and the practical minimum,it may be
advantageous to install several smaller turbines instead of a one large. The turbines would be sequentially
started, so all of the turbines in operation except one will operate at their nominal discharges and therefore
will exhibit a higher efficiency. On the other hand using double regulated turbine like Kaplan type turbine has
a minimum practical discharge close to 20% of the rated discharge whereas in a single regulated like Francis
type it is close to 40%,whenever a scheme has to cope with flows as low as 40% of the nominal one,
thedouble regulated turbine should be selected.(Celso P. 2004)

Fig1.Turbines’ type field of application

981
Also compare with Pelton and Francis turbines, Pelton turbines can work higher efficiency in partial loads.
However Francis turbines need lesser installation space, have greater runner speed and also can attain
greater efficiencies for high power values.

Table 3. Flow and head variation acceptance

Turbine Type Acceptance of flow Acceptance of head
variation variation
Pelton High Low
Francis Medium Low
Kaplan double regulated High High
Kaplan single regulated High Medium
Propeller Low Low

Selection of Specific Speed

The specific speed constitutes a reliable criterion for the selection of the turbine, without any doubt more
precise than the conventional enveloping curves, just mentioned in figure 1.
The runner size as well as its shape and characteristics vary depending on the head or flow, even with the
same type turbine. However, if the runner shape is homologous geometrically, it is regarded to have the
same characteristics without regard to its size. (Başeşme H. 2003, Ramos. H 1996, Özgür. C. 1973,)
The specific speed of a certain turbine is the rotating speed of an hypotetical turbine which is homologous to
the original turbine and produces unit output 1kW under unit head 1m.
The specific speed of the turbine is defined by the following equation.

(2)

For synchronous generators, its speed depends upon the frequency of the electric grid.

(3)

For 50 Hz grid frequency rotation speed of the turbine and generator should be 1500-1000-750-600-500-
428-375-250-214-187,5-166,7-150-125-107-93-75 and 60.
When the head and output are given, selection of a turbine of higher speeds or higher makes it possible to
realize smaller turbine and generator as well as a smaller powerhouse building. There is, however, a limited
to such miniaturization in the aspect of cavitation performance or equipment strength.
The limiting line tends to be gradually extended by today’s technological progress, bun an applicable limit
for the effective head H is experimentally determined. The rotating speed must be selected in the range of
the specific speed applicable to each turbine model as shown Table 4.
In general turbine manufacturers specify the specific speed of their turbines. A large number of statistic
studies undertaken by De Siervo and Lugaresi. Here under some of the correlation formulae represented.

PeltonTurbine (1 jet)

(4)

Francis Turbine

(5)

Kaplan Turbine

(6)

982
Table 4.Range of specific speed for each turbine type
Turbine Type Specific Speed (ns)
Pelton, Turgo,Cross-Flow 8~25
Francis 50~350
Kaplan , Propeller 250~700

BERDAN HPP ALTERNATIVE TURBINE SELECTION

Project Area
Berdan Dam is constructed on the Berdan River in Mersin Province. The Berdan River, also called the
Tarsus located south Turkey.
The main headwaters are in the Toros Mountains. There are two main tributaries: Kadıncık and Pamukluk.
Total length of the river is 124 kilometersAlthough the river is quite short, the average flow rate is 42 m3/s,
which is higher than most short rivers in the vicinity. The drainage basin is 1592 km2. It flows to the
Mediterranean Sea.
There are four dams on Berdan. These are used both for controlling floods and for producing electricity.But
the lower reaches of the river flow in an agricultural area, and because of the pollution caused by fertilizers
the dams in the lower reaches are not used for drinkingwater. (http://en.wikipedia.org/wiki/Berdan_River )

Fig.2Berdan HPP Project Area

Characteristics of Berdan Dam

Berdan Dam is a embankment dam on the Berdan River. It was built between 1975 and 1984 It supports a
power station and provides water for the irrigation of 27,050 hectares.
The hydroelectric power plant (HEPP) has a two unit Francis turbine with 10,4 MW installed power capacity
of 10,4 MW and generates 47,6 GWh electricity annually. (http://www2.dsi.gov.tr/bolge/dsi6/mersin.htm )

Table 5 Characteristics of BerdanHPP

Height 41,60 m
Installed Capacity 10,4 MW
Annual Production 47,50 GWh
Turbine Type Vertical Francis Turbine
Penstock Diameter 3,45 m
3
Project Flow 48 m /s
3
Unit Flow 24 m /s
Net Head 24 m

983
 2
The reservoir Lake Berdan Dam, extending over an area of 6,7 km with a water volume of 87,50 hm³ The
reservoir water level varies between 26,40 and 15,60m from tail water. The full reservoir level is 28,60 m at
and the minimum operation level is 15 m from tail water.

Turbine Selection and Specific Speed

Kaplan and propeller turbines are axial-flow reaction turbines; generally used for low heads from 2 to 40 m.
The Kaplan turbine has adjustable runner blades and may or may not have adjustable guide vanes. If both
blades and guide-vanes are adjustable it is described as "double-regulated". If the guide-vanes are fixed it is
"single-regulated". Fixed runner blade Kaplan turbines are called propeller turbines. (Celso P. 2004)
For a small/medium low head power units reaction turbine are used. Selection of turbine type is essential
based on specific speed.Berdan HPP has two unit vertical Francis turbine. According to head and flow it is
possible to design this plant with Kaplan type turbine. Also it is possible to use single or two unit Kaplan type
turbine.
The double regulation allows, at any time, for the adaptation of the runner and guide vanes coupling to any
head or discharge variation. It is the most flexible Kaplan turbine that can work between 15% and 100% of
the maximum design discharge. In case of using double regulating Kaplan turbines in Berdan HPP efficiency
will be increased in partial flows.
By using given project characteristics specific speed was calculated for both one and two unit Kaplan type
turbine by using Eq.(2).

Specific speedin case of using Kaplan type turbine:

For Kaplan type turbines ns speed range is 250-700. This specific speed is in Kaplan type turbine operating
range. As seen on Fig.1 Berdan HPP also suitable for both Francis and Kaplan type turbine.

Rotation speed of 2 unit Kaplan turbine;

The rotational speed nearest the design speed is:

Rotation speed of single unit Kaplan turbine

The rotational speed nearest the design speed is:

Turbine Characteristics for Two Unit Kaplan

Berdan HPP turbine characteristics were determined for 2 Unit Kaplan type turbine by using TURBINPRO
program and efficiency curve was plotted. Program outputs and hill curves can seen on Table 6 and Fig. 3

Table 6. Turbine characteristics

3
Unit Flow 24 m /s
Net Head 24 m
Turbine Efficiency % 93,8
Specific Speed 513,9
Rotation Speed 375 rpm
Unit Power 5.299 kW
Total Power 10.598 kW

984
Fig.3 Kaplan turbine performance curves (head-discharge)

Turbine Characteristics for Single Unit Kaplan

Berdan HPP turbine characteristics were determined for one unit Kaplan type turbine by using TURBINPRO
program and efficiency curve was plotted. Program outputs and hill curves can seen on Table 7 and Fig 4.

Table 7. Turbine characteristics

Unit Flow 48 m3/s
Net Head 24 m
Turbine Efficiency % 94,1
Specific Speed 485,2
Rotation Speed 250 rpm
Unit Power 10.629 kW
Total Power 10.629 kW

Fig.4 Kaplan turbine performance curves (head-discharge)

985
ENERGY PRODUCTION COMPARISONS OF ALTERNATIVE TURBINE SELECTIONS

Comparison of Turbine Efficiencies

Comings flows distributed on the turbines equally and corresponding turbine efficiencies noted from Fig3 and
Fig4. As a result plant compound turbine efficiency curve was plotted. Fig5 also shows the useable region of
the flow curve. Every selected turbine has a minimum technical flow (with a lower discharge the turbine
either cannot operate or has a very low efficiency) and its efficiency is a function of the operating discharge.
Till the %50 percent of total flow both turbine configuration have similar efficiency however single unit
configuration efficiency decrease dramatically after %25 flow rate.

Fig.5 Combined plant efficiency curve

This curve is used for calculating annual energy production of each turbine selection and calculated energy
production in each situation compared with actual plant energy production.

Comparison of Energy Production

Berdan HEPPs alternative energy production is calculated by using EqX.Monthly average flows were used
for calculating energy production. For every month average turbine efficiency was interpreted from Fig.5
Overall plant efficiency calculating by using Eq.(6). Combined plant efficiency multiplied by generator and
transformer efficiencies as a result overall plant efficiency was determined.

(7)

The average annual energy production (E in kWh) is a function of:

When the average monthly flow multiplied with overall efficiency and average working hour of the month, we
can calculate the energy production of power plant by using Eq. (8).Monthly and total energy production of
each alternative turbine selections was given on Table 8.

(8)

986
Table 8.Monthly energy productions

Average 2 Unit Kaplan 1 Unit Energy Energy

MonthWorking Plant Overall Plant Kaplan Production Production
Hour (t) Discharge Efficiency Overall Plant of 2 Unit of Single
3
( m /s) Efficiency Kaplan Unit Kaplan
Turbine Turbine
(kWh) (kWh)

Jan 320 12,34 0,80 0,79 637.577 626.770

Feb 350 18,56 0,90 0,84 1.186.201 1.106.865
March 350 31,06 0,89 0,87 2.024.437 1.966.311
April 350 34,39 0,90 0,87 2.737.809 2.659.184
May 350 34,35 0,90 0,87 3.168.486 3.077.493
June 350 43,04 0,91 0,88 3.587.106 3.480.210
July 350 32,78 0,89 0,87 2.661.753 2.588.223
Agu 350 23,25 0,91 0,85 1.802.892 1.689.447
Sept 250 24,04 0,91 0,85 1.189.075 1.114.297
Oct 210 21,73 0,91 0,85 921.094 859.161
Total Energy Production 19.916.431 19.167.962

CONCLUSION
There are many different types of turbine configurations available for low-head hydro power plants. Both
axial flow (propeller) and radial flow (Francis) turbines can be used in low head hydro power plants.
In this study calculations for turbines of Kaplan type which may be an alternative to the Francis type are
made according to the data of net head and flow that takes place in the study report of EIE in 2005 and the
total energy output values under the same circumstances are calculated for both two unit Kaplan and single
unit Kaplan turbine. As a result it is found out and shown that if the turbines of Kaplan type which are more
efficient with partial loads have been used instead of Francis type, the total electricity output would have
been 600.000 kW more in 2005 compare the current turbine settlement. On the other hand for single unit
Kaplan type turbine settlement energy production equal to current situation. In case of choosing single unit
Kaplan unit Berdan HPP generates the same energy with lower investment with half size power plant.

NOMENCLATURE
power, kW
efficiency
3
water density, kg/m
2
gravity, m/s
3
flow, m
head, m
specific speed
rotation speed, rpm
frequency, Hz
number of pole
energy, kWh
time, h

REFERENCES
CelsoP ,DrIngeniero de Minas .2004.Laymans Handbook on How to develop a small Hydro Site: ESHA
Başeşme H. 2003.HidroelektrikSantrallerveHidroelektrikSantralTesisleri.EÜAŞYayınları,
Ramos. H, A.Betamio de Almeida, M. Manuela Portela. 1996. Guidelines for Design of Small Hydropower
Plants, Lisbon.
Özgür. C. 1973. Su MakinelarıDersleri:ÇağlayanBasımevi.

987
V. J. Zipparro , H. Hans.1993. Davis’ Handbook of Applied Hydraulics:McGraw-Hill, New York.
Pash.O.2002. Small hydro power: technology and current status.Renewable and Sustainable Energy
Reviews6:537-556
Website: http://en.wikipedia.org/wiki/Berdan_River
Website: http://www2.dsi.gov.tr/bolge/dsi6/mersin.htm

988
 COMPARING CORN DRYING IN FLUIDIZED BED DRYER AND CONVECTIVE TRAY DRYER
a,* b
M. Gur , M. Gur
a
Department of Computational and Systems Biology, School of Medicine, University of Pittsburgh, Pittsburgh, PA 15213,
U.S.A.
b
Department of Mechanical Engineering, Faculty of Mechanical Engineering, Istanbul Technical University, Istanbul,
34437, Turkey
meg78@pitt.edu, mesutgur@itu.edu.tr
Correspondence: Mert Gur
Tel: (412) 383-5804
Fax: (412) 648-3163

ABSTRACT
In this study, the thin layer drying characteristics of corn was experimentally investigated in a fluidized bed dryer
and a convective tray dryer. Corn with an initial moisture content of 25 % dry basis was used in the experiments.
Experiments were performed at 3 different drying temperatures: 55, 65 and 75° C. In each experiment one layer of
corn was inserted into the dryer and then drying was performed for 90 minutes. The drying air velocities were set
to 0.45 m/s for the convective tray dryer and 2.5 m/s for the fluidized bed dryer. Single exponential and Page
models were fitted onto the measured drying data. Effective diffusivities were calculated using Fick’s first law. The
equilibrium moisture contents were evaluated using the Guggenheim-Andersen-de Boer model. From the
experiments it was observed that drying rates in the fluidized bed dryer were considerably higher than drying rates
in the convective tray dryer. Effective diffusion coefficients obtained in the fluidized bed dryer were approximately
%70 higher than those in the convective tray dryer. This increase is mainly attributed to the homogeneous mixing
motion in the fluidized bed dryer, which enhances heat and mass transfer between the corn and the drying air. In
addition, the temperature dependence of the effective diffusion coefficients was described by an Arrhenius-type
relationship.

INTRODUCTION
Corn is one of the main agricultural products in Turkey (Doymaz and Pala, 2003) and its preservation is of great
importance. As is the case for almost all agricultural products in Turkey, corn is most commonly preserved by sun-
exposed drying in the open air. In sun-exposed drying, food is in continuous contact with the open air for an
extended period of time. During this period, the food may easily become contaminated due to the soil transferred
by the air (Oztekin et al., 1999). In order to reduce contamination and drying time, the use of industrial dryers is
required. Many types of industrial drying methods for food have been investigated in the literature. Among these
perform fluidized bed drying stands out due to its high energy savings and short drying times which are caused by
the effective contact between the drying mediums (Wang and Chen, 2000 , Tatemoto et al., 2001).

In order to successfully predict drying behavior and optimize drying parameters, the drying kinetics of corn must be
understood. Many thin layer models have been proposed in the literature (Henderson and Pabis, 1961 , Wang
and Singh, 1978 , Zhang and Litchfield, 1991 , Liu and BakkerArkema, 1997 , Yaldiz et al., 2001 , Doymaz and
Pala, 2003). Thin layer characteristic for foods are generally derived from conventional convective drying
experiments, where food is put on a perforated tray. For corn thin layer drying characteristics were extensively
investigated in convective tray dryers (Overhults et al., 1973 , Li and Morey, 1984 , Zhang and Litchfield, 1991 ,
Doymaz and Pala, 2003) and corresponding effective diffusion coefficients, Deff, were proposed (Doymaz and
Pala, 2003). Corn drying was also studied in batch fluidized bed dryer (Jumah et al., 1996 , Soponronnarit et al.,
1997 , Soponronnarit et al., 1997 , Syahrul et al., 2003 , Hatamipour and Mowla, 2006) . However, in none of
these studies fluidized bed drying experiments were conducted under thin layer drying conditions. Thus, there is a
lack of work in the literature regarding the thin layer drying of corn in the fluidized bed dryer.

Recently Madhiyanon et al. (Madhiyanon et al., 2009) have investigated the thin layer behavior of coconut in
fluidized bed dryer and have determined the effective diffusivities from the experimental results. The value of the
effective diffusivities were 1-4 orders of magnitude larger than the thin layer results obtained for other food and
biological materials dried in conventional convective tray dryers (no fluidization). The large discrepancy was
attributed to the homogeneous mixing motion in the fluidized bed dryer, which boosts heat and mass transfer
between corn and air. Consequently, from this work it can be concluded that the thin layer kinetics distinguish
significantly between the fluidized bed dryer and the convective tray dryer.

The main objective of the present study is to determine the thin layer drying characteristics of corn in the fluidized
bed dryer and compare them with the drying characteristics in the convective tray dryer. Furthermore, the Page
and the single exponential models are investigated on the experimental data in order to explore their validity for
corn drying in the fluidized bed. The effective diffusivities of corn are determined at each drying temperature for
both the fluidized bed dryer and the convective tray dryer. It is observed that M eq values are widely omitted in

989
literature due to the very small values they attain at high drying temperatures. In this work the equilibrium moisture
contents M eq are evaluated using the Guggenheim-Andersen-de Boer model. The effect of the incorporation of
the equilibrium moisture content in the calculations on the evaluated effective drying coefficient values is
investigated. In addition, an Arrhenius type relation is used to explain the dependence of the effective diffusivities
on temperature.

MATERIALS AND METHODS

A special dryer, with adjustable drying air velocity, has been constructed for the corn drying experiments.
Schematic diagram of the test apparatus is shown in Fig.1. The drying chamber is of cylindrical shape with an
inner diameter of 10 cm and a height of 50 cm. The bottom of the drying chamber is a perforated tray with holes of
2 mm in diameter. The chamber is constructed of two telescoped identical cylinders which make it possible to
insert and remove the inner cylinder. The drying air which enters the cylindrical chamber trough the perforated tray
is provided by a centrifugal fan, the speed of which is controlled by the Siemens Micromaster 6SE92 (7.5 kW)
frequency inverter. As shown in Fig.1, before entering the fan (Part 2), air is preheated by an electrical heater of
7.6 kW power capacity (part 1). After leaving the fan, the preheated drying air is heated to the desired temperature
by the main heater, which has an 8 kW power capacity (part 3). The power settings of the heater are adjusted for
each drying experiment at 55, 65 and 75°C to attain the required drying temperatures. A pitot tube and a
manometer are used (Part 4) to measure the drying air velocity. The temperature and the humidity of the air are
measured every 5 minutes. Measurements are performed with a Testo 454 multifunction instrument, Testo 905-T1
temperature measure sensor and Testo 605-H1 humidity measure sensor (Part 5). The moisture content of the
corn samples are measured by gravimetric methods with Sartorius MA 30 moisture analyzer. According to the
instrument specification, temperature and humidity measurements were performed with an accuracy of %0.3 for
the Testo 454-P1 instrument, 0.2% for the Testo 454-H1 instrument and with 0.05% for the Sartorius 30 moisture
analyzer.

Fig.1 Schematic diagram of fluidized bed dryer (1: Preheater, 2: Fan, 3: Main-heater, 4: Pitot tube, 5: Perforated plate, 6: Temperature and
humidity sensors).

Corn “Zea mays indentata” were provided from the Sakarya region of Turkey and were characterized as being of
spherical shape with an average diameter of 6.8 mm. All experiments were conducted under thin layer drying

990
 conditions. In each experiment corn samples of 40 gr in total weight were inserted into the drying chamber forming
a single layer of corn on the perforated tray. Two different drying air velocities were used in the experiments: 0.45
m/s and 2.5 m/s. For the air velocity of 0.45 m/s no fluidization was observed. Therefore, drying conditions
resembled those of convective tray dryers. For the drying air velocity of 2.5 m/s, the corn was brought into
fluidization state and fluidized bed dryer conditions were attained. At both drying air velocities, experiments were
conducted for 90 minutes at 3 different temperatures; 55, 65 and 75°C. The initial moisture of the corn was
measured by gravimetric methods with a Sartorius MA 30 moisture analyzer. The moisture contents during drying
were determined every 5 minutes by taking the inner cylinder out off the outer cylinder and measuring its weight on
a scale. The weight of the completely dried corn was determined to be 32 gr on average.

MATHEMATICAL MODEL
The Moisture Ratio (MR), the dimensionless moisture content, can be defined as

M(t) − M eq
MR(t) = (1)
M o − M eq

where M(t) denotes the moisture content at time instant t , M eq denotes the equilibrium moisture content and
M o denotes the initial moisture content of the corn.
Most of the food drying experiments performed in literature have shown that food drying occurs mainly in the
falling rate period. The falling rate period is the second phase of drying, in which the moisture transfer is driven
mainly by the moisture diffusion inside the corn particles. In order to model the thin layer drying kinetics in the
fluidized bed dryer and the convective tray dryer, drying will be assumed to occur only in the falling rate period.

The drying process is modeled using a single spherical shaped corn particle, representing any corn particle inside
the dryer. Mass diffusion inside the corn is assumed to take place only in the radial direction. Using Fick’s first law
defined for spherical coordinates and the mass balance, Fick’s second law for spherical coordinates is written as;

∂MR(t) 1 ∂ ∂MR(t)
= (Deff r ) (2)
∂t r ∂r ∂r

where Deff represents the effective diffusion coefficient inside the corn particle. Assuming diffusive moisture
transfer, negligible volume shrinkage, constant temperature and diffusion coefficients, an analytic solution to Eq.(2)
was proposed as follows (Madhiyanon, et al., 2009):

∞
Deff n 2 π 2 t
∑
6 1
=
MR(t) exp( − ) (3)
π2 n =1 n
2
r2

Considering only the first term (Ramesh et al., 2001), Eq.(3) is simplified into the following form:

6 Deff π 2
=
MR(t) exp( − t) (4)
π2 r2

Taking the natural logarithm of both sides we end up with the linear form Eq. (4):

6 Deff π 2
= ln
ln MR(t) + (− )t (5)
π2 r2
The theoretical form of Eq. (4) is referred to as the single exponential equation. In addition to this theoretical
model, plenty of empirical and semi-empirical models can be found in the literature (Diamante and Munro, 1993 ,
Madamba et al., 1996). Among these methods, the Page equation, the two term model, and the Henderson &
Pabis model were used for thin layer drying in fluidized bed for coconut (Madhiyanon, et al., 2009). The Page
model has already shown to be able to successfully determine the drying kinetics of corn in conventional tray dryer
(Doymaz and Pala, 2003) and is formulated as follows (Diamante and Munro, 1993 , Madamba, et al., 1996)

991
 = exp(− xt y )
MR(t) (6)

Therefore, to keep the work as simple as possible and in accord with the related previous works, only the Page
model and the single exponential model are used to determine the drying kinetics for both the fluidized and non-
fluidized corn drying, convective tray dryer. Effective diffusivities were obtained via the single exponential model.

The equilibrium moisture content M eq is defined as the average moisture content of corn at which no moisture
transfer takes place between the corn particle and the drying air. This equilibrium moisture content is highly
dependent upon the temperature and relative humidity of the drying air. Samapundo et al. (Samapundo et al.,
2007) have investigated the sorption isotherms of corn at 25, 30 and 37°C. They concluded that the equilibrium
moisture content could be best estimated by the Guggenheim-Andersen-de Boer (GAB ) model (Berg, 1985)

M m CKa w
M eq = (7)
[(1 − Ka w )(1 − Ka w + CKa w )]

Where a w is the water activity (equilibrium relative humidity of air), M m is the monolayer moisture content
(kg/100 kg dry matter), and C and K are constants. In their work it can be observed that their sorption isotherms
tend to come very close at low water activity values. Since water activity values in our experiments were quite
small, the parameters proposed by Samapundo et al (Samapundo, et al., 2007) for the GAB model at 37° C were
used to estimate the equilibrium moisture content at each drying temperature. Although the moisture content of the
drying air is the same at each drying temperature, it was observed that the equilibrium moisture content of corn
decreases with increasing drying temperatures. The reason for this decrease is the lower water activity values
attained at higher drying temperatures.
The relationship between the effective diffusivities Deff and the drying temperature T is provided by a simple
Arrhenius equation (Madamba, et al., 1996 , Ramesh, et al., 2001) as

E 
Deff = Do exp  a  (8)
 RT 

Here E a is the activation energy, or the energy required for drying reaction to occur at that specific drying
condition. Do is a pre-exponential factor and R is universal gas constant.

RESULTS AND DISCUSSION

The drying curves of corn obtained from the experiments performed in fluidized bed dryer and convective tray
dryer are shown in Fig.2 in terms of dry basis (db) moisture content versus time. Even a cursory comparison of the
curves in Fig.2 demonstrates significantly higher drying rates for the fluidized bed dryer compared to those for the
convective tray dryer at each drying temperature. For example, the moisture contents after 90 minutes of drying at
55°C in the fluidized bed dryer can be achieved by the convective tray dryer only if the drying temperature is
increased to 75°C. From the continuously decreasing slopes of the curves it can be inferred that falling drying
rates are observed which satisfies the assumption used to develop the mathematical model in the previous
section. In the falling rate phase it was concluded that drying moisture transfer is controlled by the interior
resistance to moisture diffusion (Keey, 1972). Therefore, Eq.(3), which was developed for the diffusive moisture
drying, can be applied to our experimental data.

992
Fig.2 Effect of fluidization on the drying curves at different temperatures.

The equilibrium moisture contents were evaluated using the GAB model with the parameters provided for 37°C by
Samapundo et al.(Samapundo, et al., 2007). The water activity values in Eq.(7) correspond to the measured
drying air humidity (Samapundo, et al., 2007). Using the water activity values measured in the experiments, the
equilibrium moisture content M eq was evaluated via Eq.(7) as 4.39% (db) at 55 °C, as 4.13% (db) at 65°C and as
3.43% for 75°C. Using these equilibrium moisture content M eq values, the moisture ratios (MR) were evaluated
via Eq.(1) and their natural logarithms are shown in Fig.3 with respect to drying time. In Fig.3 it is observed that the
slope of drying curves increase with elevated temperatures. A drastic increase in slope is also observed when
switching from conventional tray dryers to fluidized bed dryers. Recalling Eq.(5), larger slopes result in higher
effective diffusion coefficients and hence, in higher drying rates. The rate of drying can be directly linked to larger
effective diffusion coefficients which result in more homogeneous temperature distribution inside the corn.
Therefore, it can be concluded that the enhanced homogeneity of the corn’s temperature distribution in the
fluidized bed dryer induces higher drying rates than the convective tray dryer.
In accordance with the related previous work (Doymaz and Pala, 2003), the effective diffusivities M eq were first
assumed to be zero. Using Eq.(4) the effective diffusivities were evaluated from the slope of the linear portion of
the logarithmic drying curves (Goksu et al., 2005) at 55, 65 and 75°C. The results are listed and compared in
Table 1 with the Deff values obtained by I. Doymaz & M. Pala (Doymaz and Pala, 2003) for untreated corn dried in
cabinet dryers (convective tray dryer) at an air velocity of Vair = 0.43 m / s with zero equilibrium moisture content
assumption. It is observed that our results for an air velocity of Vair = 0.45 m / s are consistent with the work of I.
Doymaz & M. Pala (Doymaz and Pala, 2003) under the same assumptions. Moreover, the relative increase in the
effective diffusion coefficient ΔDeff , shown in Table 1, is defined as,

993
 Deff FBD − Deff CTD
ΔDeff = (8)
Deff CTD

where Deff CTD is the effective diffusion coefficient obtained in the conventional tray dryer and Deff FBD is the
effective diffusion coefficient obtained in the fluidized bed dryer. Have given the results assuming M eq to be zero,
we will now pursue our analysis by taking M eq into account. In that case, the equilibrium moisture content values
are not zero and take the values evaluated by Eq.(6). Hence, the effective drying coefficients are evaluated from
the slopes of the drying curves in Fig.3 and are presented in Table 2. Including the nonzero equilibrium moisture
contents in the calculations resulted in a substantial increase of the effective drying coefficient Deff values.
Furthermore, the relative increase in the effective diffusion coefficient ΔDeff induced due to fluidization, has
shown to rise from the range of % 50- 61 to the range of %71-74. When the equilibrium moisture content is
assumed to be non-zero, the effect of the fluidized dryer on the drying rates is more evident. In fact, the
equilibrium moisture content M eq is never zero and always has a specific value depending on the drying air
conditions. Therefore the correct detection of M eq values shows importance in the prediction of the correct drying
kinetics.

Table 1 Effective diffusion coefficients and relative increase in the effective diffusion coefficients evaluated with zero equilibrium moisture
content assumption.
Temperature (I. Doymaz & M. Pala) D (m 2 / s)
eff D (m 2 / s)
eff
ΔD (%)
eff
Deff (m 2 / s) Vair = 0.45m / s Vair = 2.5m / s
M eq = 0
M eq = 0 M eq = 0

55 °C
9.488 ⋅10−11 9.3128 ⋅10−11 1.4967 ⋅10−10 %61

65 °C
1.153 ⋅10−10 1.1020 ⋅10−10 1.7540 ⋅10−10 %59

75 °C
1.768 ⋅10−10 1.6510 ⋅10−10 2.4753 ⋅10−10 %50

994
Fig.3 Logarithmic drying curves for corn in conventional tray dryer (CTD) and fluidized bed dryer (FBD).

Table 2 Effective diffusion coefficients and relative increase in the effective diffusion coefficients evaluated with the equilibrium moisture
content evaluated by the Guggenheim-Andersen-de Boer model.
Temperature D (m 2 / s)
eff D (m 2 / s)
eff
ΔD (%)
eff

Vair = 0.45m / s Vair = 2.5m / s

M eq ≠ 0 M eq ≠ 0
55 °C
1.2017 ⋅10−10 2.0704 ⋅10−10 %72
65 °C
1.4208 ⋅10−10 2.4763 ⋅10−10 %74

75 °C
2.1140 ⋅10−10 3.6147 ⋅10−10 %71

In Fig.4 the moisture ratios are shown for the drying at air velocity of Vair = 0.45 m / s as squares whereas for
Vair = 2.5m / s they are shown as triangles. As shown in Fig.4, Page’s equation was fitted onto the drying data of
the convective tray dryer (dashed line) and the data of the fluidized bed dryer (solid line). Curve fitting was utilized
using the Ezyfit toolbox for Matlab to determine the constants of the best fitting Page’s equation. All coefficients of
determination R 2 of these fits were found to be larger than 0.99.

995
Fig.4 (a) Comparison of the Page model fitted for conventional tray dryer (CTD) and fluidized bed (FBD) dryer at 55°C. (b) Comparison of
the Page model fitted for conventional tray dryer (CTD) and fluidized bed (FBD) dryer at 65°C. (c) Comparison of the Page model fitted for
conventional tray dryer (CTD) and fluidized bed (FBD) dryer at 75°C.

To describe the temperature dependence of the effective diffusion coefficients the Arrhenius type relationship
indicated by Eq.(7) is used. For that purpose the calculated effective diffusivities were plotted in Fig.5 as a function
of the reciprocal of the absolute temperature. Using Eq.(7) and the slopes of the best fitting straight lines in Fig.5,
the activation energy E A was found to be 26.7 kJ/kg mol for drying in the conventional tray dryer and
26.4 kJ/kg mol for drying in the fluidized bed dryer. Both of these values are close to E A = 29.56 kJ/kg mol
which was proposed previously by I.Doymaz & M.Pala (Doymaz and Pala, 2003).

Fig.5 Relationship among effective diffusivity and the reciprocal of the drying temperature.

996
CONCLUSION
The experimental results have shown that corn drying proceeds only in the falling drying rate period where drying
is controlled only by the moisture diffusion inside of the corn particle. Moreover, from the experiments it was found
that the drying rates of corn in the fluidized bed dryer are substantially higher than the convective tray dryer at all
drying temperatures. The effective drying coefficients are increased by an average of %72 by the improved drying
−10
rates. The values of effective diffusivities were found to be in the range 2.1-3.6 ⋅10 m2 /s for the fluidized bed
−10
dryer and 1.2-2.1 ⋅10 m2/s for the conventional tray dryer. The main reason for this behavior is the
homogeneous mixing motion in the fluidized bed dryer, which enhances heat and mass transfer between corn and
air significantly. The Page model was able to fit the drying data of the conventional tray dryer and the fluidized bed
dryer successfully at all drying temperatures (55, 65 and 75ºC). Calculations revealed a significant dependence of
the evaluated effective drying coefficients on the equilibrium moisture contents. Setting equilibrium moisture
contents to zero decreases the numerical values of the effective diffusion coefficients. Therefore, the equilibrium
moisture is required to be determined correctly in order to evaluate the effective drying coefficients most
accurately. Moreover, effective diffusion coefficients have been shown to increase with higher drying temperatures
and pursued an Arrhenius type relationship. The activation energy E A was found to be 26.4kJ/kg mol for the
convective tray dryer and 26.7kJ/kg mol for the fluidized bed dryer. These values are consistent with the
activation energies proposed by I. Doymaz & M.Pala (Doymaz and Pala, 2003).

REFERENCES
[1] Doymaz, I. and M. Pala, Nov 2003. The thin-layer drying characteristics of corn, Journal of Food Engineering,
vol. 60, pp. 125-130.
[2] Oztekin, S., A. Bascetincelik, and Y. Soysal, Jan-Apr 1999. Crop drying programme in Turkey, Renewable
Energy, vol. 16, pp. 789-794.
[3] Wang, Z. H. and G. H. Chen, May 2000. Heat and mass transfer in batch fluidized-bed drying of porous
particles, Chemical Engineering Science, vol. 55, pp. 1857-1869.
[4] Tatemoto, Y., Y. Bando, K. Yasuda, Y. Senda, and M. Nakamura, 2001. Effect of fluidizing particle on drying
characteristics of porous material in fluidized bed, Drying Technology, vol. 19, pp. 1305-1318.
[5] Henderson, S. M. and S. Pabis, 1961. Grain drying theory I. Temperature effect on drying coefficient, Journal of
Agricultural Engineering Research, vol. 6, pp. 169-174.
[6] Wang, C. Y. and R. P. Singh, 1978. Use of variable equilibrium moisture content in modelling rice drying, ASAE
Paper No. 78– 6505,
[7] Zhang, Q. and J. B. Litchfield, 1991. An Optimization of Intermittent Corn Drying in a Laboratory Scale Thin-
Layer Dryer, Drying Technology, vol. 9, pp. 383-395.
[8] Liu, Q. A. and F. W. BakkerArkema, Apr 1997. Stochastic modelling of grain drying .2. Model development,
Journal of Agricultural Engineering Research, vol. 66, pp. 275-280.
[9] Yaldiz, O., C. Ertekin, and H. I. Uzun, May 2001. Mathematical modeling of thin layer solar drying of sultana
grapes, Energy, vol. 26, pp. 457-465.
[10] Overhults, D. D., G. M. White, M. E. Hamilton, and I. J. Ross, 1973. Drying soybeans with heated air,
Transactions of the ASAE, vol. 16, pp. 195-200.
[11] Li, H. Z. and R. V. Morey, 1984. Thin-Layer Drying of Yellow Dent Corn, Transactions of the Asae, vol. 27, pp.
581-585.
[12] Jumah, R. Y., A. S. Mujumdar, and G. S. V. Raghavan, Aug 1996. Batch drying kinetics of corn in a novel
rotating jet spouted bed, Canadian Journal of Chemical Engineering, vol. 74, pp. 479-486.
[13] Soponronnarit, S., A. Pongtornkulpanich, and S. Prachayawarakorn, 1997. Corn quality after drying by
fluidization technique at high temperature, Drying Technology, vol. 15, pp. 2577-2586.
[14] Soponronnarit, S., A. Pongtornkulpanich, and S. Prachayawarakorn, 1997. Drying characteristics of corn in
fluidized bed dryer, Drying Technology, vol. 15, pp. 1603-1615.

997
[15] Syahrul, S., I. Dincer, and F. Hamdullahpur, Jul 2003. Thermodynamic modeling of fluidized bed drying of moist
particles, International Journal of Thermal Sciences, vol. 42, pp. 691-701.
[16] Hatamipour, M. S. and D. Mowla, Sep 2006. Drying behaviour of maize and green peas immersed in fluidized
bed of inert energy carrier particles, Food and Bioproducts Processing, vol. 84, pp. 220-226.
[17] Madhiyanon, T., A. Phila, and S. Soponronnarit, Oct 2009. Models of fluidized bed drying for thin-layer chopped
coconut, Applied Thermal Engineering, vol. 29, pp. 2849-2854.
[18] Ramesh, M. N., W. Wolf, D. Tevini, and G. Jung, Jul 2001. Influence of processing parameters on the drying of
spice paprika, Journal of Food Engineering, vol. 49, pp. 63-72.
[19] Diamante, L. M. and P. A. Munro, Oct 1993. Mathematical-Modeling of the Thin-Layer Solar Drying of Sweet-
Potato Slices, Solar Energy, vol. 51, pp. 271-276.
[20] Madamba, P. S., R. H. Driscoll, and K. A. Buckle, Jul 1996. The thin-layer drying characteristics of garlic slices,
Journal of Food Engineering, vol. 29, pp. 75-97.
[21] Samapundo, S., F. Devlieghere, B. De Meulenaer, A. Atukwase, Y. Lamboni, and J. M. Debevere, Mar 2007.
Sorption isotherms and isosteric heats of sorption of whole yellow dent corn, Journal of Food Engineering, vol.
79, pp. 168-175.
[22] Berg, V. D., 1985. Development of B.E.T. like models for sorption of water of foods; theory and relevance, In D.
Simatos & J. L. Multon ( Eds.) , Properties of water in foods, pp. 119– 135.
[23] Keey, R. B., Drying principles and practice, [1st ed. Oxford, New York,: Pergamon Press, 1972.
[24] Goksu, E. I., G. Sumnu, and A. Esin, Feb 2005. Effect of microwave on fluidized bed drying of macaroni beads,
Journal of Food Engineering, vol. 66, pp. 463-468.

998
 MATHEMATICAL MODELING OF A SMALL SCALE COMPRESSED AIR
ENERGY STORAGE SYSTEM

M. Kiliç* Z. K. Kocabiçak** E. E. Topçu* and M. Mutlu*

Uludağ University, Faculty of Engineering& Architecture

Department of Mechanical Engineering* Department of Automotive Engineering **
mkilic@uludag.edu.tr; zkamis@uludag.edu.tr; erzan@uludag.edu.tr; mustafamutlu@uludag.edu.tr

ABSTRACT
Using compressed air to store energy is one of the energy storage methods. In this study, a small scale
compressed air energy storage (CAES) system is designed and modeled. The energy storage capacity of
designed CAES system is about 2 kWh. The system contains a hydraulic pump unit, expansion-compression
liquid pistons, valves, a tank and a control unit. The aim of the designed system is basically to store air under
a defined pressure. The designed CAES system is modeled and simulated by MATLAB/Simulink program.
Pressure changes in the tank and pistons are obtained. Besides, energy storage capacity of the system for
different pressures is investigated in isothermal conditions.

Keywords: Energy storage, CAES, Hydro-pneumatic system, MATLAB/Simulink modeling

INTRODUCTION
The development and use of renewable energy has shown a rapid growth over the past few years. In the
next 20–30 years, all sustainable energy systems will have to be based on the rational use of traditional
resources and greater use of renewable energy. Renewable energy sources are plentiful and conversion
systems are becoming more and more affordable. Their significant contribution to sustainable energy use will
however require considerable further development of storage methods. This will open up a new field of
application, especially due to the growth of electrical production from renewable energy, along with
decentralized production. Thus, the growth of this decentralized production means greater network load
stability problems and requires energy storage.

There are various energy storage systems technologically. These systems are classified generally according
to the technology, size of power and energy and application area. Beside the high-capacity storage facilities
based on hydro-power technologies, electrochemical solutions are the today's candidate to store for
renewable energy sources. However, limited life-cycles and sustainability of batteries are often inhibiting
factors. Compressed air technology is another renewable energy storage system which has many
advantages. They can be summarized that air is available everywhere, compressed air can be stored for
long time in a tank and pneumatic energy conversion doesn’t produce waste or greenhouse emissions.

There are many studies in the literature on energy storage. Ibrahim et al. (2008) comparatively investigated
the basic characteristics and application areas of different energy storage techniques and analyzed to
determine the most appropriate technique. Energy networks for the electrical energy storage systems are
investigated by Kondoh et al. (2000). In this study, the characteristics of energy storage systems are
examined and hydraulic, compressed air, secondary batteries, super-conducting magnets, flywheels or
capacitors are investigated. Lund and Salgi (2009a) discussed the compressed air energy storage systems
in their studies. Lund and Salgi (2009b) describe three different computer-aided methods to determine the
optimal operating strategy for a given CAES device. Glendenning (1976) examined long-term projections for
compressed air storage systems. Denholm and Kulcinski (2004) investigated greenhouse gas emissions of
large-scale energy storage systems and the energy requirements for their life cycle. The use of compressed
air energy storage systems working wind energy were discussed in several studies (Denholm (2006), Swider
(2007)).

Grazzini et al. (2008) made thermodynamic analysis for a CAES system working under adiabatic condition.
Baker (2008) examined the developments in energy storage systems and worked on electrochemical,
mechanical, and thermal energy storage systems. Milazzo (2008) investigated CAES system with thermal
energy storage (CAES-TES) for the compression stage. Using the second law of thermodynamics,
thermodynamic analysis of a compressor unit is examined. Rufer and Lemofouet (2006) studies on a long-
lasting, hybrid energy storage system based on compressed air energy. In the study, the general concept of
the system, the power-electronic interface circuits, the maximum efficiency point tracking algorithm and the

999
output power strategy are described. In addition, the principles of energy storage system with hydro-
pneumatic transducer are defined. Van de Ven and Li (2009) are investigated liquid piston gas compression
systems in detail.

Although, a number of study about the CAES can be seen in the current literature, there is not too much
study and applications present in the context of a standalone renewable energy storage with the CAES
system for domestic applications. This study aims to a detailed investigation of a small scale energy storage
system with high life cycle, based on Compressed Air Energy Storage (CAES). A mathematical model is
constructed for the designed small scale compressed air energy storage system and simulated by
MATLAB/Simulink program. In order to obtain maximum efficiency, it is assumed that air is compressed in
isothermal conditions. Pressure changes in pistons and the tank is investigated. The method presented and
the results obtained in this study will be useful for the further studies and the design and construction of the
small scale CAES systems.

THE DESIGNED CAES SYSTEM

The schematic view of the designed system is shown in Figure 1. The system consists of two liquid pistons,
four electro-pneumatic valves, a tank, a hydraulic pump, an electric motor and a control unit basically.

In the system, the compression and expansion processes occur in reciprocating liquid piston working
chambers. One liquid piston is sucking the atmospheric air while the other is sending the compressed air to
the tank such as internal combustion engines. During this process the tank pressure increases by sending
the compressed air to the tank. The greater stored air in the tank means more pressure and also more
energy. The amount of energy stored in the system can be improved by increasing tank pressure or tank
volume.

The hydraulic pump is used to supply the flow of the liquid pistons and to set the working speed of the
system. The speed of the system can be changed according to the liquid piston pressure can be necessary
to save time and to eliminate overheating. This working principle can be provided with the adjustment of the
geometric displacement or speed of pump. Due to developments in electronic technology in recent years,
obtaining variable flow rate by changing electric motor speed can be possible easily. In this system, a
variable speed pump is selected to adjust the speed of liquid pistons and to direct hydraulic oil in both
directions without using any directional control valves. Thus, partial energy loses arising from friction and
throttle losses at the directional control valves are eliminated and energy recovery is increased.

The air flow in liquid pistons is controlled by electro-pneumatic valves. Two of them (B,C) are used to send
the compresses air to tank, the others (A,D) are used to suck the atmospheric air. The valves are controlled
by the computer control unit.

Fig. 1. Designed CAES system

1000
MATHEMATICAL MODELING OF THE SYSTEM
In the designed system the energy storage capacity of the designed CAES system is defined about 2 kWh.
Liquid piston diameter (D), length and dead length (L, Ldead) is determined respectively 0.2 m, 1.1 m and 0.05
3
m. The air tank capacity (Vtank) is 0,5 m . The equations used in system design and modeling are given
below.

Stored air energy per unit mass depending on the air pressure can be calculated from Eq. (1).

Ptan k
W  RT ln (1)
Patm

Air mass stored in the tank can be calculated from ideal gas equation.

Ptan k Vtan k  mRT (2)

According to the Eq. (2), pressure change in the tank can be defined as;

dPtan k RT dm dV
  Ptan k tan k (3)
dt Vtan k dt dt

With the acceptance of isothermal working conditions, the relationship between pressure and volume can be
given by Eq. (4);

P1V1  P2 V2 (4)

The mean strokes of liquid pistons to generate the desired pressure in the tank can be calculated from Eq.
(5);

Ptan k Vtan k
n  (5)
Patm  V  Vdead 

The stroke numbers to create the desired pressure in the tank is given in Table 1. Energy stored in the tank
per unit mass and the total stored energy for different pressures, which is calculated from above equations,
is given Table 1.

Table1. Energy values stored in the tank for different pressures

Pressure in the tank Stored energy per unit mass Stroke Total stored energy
Ptank [bar] W [kJ/kg] n E [kWh]
10 193.626 148 0.32
20 251.914 290 0.83
30 286.010 442 1.40
40 310.201 595 2.04
50 328.965 738 2.70

Working speed of the pistons can be arranged by changing the flow rate of the pump as mentioned earlier.
The specifications of the selected pump are given in Table 2. The pump flow rate is given below;

Q=Dp  n pump  ηv (6)

Table 2. Hydraulic pump characteristics of the system

Dp M Pmax nmin nmax Qmin Qmax
3
[cm /dev] [Nm] [bar] [dev/dk] [dev/dk] [lt/dk] [lt/dk]
20 50 160 150 3200 2.8 60.8

1001
MATLAB/SIMULINK MODEL OF THE SYSTEM
Compressed air storage system is modeled by MATLAB/Simulink program for isothermal condition. The flow
diagram of the system is shown in Fig. 2.

Fig. 2. The flow diagram of the model

Both of pistons are used for compressing and expanding phases. As one of them compresses the air, the
other expands and sucks the air from atmosphere. When compressing piston pressure is reached to the tank
pressure, the valve (B or C) is opened, compressed air is sent to the tank and the tank pressure is increased.
In the expanding piston, the valve (A or D) is opened through the stroke and liquid piston’s pressure drops to
atmospheric pressure.

The MATLAB/Simulink model given in Fig. 3 calculates the pressures in the tank and liquid pistons. Some
variables such as the liquid level, the amount of mass sending the tank, liquid piston speed, etc. can also be
obtained by the model.

The liquid levels of both liquid pistons are determined by “n_x1_x2” submodel (Fig. 4). The air mass is
calculated by the “mass_calculation” submodel (Fig. 5). Tank pressure and liquid pistons pressures are
calculated by respectively “tank_pressure” and “liquid pistons_pressures” submodels (Fig. 6, 7).

1002
Fig. 3. Simulink model of the system

Fig. 4. “n_x1_x2” submodel

Fig. 5. “mass_calculation” submodel

1003
Fig. 6. “tank_pressure” submodel

Fig. 7. “liquid pistons_pressures” submodel

RESULTS
Stored energy according to the tank pressure can be shown from Fig. 8. As shown in this figure, 2.7 kWh
energy can be stored in the tank at 50 bar. Pressure changes in liquid pistons and tank are given in Fig. 9.
Compressing piston’s pressure increases as the expanding piston’s pressure drops to atmospheric pressure,
because air is compressed and sent to the tank by compressing piston. Until the valve is opened, the
compressing piston pressure increases fast. In addition, pressure increasing in the piston and the tank are
the same but the rate of increasing decreases after opening the valve. Tank pressure remains constant
when the valve is closed. The pressures of the tank and pistons reach to the 15 bar in 10000 s at 3200 rpm.
If the time is extended, more air mass will be stored in the tank. Consequently, the tank and piston pressures
and stored energy will be raised.

2.5
Stored energy (kWh)

1.5

0.5

0
10 15 20 25 30 35 40 45 50
Pressure in the tank (bar)

Fig. 8. Tank pressure- stored energy variations

1004
The air mass sending to the tank is shown in Fig. 10. Air mass sucked by expanding piston at one stroke is
approximately 40 gr. But the air mass changes as long as the system is running because of dead volume.
As the tank pressure increases, the amount of mass in the dead volume is also increases and less mass is
sent to the tank. The air mass in the tank is 9 kg in 10000 s at 3200 rpm (Fig. 10).
The tank pressure changes and energy variations for different pump speeds are given in Fig. 11 and Fig. 12.
As shown in these figures, tank pressure and the stored energy values will increase faster if pump speed
increases.

6 Liquid Piston 1
x 10
2
Pressure (N/m2)

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
5
x 10
2.5
Pressure (N/m2)

1.5

1
0 50 100 150 200 250 300 350 400 450 500
Time (s)

6 Liquid Piston 2
x 10
2
Pressure (N/m2)

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
5
x 10
2.5
Pressure (N/m2)

1.5

1
0 50 100 150 200 250 300 350 400 450 500
Time (s)

6 Tank
x 10
2
Pressure (N/m2)

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Time (s)
Fig. 9. Pressure changes in tank and pistons

1005
 10
Mass (kg)

7.5
5
2.5
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

0.4
Mass (kg)

0.3
0.2
0.1
0
0 50 100 150 200 250
Time (s)
Fig. 10. Air mass changes in tank

6
x 10
2
n=1000 rpm
n=2000 rpm
1.8
n=3000 rpm
n=4000 rpm
1.6

1.4
Tank Pressure (N/m2)

1.2

0.8

0.6

0.4

0.2

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Time (s)
Fig. 11. Tank pressures in different pump speeds

0.7
n=1000 rpm
n=2000 rpm
n=3000 rpm
0.6
n=4000 rpm

0.5
Stored Energy (kWh)

0.4

0.3

0.2

0.1

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Time (s)
Fig. 12. Stored energy in different pump speeds

1006
CONCLUSIONS
In this study, the design and modeling of a small scale compressed air energy storage system has been
examined. The system is modeled by MATLAB/Simulink program. Isothermal conditions are taken into
account in the modeling and modeling stages are given in detail. Pressure changes in the tank and pistons
are obtained.

In this system, it is determined that 2.7 kWh energy can be stored in the tank at 50 bar. The pressures of the
tank and pistons reach to the 15 bar in 10000 s at 3200 rpm. The air mass and stored energy in the tank is 9
kg and 0.55 kWh respectively at the same period and speed. If the time is extended or speed of the pump is
raised, more air mass will be stored in the tank. In addition to these, the storage capacity can be easily
expanded by adding more storage tank without changing main compression system. Sealing the storage
tanks makes that the stored energy can be hold without any time limitation.

Although, the present study presents only isothermal compression, the mathematical model can be used for
the polytrophic or adiabatic compression processes. In the advanced stages of the study, the model will be
expanded with the evaluation of feedback control system.

The method presented and the results obtained in this study will be useful for the further studies and the
design and construction of the small scale CAES systems.

Acknowledgment
This project is supported by Scientific and Technological Research Council of Turkey, under project number:
111M015. Authors thank to the Scientific and Technological Research Council of Turkey.

NOMENCLATURE
R universal gas constant (J/kgK)
T air temperature (K)
2
Patm atmospheric pressure (N/m )
2
Ptank tank pressure (N/m )
3
Vtank tank volume (m )
3
Q flow rate (m /s)
3
Dp pump displacement (m /rev)
npump pump speed (rev/min)
3
V liquid piston volume (m )
3
Vdead dead volume (m )
D the diameter of liquid piston (m)
L the length of liquid piston (m)
Ldead the dead length of liquid piston (m)
D 2
A area of liquid piston (= )
4
ηv volumetric efficiency

REFERENCES
Ibrahim, H., A. Ilinca and J. Perron. 2008. Energy storage systems–Characteristics and comparisons.
Renewable and Sustainable Energy Reviews. 12: 1221–1250.

Kondoh, J., I. Ishii, H. Yamaguchi, A. Murata, K. Otani, K. Sakuta, N. Higuchi, S. Sekine and M. Kamimoto.
2000. Electrical energy storage systems for energy networks, Energy Conversion & Management. 41:1863-
1874.

Lund, H. and G. Salgi. 2009. The role of compressed air energy storage (CAES) in future sustainable energy
systems. Energy Conversion and Management. 50:1172–1179.

Glendenning, I. 1976. Long-term prospects for compressed air storage. Applied Energy. 2(1): 39–56.

Denholm, P. and G.L. Kulcinski. 2004. Life cycle energy requirements and greenhouse gas emissions from
large scale energy storage systems. Conversion and Management. 45(13–14): 2153–2172.

1007
P. Denholm, P. Improving the technical, environmental and social performance of wind energy systems
using biomass-based energy storage. Renewable Energy 31(9), 1355–70, 2006.

DJ. Swider, Compressed air energy storage in an electricity system with significant wind power generation.
IEEE Trans Energy Convers 22(1), 95–102, 2007.

H. Lund, G. Salgi, B. Elmegaard, A. N. Andersen, Optimal operation strategies of compressed air energy
storage (CAES) on electricity spot markets with fluctuating prices. Appl Therm Eng 29(59), 799–806, 2009.

G. Grazzini, A. Milazzo, Thermodynamic analysis of CAES/TES systems for renewable energy plants.
Renewable Energy 33, 1998–2006, 2008.

J. Baker, New technology and possible advances in energy storage. Energy Policy 36, 4368–4373, 2008.

Milazzo, A., Optimization of the configuration in a CAES-TES system. In: Shape and Thermodynamics,
International Workshop Florence 2008, 25-26 September 2008, Florence, 2008.

S. Lemofouet, A. Rufer, Hybrid energy sorage systems based on compressed air and supercapasitors with
maximum efficiency point tracking, IEEE 4, 1105 – 1115, 2006.

Cyphelly, I, A. Rufer, Ph. Brückmann, W. Menhardt, A. Reller, Usage of compressed air storage systems;
Swiss Federal Office of Energy, DIS Project Nr. 100406, May 2004. Publ. Nr: 240050.

Van de Ven, J.D., P. Li. 2009. “Liquid Piston Gas Compression”. Applied Energy. Vol. 86. pp:2183-2191.

1008
 INVESTIGATION OF THE USE OF SOLAR THERMAL BUFFER ZONE IN BUILDINGS
a a* b c
A. Jan , M. S. Hamed , G. Razaqpur , and S.Foo
a
Thermal Processing Laboratory (TPL), Department of Mechanical Engineering,
McMaster University, Hamilton, Canada
b
Department of Civil Engineering, McMaster University, Hamilton, Canada
c
Public Works and Government Services Canada (PWGSC), Ottawa, Canada

ABSTRACT
This paper presents results of a collaborative study that is being carried out by the Thermal Processing
Laboratory (TPL), the Department of Civil Engineering and Public Works and Government Services Canada
(PWGSC). The main objective of this study is to investigate the feasibility of passive means to achieve Net-Zero
Energy (NZE) in federal buildings in Canada.

An experimental and numerical investigation of the fluid flow and heat transfer inside a Solar Thermal Buffer
Zone (TBZ) has been carried out. The numerical work has been conducted using the ANSYS-CFX commercial
software package. A TBZ is an air-filled cavity envelopes the building. It acts as an additional insulation and
allows solar energy collected from the south side of the building to be distributed throughout the other sides,
hence, reducing the heating load of the building. The main objective of the current study was to determine the
optimum TBZ size.

Considering a one-third model of a typical floor in a building located in Ottawa, Ontario, Canada, our results
obtained using an average winter day solar intensity, show that the TBZ could result in a significant saving in the
building heating load, which makes the TBZ an attractive passive technology to achieve significant sustainable
energy savings in commercial buildings.

INTRODUCTION
One of the biggest problems our society will face in the coming years pertains to energy conservation and
management. Our population is growing at an increasing rate while our planet’s resources are slowly
diminishing. This is creating an increase in demand for more efficient, clean and more sustainable methods of
energy production, consumption and possible storage. In the past few decades we have witnessed an enormous
interest in the impact of human activities on the environment. Despite the fact that fossil fuel use has an adverse
effect on the environment, fossil fuels still contribute significantly towards energy production in developed
countries. As per the Natural Resources Canada’s Energy Use Data Handbook (2009), buildings are known to
use more than 40% of primary energy used in North America. The need for Sustainable development has been
on the increase in the recent past.

A large portion of energy is spent every year on space heating and cooling. In Canada, this is of great concern
due to very harsh, varying climate which often reaches extremes from year to year. Table 1 below shows that the
most amount of energy used is buildings is used for space heating.

Table 1 Total Energy Use in buildings (Natural Resources Canada, 2009)

Total Energy Use (%) 1990 2005 2006 2007 2008 2009
Space Heating 61.8 61.3 60.3 62.8 63 62.8
Water Heating 19.1 18.1 18.8 17.9 17.5 17.3
Appliances 14.3 13.8 14.4 13.4 13.9 14.4
Lighting 4 4.4 4.5 4.1 4.2 4.3
Space Cooling 0.8 2.5 2 1.9 1.5 1.2

In the 1970’s and 80’s, the aim of net zero energy buildings (NZEB) was taken into consideration earnestly due
to the rise in oil prices. NZEB is the reduction of energy obtained from fossil fuels in building operations and
maintenance. It can be accomplished by taking conservation measures or replacing fossil fuels with renewable

1009
energy sources. The main goal of the present study was to reduce the amount of non-renewable energy used for
space heating in buildings by using a passive means utilizing naturally present solar energy; namely the TBZ. To
the best of the authors’ knowledge, the present experimental approach investigating the use of TBZ in buildings
is unique.

SOLAR THERMAL BUFFER ZONE:

Due to the orientation of the sun, a building receives an uneven amount of solar gain (solar radiation energy).
The south facing side of a building receives more energy than the north facing side, which creates an uneven
temperature distribution and significant heating and cooling loads. For example, even in the winter, sometimes
the north side of the building requires heating while the south side requires cooling. A better distribution of the
solar gain could eliminate such problems. This can be done by creating a thermal buffer zone (TBZ) around the
building. A TBZ, as shown in Fig.1, consists of two layers of glass separated by an air cavity. Due to the higher
solar gain on the south-facing cavity, a density difference between the two columns of air causes an air flow
around the building envelope. The strength of such thermally-driven flow depends on the amount of solar gain,
the envelope size, and the material properties. The main objective of this study is to investigate the effect of the
cavity size on the amount of heat that can be distributed from the south side to the north side of the building.

Fig. 1 Solar Thermal Buffer Zone

LITERATURE REVIEW
A solar thermosiphon is a differentially heated open cavity. Therefore, research related to predicting performance
of solar thermosiphons can be traced back to work of natural convection over a single heated vertical plate and
flow between a differentially heated parallel plate open cavities. The air flow in solar chimneys, Trombe wall, and
airflow windows, has often been treated as separate subjects. However, the basic assumption of modelling them
as flow between parallel plates is universal in all the research done on these methods. This discussion will treat
all three methods of buoyancy induced flow under the same heading as flow between heated vertical parallel
plates. Experimental and numerical modelling of the heated vertical parallel plate channel has been carried out
extensively and the flow within the channel is well understood, Gan 2006.

There has been no research work done previously on a thermal buffer zone with a convection loop, i.e., the
ability to move the hot air from the south side around the building, passively, (R. Richman, 2010). The closest
previous work to the present configuration is the work of (Richman R. C., 2009) on the Solar Dynamic Buffer
Zone (SDBZ). The SDBZ is a curtain wall system which “works by ventilating a cavity within a wall with heated
exterior air to control moisture migration across the assembly”. The SDBZ is also employed within the spandrel
area just like the Solar Air Flow Window. During the heating season, the model used by (Richman R. C., 2009)
predicted an average overall seasonal efficiency of 35%. His experimental results showed that the SDBZ curtain
wall is an effective means of collecting solar energy in a relatively passive manner with experimental efficiency of
25 – 30% on average. However, connecting the produced air to the main air conditioning duct system is the
main practical restraint, similar to the case of solar air windows.

Another research that is similar to a thermal buffer zone is the concept of double facades in a building. A double
skin façade is a special type of envelope where a second skin, usually transparent glazing, is placed in front of
the regular building façade. The air space in between, the channel, is ventilated (naturally or mechanically) in

1010
order to diminish overheating problems in summer and to contribute to energy saving in winter, (M.A. Shameri,
2011). It has been installed and monitored for at least one year at the Siemens building in Dortmund, Victoria.
The research was carried out at the University of Dortmund. The solar gains of the permanently ventilate facade
saved approximately 15-18% of the heating load in winter, (Pasquay, 2004). The average temperature in the
façade was 10-15 degrees higher than the outside temperature. One of the drawbacks of this system is that
wind speed and wind direction have a significant influence on the ventilation in the façade gap, (Pasquay,
2004). The use of double skin facade (DSF) has also been investigated (DeHerde, 2004) and (Saelens, 2002).
Additional cost of the DSF above typical static façade system have ranged significantly from 20% to about 300%
(Arons, 2000).

THE EXPERIMENTAL SETUP

The experimental apparatus used to carry out the present study is shown in figure 2. The apparatus was
designed to investigate the effect of the TBZ gap aspect ratio, B/H and the radiation intensity on the
effectiveness of the TBZ. Only the effect of B/H is reported in this paper. The optical properties of the system
can also be changed by changing the type of glass.

Fig. 2 Experimental Setup

The heat flux was varied by changing the distance between the cavity and the solar radiation simulator source,
L. The constant heat flux used in this paper corresponds to the incident solar radiation on a south-facing window
2
in Ottawa, Canada, during the month of January. The daily average heat flux is about 380 W/m calculated
based on the standard solar resource calculations, Duffie and Beckman (1974). This value of the heat flux was
2
achieved with L = 1.38 m. The illumination system could cover a 1 m area with a maximum average radiation
2
intensity of 1000 W/m . The light source consisted of four Quarts Tungsten Halogen (QTH) light bulbs, which
o o
give a black-body spectral properties of 3000 K to 3300 K.

A custom structural frame made of softwood lumber was used to hold the various components in place to
replicate a solar thermal buffer zone. A 38.1 mm thick Expanded Polystyrene board, R-7.75, with reflective
radiant barrier facing the outside of the setup was used as insulation to reduce heat losses from the surfaces
exposed to the ambient. The air conditioned space was a 1 x 1 x 1 m cube. The height of the upper, lower, and
right gaps was kept constant at 0.1 m. The south-faced gap width, B, was varied between 0.1 and 0.7 m, which
gave an aspect ratio, B/H, of 0.1-0.7.

In order to better collect the incident solar radiation from the heat source, it was necessary for the glass used in
the physical prototype to have high absorbance ratio. The outer glass was a standard 6 mm thick clear window

1011
glass, which has a high solar absorbance ratio of 0.83. A solar shield grey glass was used for the inner glass
surface, which has lower absorbance coefficient of 0.54. Holes were drilled through the inner glass to insert the
velocity probe into the air gap at the three locations indicated by high, mid, and low in Fig. 2.

Radiation intensity was measured using a Hukseflux LP02, ISO second class, pyrometer. The unit’s accurate
operational wave length range is 305-2800 nm, which contains the majority of the energy for the solar spectrum.
The average incident radiation was calculated from measurements taken over a two-dimensional, 1 x 1 m,
square grid with 9 cm spacing between measurement locations.

Temperatures were measured using a number of 0.5 mm, Type-T, thermocouples placed at various locations,
shown in Fig. 3. Thermocouples were attached to surfaces using a clear tape and shielded from direct radiation
using tin foil. The laboratory air temperature was measured using 5 thermocouples.

Fig.3: Locations of temperature measurements

Air velocity and temperature were measured at various locations as shown in Fig. 4, using a ComfortSense hot
wire anemometer

1012
Fig. 4: Locations of air velocity and temperature measurements.

NUMERICAL INVESTIGATION
The commercial CFD software program ANSYS-CFX was used to model the flow and heat transfer in a two-
dimensional model of the TBZ. The computational domain is shown in Figure 4.
Fluid
Domain
E

Location 4

Location 3 F
Glass
Domain
B Room C
Location 2 o Location 5
A 22 C D

Location 1
G

H
Fig. 5: Numerical computational domain with labelling of boundaries and locations of measurements

Turbulence has been modeled suing the Shear Stress Transport (SST) model developed by Menter (1994).

RESULTS AND DISCUSSION

The effect of the gap aspect ratio, B/H, on the air velocity inside the south-facing gap measured at location 3
(see Fig. 5) is shown in Fig. 6. The air velocity has been non-dimensionalized using the average mass flow
velocity inside the gap. Experiments 1, 2, 3, and 4 have carried out using B/H values of 0.089, 0.18, 0.25, and
0.34, respectively.

1013
 0,14 Experiment 1 Experiment 2
Experiment 3 Experiment 4

0,12

0,1
Velocity (V)

0,08

0,06

0,04

0,02

0
0 0,2 0,4 0,6 0,8 1
Dimensionless Distance from Wall (x/B)

Fig. 6. Variation of dimensionless velocity inside south-facing gap as function of gap aspect ratio, B/H.
Measurements taken at Location 3 (see Fig. 5)

The variation of the mass flow rate as a function of B/H is shown in figure 7. Following the expected trend, the
mass flow rate increased with increasing the aspect ratio up to B/H = 0.25. Increasing B/H to 0.34, experiment 4,
resulted in a sharp decrease in the mass flow rate. The mass flow rate calculated in the north-facing gap showed
a similar trend; however it had lower values, due to leakage and measurements uncertainty. The significant
decrease in the mass flow rate observed in experiment 4 was attributed to high recirculation developed inside
the south-facing gap whereby most of the hot air stayed in the south-side and did not make it to the north-side of
the TBZ. Such recirculation is evident from the numerically obtained velocity fields for B/H = 0.089 and 0.34
shown in Fig. 8.
0,018

0,016 Experiments 1,2,3,4

- South Side
0,014 Experiments 1,2,3,4
- North Side
Mass Flow (Kg s-1 m-1)

0,012

0,010

0,008

0,006

0,004

0,002

0,000
0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4
B/H

Fig. 7 Mass flow rate per m width vs. B/H aspect ratio

1014
 (a) B/H = 0.089 (b) B/H = 0.34

Fig. 8. Numerical velocity fields obtained for experiments 1 and 4, B/H = 0.089 and 0.34.

CASE STUDY
In order to assess the effect of suing the TBZ on the energy usage of a typical building, a case study of
two offices was considered; one office is located on the south-side of the building and the other office is located
2
on the north-side, see Fig. 9. Each office is represented by a square room, having an area of 1 m . The cooling
and heating load of each room was estimated using temperatures obtained from the present experimental
results and the optical properties from the 2009 ASHARE Handbook.

The representative day for the case study was chosen to be in January. As per ASHARE Fundamentals
o
Handbook (2009), the average temperature in Ottawa in the month of January is 0 C. Assuming no humidity,
o 2
the conditioned space was kept at 21 C, and an average daily heat flux of 380 W/m was applied on the south-
facing side. In order to show the benefits of the TBZ, calculations must be compared with the conventional
building without a TBZ. For a building with no TBZ, all the calculations were based on the optical properties
obtained from the 2009 ASHRAE handbook.

Fig. 9: A floor in a building with offices on South and North side

The optimum B/H aspect ratio of 0.25, as determined experimentally, was chosen for this case study. Table 2
shows the benefit of the TBZ. Having a TBZ reduced the cooling load of the south-facing office by 70%.
Similarly, the TBZ reduced the heating load of the north-facing office by about 30%.

1015
 Table 2 Cooling and heating loads estimated with and without a TBZ

NO TBZ TBZ
South Side
2 2
Cooling (W/m ) Cooling (W/m )
Single glazing 203 46
Double glazing 195 42
Double glazing, low e coating 184
Triple glazing 204
Triple glazing, low e coating 184
North Side
2 2
Heating (W/m ) Heating (W/m )
Single glazing 105 72
Double glazing 72 54
Double glazing, low e coating 63
Triple glazing 55
Triple glazing, low e coating 44

SUMMARY AND CONCLUSION

An investigation of the benefits of using the thermal buffer zone in buildings has been carried out. From the
literature review, it was found that there were no comprehensive correlations accounting for the various
parameters that affect the performance of the TBZ. These correlations are needed for performing a year-round
building energy simulation that would estimate the benefits of the TBZ. The effect of the gap aspect ratio has
been investigated. Four aspect ratios were investigated experimentally. Results showed that the aspect ratio of
0.25 gave the maximum mass flow rate. The experimental results were used in a case study considering cooling
and heating loads of an office building located in Ottawa. Results obtained for the case study showed significant
potential of the TBZ in reducing building loads.

ACKNOWLEDGMENT

The authors would like to acknowledge the technical and financial support received from Public Works and
Government Services Canada.

NOMENCLATURE
SYMBOL DESCRIPTION UNITS
Specific heat capacity
Spacing between glass and absorbing wall
Measured voltage
Gravitational acceleration
Grashof number based on wall heating
Grashof number based on air temperature
Convection heat transfer coefficient
Total cavity height
Thermal conductivity
Spacing between glass and illumination system
Mass flow rate

1016
 Nusselt number based on height
Measured atmospheric pressure
Prandtl number
Heat flux
Rayleigh number based on height
Reynolds number
Uniform temperature of wall
Temperature of fluid far from wall
Velocity
Area averaged velocity
Thermal diffusivity
Thermal expansion coefficient
Kinematic viscosity
Dynamic viscosity
Density

REFERENCES

1. Arons, D. (2000). Properties and Applications of Double-Skin Building. Massachusetts Institute of

Technology.
2. DeHerde, C. a. (2004). Double Skin facades for Office Buildings. LUND University.
3. Friedrich, K. (2011). Experimental and Numerical Investogation of Solar Air Flow Window, Masters
Thesis, McMaster University.
4. Gan, G. H. (1998) A parametric study of Trombe walls for passive cooling of buildings. Energy and
Buildings, 27, 37-43.
5. M.A. Shameri, M. A. (2011). Perspectives of double skin facade systems in buildings and energy
savings. Renewable and Sustainable Reviews , 1468-1475.
6. Natural Resources Canada. (2009, April 1). Comprehensive Energy Use Database. Retrieved October
20. 2011, from Natural Resources Canada:
7. http://oee.nrcan.gc.ca/corporate/statistics/neud/dpa/comprehensive_tables/index.cfm?fuseaction=Select
or.showTree
8. Pasquay, T. (2004). Natural ventilation in high-rise buildings with double facades, saving or wasting of
energy. Energy and Buildings , 381-389.
9. R. Richman, K. P. (2010). Quantifying and predicting performance of the solar dynamic buffer zone
(SDBZ) curtain wall thorugh experimentation and numerical modeling. Energy and Buildings , 522 - 533.
10. Richman R. C., K. (2009). A more sustainable curtain wall system: Analytical modeling of the solar
dynamic buffer zone (SDBZ) curtain wall. Buildings and Enviornment , 1-10.
11. Saelens, D. (2002). Energy Performance assesment of single story multiple-skin facades.

1017
 PERFORMANCE EVALUATION FOR DIFFERENT CONFIGURATED HRSGs
a,b b
Uludag , Ozcelik
a
Ege Gaz A.Ş. Aliağa LNG Terminal, Izmir/TURKEY
b
Ege University, Faculty of Engineering, Department of Chemical Engineering,
Bornova- Izmir, TURKEY
onuruludağ81@gmail.com, zehra.ozcelik@ege.edu.tr

ABSTRACT

In this study, performances decreasing heat losses and plant costs are compared in different stage HRSG
with inlet parameters of under constuction power plant. Flue gas that occured after combustion in gas turbine
were passed in 1,2 and 3 stage HRSG, thermal changings, produced energies, type of equipments and
effects to environment were calculated and compared when energy and exergy analyses are applied using
ASPEN Plus simulation program. Result of these calculations and comparisons has been reached optimum
design criteria, 3 stage HRSG is the most effective system. Additionally, exergy analysis behind energy
analysis has showed that not only amount of heat also quality of heat is calculated and played important role
to choice suitible type of equipment. When using these results in another process, it might not be fix with it
but used method and calculations gives the optimum conditions. In conclusion, energy production in
combined cycle power plant will be more effective and sensitive for environment with specified and evaluated
heat recovery systems.

INTRODUCTION

Today, rising population and developed industry in the world are caused higly needed energy resource.
Therefore, too many kind of technology is improved to supply this energy. Especially, eco-friendly renewable
energy plants(sun,water,wind) and power plants that have minimum effect to global warming are started to
use for energy production. Waste heat and power that produced from power plants is used a certain time to
gain high yield energy.It is clear that heat recovery systems play very important role for decreasing cost and
environmental responsibility.For this reason, different type of systems are designed. And type of using of
these systems are depend on some different parameters. Producing electricity has been much more
effective with combined cycle power plants. The most heat loss in power plants is occured outlet of gas
turbine and it is an unavoided situation.

In this study, performances decreasing heat losses and plant costs are compared in different stage HRSG
with inlet parameters of under constuction power plant. Ege Elektrik power plant is under construction
combined cycle power plant that have 1070 MWe capacity. 1070 MW power is collected from two section
that includes combustion gas turbine with power 378 MW and 296 MW each and two steam turbine with
powers 222 MW and 174 MW each. The total net production is 9,373,200 MW annually(These values are
design criterias and can change during real operation). In this study, 600 MW section is considered. Power
plant basic flow sheet is given as illustrated in Fig. 1 and Fig. 2.

The exergy which is the maximum available energy that a system can deliver as work from its initial state to
the state of environment (dead state) is a tool of the second law analysis which, gives the quality of the
energy differing from the first law analysis techniques, that gives only the quantity of the energy.The exergy
concept was introduced to overcame limitations of the energy analysis. So with increasing of the energy
demand and reduction in the fossil energy reserves, the best (maximum) use of energy became very
important fort he humankind. At this point the exergy analysis gives the wiev for the determining the energy
loss positions and the chance to improve the efficiency (Nag P.K. et al. , 1998), ( Kotas T.J. et al. , 1985),
(Szargut J. Et al. , 1988).

1018
Fig. 1: Combined Cycle Power Plant with 1 stage HRSG

Fig. 2: Combined Cycle Power Plant with 3 stage HRSG

1019
EXERGETIC ANALYSIS

Exergy Analysis

The exergy which is the maximum available energy that a system can deliver as work from its initial state to
the state of environment (dead state) is a tool of the second law analysis which, gives the quality of the
energy differing from the first law analysis techniques, that gives only the quantity of the energy. So with
increasing of the energy demand and reduction in the fossil energy reserves, the best (maximum) use of
energy became very important fort he humankind. At this point the exergy analysis gives the wiev for the
determining the energy loss positions and the chance to improve the efficiency (Kakaras E et al. , 2004),
(Huang Y.-C. et al. , 2000).

In the analysis the kinetic, potantial, magnetic and nuclear exergies are neglected so the exergy is reduced
to the chemical and physical components only ( Kotas T.J. et al. , 1985) , (Szargut J. Et al. , 1988) , (Cengel
Y. A et al. , 2002) , (Bejan A.et al. , 1996).

E=Eph + Ech (1)

A steady state exergy balance includes;

Ei=Eo+W+Ed (2)

Ei denotes inlet exergy to the system, Eo denotes the outlet exergy, W is the work interaction between
system and environment and Ed is the exergy destructed.

With these techniques the exergy analysis is applied to the system. The referance state is choosen as 25 C
and 1 atm with basic atmospheric components.

Ege Elektirik power plant is a typical natural gas combined power plant. 1070 MW power is collected from
two combustion gas turbine with power 378 MW and 296 MW each and two steam turbine with powers 222
MW and 174 MW each. The total net production is 9,373,200 MW annually.

The power plant consists of two combined cycle power blocks which are identical to each other. This power
blocks use the same principle and some units are used in common.We will consider in this paper exergy
analysis of 600 MW part of plant. We will compare and discuss different HRSG configuration results .

Compressor: Air entering the compressor in atmospheric conditions. The following equations are used for
exergy analysis and efficiency:

E1=E3+Wcomp+Ed (3)

ε = (E1-E3)/Wcomp (4)

Combustion Chamber: It is assumed that there is a complete burning in the combustion chamber and
natural gas component is 100% mathane. The following equations are used to calculate the destructed
exergy in combustion chamber and the efficiency.

E2+E3=E4+Ed (5)

ε =E4/(E3+E2) (6)

1020
Turbine: There is expansion to 1,085 bar in turbine and the outlet gas temperature is 1001,5 K.

E5=E4+W+Ed (7)

ε =W/(E5-E4) (8)

Heat recovery steam generator: In this paper we use 2 different HRSG type to compare exergy destruction
and yield. One of these is 1 stage HRSG and the other one is 3 stage HRSG. The results are given in Table
1 and Table 2.

E5+E10=E6+E11+Ed (9)

ε =(E6-E10)/(E5-E11) (10)

Table 1: 1 stage HRSG exergy destructions

Exergy destruction MW %
Compressor 37,75 0,28
CC 10093,33 74,56
GT 2542,78 18,78
HRSG 439,00 3,24
ST 424,46 3,14
Total 13537,32 100,00

Table 2: 3 stage HRSG exergy destructions

Exergy Destruction MW %
Compressor 37,75 0,28
CC 10093,33 73,52
GT 2542,78 18,52
HRSG 548,41 3,99
HPTURBINE 149,84 1,09
IPTURBINE 194,33 1,42
LPTURBINE 161,87 1,18
Total 13728,31 100,00

Steam turbines: The temperature , pressure and flow rate values of stream entering turbines are given in
Table 3 and Table 4.

Table 3: 1 stage HRSG stream entering values

Molar flow
rate
Stream (kmol/s) T(K) P(atm)
Air inlet 44,44 298,15 1,00
Natural gas 4,17 298,15 42,00
GT out 48,61 1001,05 1,09

1021
ST inlet 16,67 850,00 83,89

Table 4 : 3 stage HRSG stream entering values

Molar flow
rate
Stream (kmol/s) T(K) P(atm)
Air inlet 44,44 298,15 1,00
Natural gas 4,17 298,15 42,00
GT out 48,61 1001,05 1,09
HPST inlet 10,03 808,00 119,61
IPST inlet 12,42 748,67 25,36
LPST inlet 17,58 564,00 5,00

Result of exergy analysis

The exergy efficiency values and energy production of combined system components are given in the Table
5 and Table 6. The mole fraction of stack gas is calculated as 1,7 %O2, 70,8 %N2, 8,6 %CO2, 18,8 %H20
with complete burning and the gases are assumed to be ideal during the calculations. It is assumed that
there is no vertical temperature difference in the HRSG.

Table 5: Exergetic efficiency and energy production of 1 stage HRSG

Unit ε (%) Energy production(MW)
Compressor 91,15 -417,35
GT 93,51 791,30
ST 85,90 169,22
HRSG 45,33
Total 543,17

Table 6: Exergetic efficiency and energy production of 3 stage HRSG

Unit ε(%) Energy production(MW)
Compressor 91,15 -417,35
GT 93,51 791,30
HPT 93,68 69,79
IPT 93,29 90,07
LPT 97,76 79,06
HRSG 69,47
Total 612,87

As it seen the largest exergy destruction occurs in the combustion turbine..So by the exergy analysis, it is
seen that the exergy destruction is occurs at the combustion chamber with 74%, the GT with 19% and from
the HRSG with 4%. So the efficiency of the power plant mostly depends on the efficiency of these
equipments so any improvement on these equipments will increase the overall efficiency of the plant
noteworthy. Also chimney temperature plays important role to determine of using energy efficiently. 1 stage
HRSG chimney temperature is about 120 C and 3 stage HRSG chimney temperature is about 100 C.

1022
CONCLUSIONS

As a result, the exergy values of the streams that belong to the equipments of the combined cycle, the
exergy losses, the efficiency values related with each streams and units in the power plant are listed. The
highest exergy losses occur in the combustion chambers. The value is in harmony with the literature value
and continues as a big problem in power plants. The improvements in this part will result an effective
increase in the overall efficiency. On the other hand when exergy destruction and energy production for 2
different configurated HRSG is compared, it can be seen that 3 stage HRSG has much more efficiently
energy production and lower exergy destruction than 1 stage HRSG. As an energy production, it can be
observed that 3 stage HRSG has higher yield than 1 stage HRSG. Also it can be observed that 3 stage
HRSGs chimney temperature is lower than 1 stage HRSGs chimney temperature. According to this results, it
can be said that, if under construction power plant has 3 stage HRSG instead of 1 stage HRSG, exergy loss
will be minimum and energy production will be maximum in given design criteria.

NOMENCLATURE

E exergy ,MW
W work ,MW

Greek Letters

ε efficiency ,%

Subscripts

i inlet
o outlet
d destructed
ph physical
ch chemical
comp compressor

REFERENCES

Nag P.K., De S. , 1998. Study of thermodynamic performance of an integrated gasification combined cycle
power plant. Proceedings of the Institution of Mechanical Engineers. Part A. Journal of power and energy ,
vol. 212 , pp. 89-95

Kakaras E. , Doukelis A. , Leithner R. , Aronis N. , 2004. Combined cycle power plant with integrated low
temperature heat (LOTHECO). Applied thermal engineering, vol. 24, pp. 1677-1686.

Huang Y.-C. ; Hung C.-I. ; Chen C.-K. 2000. Exergy analysis for a combined system of steam-injected gas
turbine cogeneration and multiple-effect evaporation. Proceedings of the Institution of Mechanical Engineers.
Part A. Journal of power and energy, vol. 214, pp. 61-73

Kotas T.J. , 1985. The Exergy Method of Thermal Plant Analysis , Anchor Brendon Ltd. , Tiptree , Essex ,
Great Britain.

Szargut J. , Morris D. R. , Steward F.R. , 1988. Exergy Analysis Of Thermal Chemical And Metallurgical
Processes, Hemisphere Pub. Corp. , New York.

Cengel Y. A. , Boles M. A. , 2002. Thermodynamics: An Engineering Approach, Mc-Graw Hill, 4th Edition.

Bejan A. , Tsatsaronis G. , Moran M. , 1996. Thermal Design And Optimization, Wiley-Interscience

Publication, New York.

1023
 THERMODYNAMIC PERFORMANCE EVALUATION OF A GEOTHERMAL DRYING SYSTEM

H. U. Helvacı, G. G. Akkurt

Mechanical Engineering Department, Izmir Institute of Technology, 35430 GulbahceKoyu, Urla-Izmir, Turkey
huseyinhelvaci@iyte.edu.tr, guldengokcen@iyte.edu.tr

ABSTRACT
Renewable energy sources such as geothermal energy can be used in drying processes as a heat source due to
the high energy costs of fossil fuels.
In this study, geothermal cabinet type dryer was constructed and placed in Balçova-Narlıdere geothermal field
where the clean city water of district heating system is used as an energy source for the dryer. The dryer was
tested on site for drying of olive leaves. The energy and exergy analyses of the drying process conducted under
two cases: 1. Exhaust air is rejected to the environment. 2. A portion of exhaust air was re-circulated. Energy
Utilization Ratio (EUR) was determined as 7.96 for case 1 and 50.36 for case 2.The greatest exergy destruction
occurs in the fan, followed by heat exchanger and the dryer, accounting for 0.2913, 0.05663, 0.0115 kW,
respectively. Exergetic efficiency of the drying chamber was calculated as 89.66%. Re-circulating the exhaust air
decreased the exergy value at the outlet of the dryer from 0.1013kW to 0.08104 kW, means that re-using the air
increases the performance of the dyer.

INTRODUCTION
Herbal, medicinal and aromatic plants have been widely used for preventing and treating specific ailments and
diseases provide the raw material for the food, cosmetic and pharmaceutical industries to produce spice,
essential oils and drugs te ın et al.,1999).Olive leaves have beneficial effects on human metabolism such as
combating fevers and other diseases which are attributed to the phenolic compounds structure (Benavente-
Garcia et al., 2000; Bouaziz et al., 2005). Drying of olive leaves is the most important process prior to
consumption and e traction by reducing moisture content and avoiding the interference of water to the process
oysal and te ın, 2001; Bahloul et al., 2009). Although conventional drying methods such as open air sun
drying is still the most common method in the World, is not suitable for drying of herbs and spices due to
contamination with dust, soil and insects. Furthermore, unmanageable drying parameters such as temperature
and velocity which causes over drying results in loss of quality of dried product. Therefore, the drying process
should be undertaken in closed and controlled environment such as hot air dying in a tunnel or cabinet dryer. An
improvement in the quality of dried product and decrease in time necessitates better technical drying process
which results in a huge amount of thermal and electrical energy use. Therefore, use of renewable energy
sources such as solar and geothermal energy plays an important role in the utilization of controlled conventional
drying methods (Fargali et al., 2008). The daily and seasonal changes of solar intensity, limited time that sun is
available during the day make the drying process intermittent for solar drying. Geothermal resources are good
alternatives to the solar energy. Geothermal fluids can be used indirectly by a heat exchanger or heat pumps in
drying applications. Numerous studies conducted on drying where the geothermal energy was used as a heat
source of the dryer (Andritsos, 2003; Kumoro and Kristano 2003; Hirunlabh 2004; Sumatarto, 2007).
Thermodynamic analysis, mainly exergy analysis plays an important role for system design, analysis and
optimization of thermal systems (Dincer and Sahin, 2004). Since the maximum moisture removal for desired final
conditions of dried product is the main objective with the use of minimum amount of energy in drying industry,
both quantity and quality of energy should be considered throughout the drying process. In other words, energy
and exergy analysis of drying system should be conducted by using the first and second laws of thermodynamics
(Akpinar et al., 2006). Although several studies have been conducted on energy and exergy analysis of solar
drying systems (Midilli and Kucuk, 2003; Akbulut and Durmus, 2010; Chowdhury, 2011) very few papers have
been undertaken on energy and exergy analysis of geothermal drying system which were based on the
geothermal heat pump drying (Kuzgunkaya and Hepbasli, 2007; Colak et al., 2008). Therefore, the primary
objective of this study is to perform energy and exergy analysis of geothermal drying system.

MATERIALS AND METHODS

A cabinet type dryer was constructed and placed Balçova-Narlıdere geothermal field where the geothermal
district heating system exists. Since geothermal fluid temperatures are much higher than necessary for a drier,

1024
circulation water from district heating system was decided to be used. The dryer was placed one of the heat
centres, Yenikaleheat centre. A schematic diagram of the heat centre and the connection of the drier to the
circulation water line were shown in Figure 1.

Fig.1. A schematic diagram of geothermal dryer andYenikale Heat Centre.

Heat centres are the stations where geothermal fluid (1) transfers its heat by primary plate type heat exchangers
(3), to the clean city water (4) which circulates through the city and provides heat to the buildings. In Yenikale
heat centre, the clean city water returns to the heat exchanger (5) at a temperature of 62-65 °C where the dryer
heat input is supplied (6). The geothermal dryer (Figure 2 - 3) is a cabinet dryer, consists of three sections; the
fan unit, heating unit and drying tunnel sections. A schematic diagram of the dryer was given in Figure 4.

Fig. 2.A view of geothermal drier.

1025
Fig. 3.A view of drying trays.

Fig. 4.Schematic view of geothermal dryer.

As seen in Figure 4, drying air is regulated by centrifugal fan (1) which draws the ambient into the dryer. The
ambient air is mixed with re-cycled air before entering the water–to-air heat exchanger (2). The heated air
leaving the water-to-air heat exchanger reaches to the drying chamber 4) which was made of appro imately 1”
plywood. The inner surface of the dryer is insulated by fiber glass and covered with an aluminum sheet in order
to reduce heat losses. The dryer contains 6 trays (3) which are 600 mm in length and 500 mm in width and there
is an air re-circulating unit (5) to recover the exhaust air. The clean water returning to the primary heat exchanger
in the Yenikale heat center is directed to the dryer heat exchanger (Figure 1) as a heat source. During the
experiments, temperature [ambient (12), tray inlet (10), tray exit (11)], relative humidity [ambient (12), tray inlet
(10), tray exit (11)] and velocity [tray inlet (9)] data were measured and recorded by a data logger. Drying
experiments were conducted in April, 2012. Olive leaves (Oleaeuropaea L.) which were dried in the dryer were
collected from the trees in the Izmir Institute of Technology Campus area and were stored for 3 days in the air
proof plastic bags at 4°C. Collected leaves were spread on to the trays as a thin layer. Approximately a mass of
150 g fresh olive leaves was used for each drying experiment. The leaves were dried until the drying rate
reached zero when the moisture content of the leaves were about 4-6% (w-b). To determine changes in weight
of leaves, a digital weighing apparatus ±0.01 g) was used during the drying process.

1026
ANALYSIS
Energy and exergy analysis were investigated using measured temperature, relative humidity and velocity data
at tray inlet, outlet and environment. Energy and exergy analysis were conducted for two cases.
 Case 1 without air re-circulation
 Case 2 with air re-circulation

Energy analysis
Drying process was considered as a steady-state, steady-flow process in the analysis of conservation of mass
(air, water) and energy.

Fig. 5. Schematic illustration of drying system (no re-circulation)

da,env = air,(1) (1)

da,env × hda,env = air,(1) × hair,(1) (2)

1027
Fig. 6. Schematic illustration of drying system (re-circulation)

da = da,env + re,air (3)

da,env × (1-r) + rec,air × (r) = air,(1) (4)

da,env × hda,env× (1 - r) + re,air × hre,air × (r) = air,(1) × hair,(1) (5)

Following equations are generally applied to compute the mass and energy conservation of energy (Midilli and
Kucuk, 2003; Cengel and Boles, 2011)

Fan

air,(1) = air,(2) = da (6)

The general equation of energy conservation can be expressed as

– = ∑( da)in (hout + )-∑( da)out (hout + ) (7)

) (8)

Heat exchanger

a) Water

water,in = water,out (9)

water (hwater,in - hwater,out) (10)

1028
b) Air

air,(2) = air,(3) = da (11)

da (h3 - h2) (12)

Drying chamber
General equation of mass conservation of drying air for drying chamber

∑ da,(3) =∑ da,(4) = da (13)

General equation of mass conservation of moisture

∑( water)in + ( water)product =∑( water)out (14)

Or

∑(ωda)in da,(3) + ( water)product =∑(ωda)out da,(4) (15)

Equation 16 is generally used to transform the relative humidity to humidity ratio of the air at inlet and outlet of
the drying chamber.

 =0.622× (16)

Total mass flow of air needed for drying process was calculated using Equation 17.

da = ρda×Vda× A (17)

The enthalpy of drying air

h = (Cp)daTda +w.hsat@T (18)

The energy utilization (EU) which can be expressed as the amount of energy employed during moisture removal
of the product was calculated using Equation 19.

EU = da . [hda,(3) – ( hda,(4)] (19)

The energy utilization ratio (EUR) was defined as the ratio of the energy utilization to the energy given from the
heat exchangers of dryer.

EUR = × 100 (20)

Exergy analysis
Total exergy of inlet, outlet and losses through the system components were determined in the scope of the
second law analysis of thermodynamics (Akpinar et al., 2006; Cengel and Boles, 2011).The exergy values are
calculated by using the characteristics of the working medium from a first-law energy balance (Szagut et al.,
1988).

The specific exergy is determined as

 = (h-h0) – T0 × (s - s0) (21)

1029
Exergy rate is defined as
(22)

Or

= daCpda [(T- T0) – T0ln ] (23)

The dead state temperature (T 0) was taken as the measured environment temperature (T env). Exergy
destructions as well as exergetic efficiencies were obtained by exergy balances for each component.
Fan

  (24)

 
(25)

Heat exchanger

a) Water

  (26)

b) Air

  (27)

 
he = (28)
 

Drying chamber

  (29)

dc = (30)

Or

 
dc =1- (31)


Assumptions
The following assumptions were made;
 All the process was taken as a steady state and steady flow.
 Heat transfer to the system and work transfer from the system are positive.
 Potential and kinetic energy effects were neglected.
 Air is an ideal gas with constant specific heat.

RESULTS AND DISCUSSION

Energy and exergy analysis of the drier was conducted using measured temperature, velocity and relative
humidity data. Mass flow rates, enthalpy values and temperature of the air at inlet of the heat exchanger were
calculated. Temperature, relative humidity, mass flow rate data, for the air and water are given in Table 1.
Energy and exergy rates are also calculated for each state and represented in Table 1. In this study, the

1030
reference-dead state conditions were determined as T 0= Tenv, RH0= RHenv, P0= 101.325 kPa. The
thermodynamic properties of air and water were obtained by using Engineering Equation Solver (EES, Inc. 2012)
software package.

Table 1. Energy and exergy analysis data

Specific Exergy Energy
State T RH h s
Fluid Exergy Rate rate
No °C) (%) (kJ/kg) (kJ/kgK) kg/s
(kJ/kg) (kW) (kW)
0 Air/Dead
22 42 39.67 5.75 _ _ _ _
State
1 Air/Fan
37 19.4 56.79 5.8 0.20 0.3175 0.066 11.8
inlet
2 Air/Fan
outlet- 37.9 18.5 57.67 5.81 0.20 0.3615 0.076 11.97
HE inlet
3 Air/HE
outlet-
40.9 15.7 60.81 5.82 0.21 0.538 0.1128 12.78
Dryer
inlet
4 Air/Dryer
39.8 16.4 59.23 5.815 0.21 0.4823 0.1013 12.43
outlet
W in Water/
62 _ 259.6 0.8562 0.03 63.11 2.272 9.34
HE inlet
W out Water/
56.5 _ 236.5 0.7866 0.03 60.51 2.178 8.52
HE outlet

Table 2 presents exergy destruction and exergetic efficiencies for each component of drying system. Energy
Utilization (EU) and Energy Utilization Ratio (EUR) values are also represented in Table 2.

Table 2.Exergy destruction, exergetic efficiency, EU and EUR data of the dryer.
Component Rate of Exergy Destruction Exergetic efficiency
Exdestruction (kW) (%)
Fan 0.2913 3.08
Heat exchanger 0.05663 39.52
Dryer 0.0115 89.66
EU = 0.3316 kW
EUR (Case 1) = 7.96 (%)
EUR (Case 2) = 50.36 (%)

The EU, which indicates the amount of energy employed for the reduction of moisture content of the product,
was calculated as 0.3316 kW. Energy Utilization Ratios were obtained as 7.96 % and 50.36 % for both Case 1
and Case 2, respectively. The EUR indicates the ratio of the energy utilization to the amount of energy given
from the heat exchangers of the dryer. The lower EUR values obtained from the experiments that air was not
circulated compared to the experiments with air recirculation (Figure 7). This indicates that recycling the exhaust

1031
air allowed us to reuse the energy leaving the dryer which decreases the energy input from the heat exchanger
and increases the energy efficiency (Corzo et al., 2008).

EUR EUR
7.96 %
50.36 %

EUR EUR

Fig. 7. a) Case 1 b) Case 2

Exergy Destruction
(kW)
0,0115 (kW)
0,05663 (kW)

Fan
0,2913 (kW) HE
Dryer

Fig. 8.Exergy destruction values for each component of the drying system.

It is clear from the Figure 8 that the greatest exergy destruction occurs in the fan, followed by heat exchanger
and the dryer, accounting for 0.2913, 0.05663, 0.0115 kW. Exergetic efficiency can be defined as the ratio of
exergy loss (exergy inflow – exergy outflow) to exergy inflow to the drying chamber. In this study exergetic
efficiency of the drying chamber was found 89.66 %. This result was similar with the results in the literature
(Corzo et al., 2008; Erbay and Icier, 2011). According to the results by recirculating the exhaust air exergy value
at the outlet of the dryer decreased from 0.1013 kW to 0.08104 kW. That means re-circulating some portions of
the exhaust air at the outlet of the dryer which still has availability.

CONCLUSIONS
Thermodynamic analyses of geothermal drying process of olive leaves were performed in the aim of this study.
The results of the energy and exergy analysis indicated that the greatest exergy destruction occurred in the fan,
followed by heat exchanger and dryer. The exergetic efficiencies for the whole geothermal drying unit varied
between 3.08 and 89.66. Re-circulating exhaust air increased EUR values from 7.96% to 50.36%, and
decreased the exergy by 0.02026 kW at the outlet of the dryer, means that re-using the air enables to use
energy and exergy more effectively and this increases the performance of the dryer. An economic analysis
should be conducted to show the viability of the recirculation for commercial drying facilities. In geothermal

1032
dryers, energy source (geothermal fluid or waste water of a plant) is almost free but if the energy source is a
fossil fuel, drying process is more profitable.

Acknowledgement
The authors thankIzmir Jeotermal for technical supports.

NOMENCLATURE
2
A area, m
Cp specific heat, kJ/kg K
EU energy utilization, kW
EUR energy utilization ratio, %
exergy rate, kW
h specific enthalpy, kJ/kg
mass flow rate, kg/s
P pressure, kPa
heat rate, kW
RH relative humidity, %
r recirculation ratio, %
s specific entropy, kJ/kg K
T temperature, °C
V velocity, m/s
work rate, kW
3
ρ density, kg/m

Greek Letters
ψ specific exergy, kJ/kg
ε exergetic efficiency, %
 specific humidity ratio, kg water/kg air

Subscripts
da drying air
dc drying chamber
des destroyed
env environment
He heat exchanger
in inlet, inflow
out outlet, outflow
re recirculated
sat saturated

Superscripts
0 dead (reference) state

REFERENCES
Akbulut, A. and A. Durmus. 2010. Energy and Exergy Analyses of Thin Layer Drying of Mulberry in a Forced
Solar Dryer. Energy 35:1754-1763.

Akpinar, E. K., A.Midilli and Y. Bicer. 2006. The first and second law analyses of thermodynamic of pumpkin drying
process. Journal of Food Engineering 72:320-331.

Andritsos, N., P. Dalampakis, and N. Kolios. 2003. Use Of Geothermal Energy For Tomato Drying. GeoHeat
Center QuarterlyBul 24:9-13.

1033
Bahloul, N., N. Boudhrioua, M. Kouhila, and N. Kechaou. 2009. Convective Solar Drying Of Olive Leaves.
Journal of Food Process Engineering 34:1338-1362.

Benavente- arcıa, O., J. Castillo, J. Lorente, A. Ortuño, and J. A. Del Rio. 2000. Antio idant Activity of Phenolics
Extracted from OleaEuropaea L. Leaves. Food Chemistry 68:457-462.

Bouaziz, M. and S. Sayadi. 2005. Isolation and Evaluation of Antioxidants from Leaves of a Tunisian Cultivar
Olive Tree. European Journal of Lipid Science and Technology 107:497-504.

Cengel, Y. and M. A. Boles. 2011. Thermodynamics: an engineering approach. New York: McGraw- Hill.

Chowdhury, M.M.I., B.K. Bala, and M.A. Haque. 2011. Energy and Exergy Analysis of the Solar Drying of
Jackfruit Leather. Biosystems Engineering 110:222-229.

Colak, N., E. Kuzgunkaya, and A. Hepbaşlı. 2008. E ergeticAssesment Of Drying Of Mint Leaves In a Heat
Pump Dryer. Journal of Food Process Engineering 31:281-298.

Corzo, O., N. Bracho, A. Vasquez, and A.Pereira. 2008. Energy and exergy analysis of thin layer drying of coroba
slices. Journal of Food Engineering. 86:151-161.

Dincer, I. and A. Z. Sahin. 2004. A new model for thermodynamic analysis of a drying process. International Journal
of Heat and Mass Transfer 47:645-652.

Erbay, Z., F. İçier. 2010. Energy and e ergy analysis on drying of olive leaves (Oleaeuropaea L.) in tray drier.
Journal of Food Process Engineering. 34:2105-2123.

Fargali, H. M., A. A. Nafeh, F. H. Fahmy, and M. A. Hassan. 2008. Medicinal Herb Drying Using a Photovoltaic
Array and a Solar Thermal System. Solar Energy 82:1154-1160.

Hirunlabh, J., S. Thiebrat and J. Khedari. 2004. Chili And Garlic Drying by Using Waste Heat Recovery From a
Geothermal Power Plant. GeoHeat Center QuarterlyBul25:25-27.

Kumoro, A.C. and D. Kristanto. 2003. Preliminary study on the utilization of geothermal energy for drying of
agricultural product. International Geothermal Conference 14:1-7.

Kuzgunkaya, E.H., and A. Hepbasli. 2007. Exergetic Performance Assessment of a Ground source Heat Pump
Drying System. International Journal of Energy 31:760-777.

Midilli, A. and H. Kucuk. 2003. Energy and Exergy Analyses of Solar Drying Process of Pistachio. Energy 28:
539-556.

O te in, ., A. Başçetinçeli , and Y. oysal. 1999. Crop Drying Programme in Turkey. Renewable Energy 1:
789-794.

oysal, Y. and . te in. 2001. PH—Postharvest Technology: Technical and Economic Performance of a Tray
Dryer for Medicinal and Aromatic Plants. Journal of Agricultural Engineering Research 79:73-79.

Sumotarto, U. 2007. Design of a geothermal energy dryer for beans and grains drying in Kamojang geothermal
field, Indonesia. GeoHeat Center Quarterly Bul28:13-18.

Szargut, J., D. R. Morris and F. R. Steward. 1988. Exergy analysis of thermal, chemical and metallurgical
processes. New York: Hemisphere Publishing Corp.

1034
 CHEMICAL THERMODYNAMICS OF HYDROCARBON COMPOSITIONS
IN NATURAL GAS FIELD, NORTHEAST OF THAILAND
a b
S. Promkotra , T. Kangsadan
a
Department of Geotechnology, Faculty of Technology, Khon Kaen University, Khon Kaen 40002, Thailand
b
Chemical and Process Engineering Program, Department of Mechanical and Process Engineering,
The Sirindhorn International Thai-German Graduate School of Engineering (TGGS),
King Mongkut’s University of Technology North Bangkok (KMUTNB), Bangsue, Bangkok 10800, Thailand
1 2
e-mail: sarunya@kku.ac.th, tawiwan.k.cpe@tggs-bangkok.org

ABSTRACT
The relationship of hydrocarbon species in natural gas is predicted the phase envelope by thermodynamic simulation.
Conceptual thermodynamics depends on physical and chemical characteristics in equilibrium condition. The API
gravity of condensate is expressed in a range of 55-63o. Its average density is similar to the specific gravity of 0.74.
Pressures in each gas processing process are decreased from the gas well (74.754 bar.g) to the condensate tank
(0.02 bar.g). Mostly methane is discovered approximately 96%mole. Due to physical and chemical properties, this
petroleum reservoir indicates the dry gas reservoir. Chemical reactions depending on hydrocarbon compositions can
be valuable for evaluating phase equilibrium. They derived from different hydrocarbon types are fabricated by
chemical species as reactants and products. For creating them, all compositions related to the genesis of petroleum
in the same environment are combined as a reactant. They are possibly defined phase related to Gibbs free energy.
Carbon dioxide, hydrogen and nitrogen gas in petroleum reservoir are 10, 8, and 2 mole percent, respectively. The
chemical thermodynamics is simulated by using Prode Properties, a computer software program, in which Peng-
Robinson equation of state is used as the chemical thermodynamic model. Characteristics of phase envelope consist
of the critical temperature and pressure at phase transformation between liquid and gas. Two types of conceptual
schemes are varying the invariable and variable mole fraction to evaluate the phase envelope. For their invariable
mole fraction, the carbon ratio is defined at 60:20 percentages.Temperatures of natural gas from the phase envelope
are inversely proportional to equilibrium constant. High temperature at low equilibrium constant specifies the forward
chemical reaction. For their variable mole fraction, the increased carbon atoms affect to gradually intensify the
temperature and pressure. Hydrocarbon types and mole fraction of these components concern to the possibility of
phase envelope at suitable phase.

INTRODUCTION
Natural gas systems are generally analyzed by using hydrocarbon thermodynamics to understanding the interaction
between natural gas fluids and assorted environments. The outcomes lead to conception in engineering practice
mainly in the reservoir, pipeline and production process. Fluid thermodynamic behavior can be assessed behaviors of
any feasible fluid phases as a function of temperature, pressure and hydrocarbon components. To optimal propose
and proficient process of the gas processing plant, the phase behavior of hydrocarbon reservoir fluids will be
fabricated under a series of phase in equilibrium condition over the two-phase pressure-temperature range. Thus, the
relationship of hydrocarbon species in the natural gas field is predicted the phase envelope by chemical
thermodynamic concept. This inquiry is worthy research to comprehend the phase equilibria of petroleum fluid.
Reservoir fluids include a wide variety of matters of chemical nature that comprise of hydrocarbons a n d non-
hydrocarbons. H ydrocarbon fluids in petroleum reservoir are very complex substances. Although, the complexity of
hydrocarbon fluids generates in underground reservoirs, equations of state (EOS) can express the phase-behavior
computations of these complex fluids. An equation of state is a methodical representation concerning pressure to
temperature and volume. The appearance is used to reveal the vapor-liquid equilibrium (VLE), and the thermal
properties of pure materials and mixtures. Chemical thermodynamic model depends on the category of the phases
in equilibrium using the Peng-Robinson cubic equation of state for the vapor-liquid phase (Freze et al., 1983).
Chemical thermodynamics of natural gas and condensate is the useful research to understand phase behaviors
of different hydrocarbon criteria (Ayala, 2006). Normally, heavy hydrocarbon compositions affect the thermo-
physical properties of the gas, and the range is characteristically undefined. Most equations of state is
importance to describe the phase equilibria of natural gas, however cannot correspond the phase loops for
pressures in VLE, especially at high pressure density. The developed prediction will be utilized by the specific
EOS for the performance of gas processing facilities (Atilhana et al., 2010).
For the construction of complete vapor—liquid phase envelopes, there was qualified of precisely defining the
critical points and maximum temperature and pressure by an algorithm (Michelsen, 1980). Currently improved
methods for absolute computation of critical points are essentially considered, and a thermodynamic simulation is
suggesting more convenient creation. Cubic equations of state, for example the Peng-Robinson EOS, are
regularly functioned by the petroleum industry for the intention of recovery and processing operations of gases at
high pressures. Alfradique and Castier (2007) selected the Peng-Robinson EOS in the calculation of dew points,
bubble points, and critical points of natural gases. For this reason, the Peng-Robinson EOS is possibly those to

1035
complete this aim with widest acceptance. Thus, the goals of this research are to identify chemical compositions
of natural gas in equilibrium condition and to evaluate phase envelope of natural gas and condensate by
thermodynamic simulation.

STUDY AREA
Study area is located in the natural gas field in Udon Thani province, Thailand where is a situated production well of
the gas field. The distance from production wells to the gas processing plant (GPP) in Khon Kaen province is
approximately 62 km. Natural gas and condensate are collected from GPP in Num Phong District, Khon Kaen
province where is located in the northeastern part of Thailand. Generally, petroleum regions in Thailand are
geographically divided into Northern and Central Plains, Northeastern, Southern, Andaman Sea and Gulf of Thailand
(Ministry of Energy, 2007). Major produced reservoirs are supplied in different parts of the country. Commercial
hydrocarbon potentials were found in the Northeast and one of them can produce petroleum gas from Permian
carbonates. Exploration data reveals that Permian carbonates are considerable primary reservoirs. High-potential
expectations are situated in surroundings of Nam Phong and Sin PhuHorm Gas Field, including Phu Khieo Prospect.
Dry gas is expected to be the petroleum resource in a profitable scale. Exploitation of the natural gas can be
supported for the power plant (Sattayarak, 2005).

METHODOLOGY

Gas Processing Plant

Petroleum fluid is transported from a petroleum reservoir to the ground surface (160 bar.g pressure, 60oC
temperature) and flow to the gas processing plant (GPP) in Khon Kaen province. In the first part of GPP, petroleum
fluid flows to the slug catcher (74.54 bar.g pressure, 31.67oC temperature). Fluid in this locationis separated from
natural gas. For the natural gas line, gas flows to the inlet coalescing filter and mercury absorber in order to eliminate
dust and mercury whereas water in natural gas is separated by tri-ethylene glycol at the glycol contactor with the
pressure of 45.01 bar.g and temperature of 38.99oC. Then, gas flows to the gas/gas exchanger and pass to the low
temperature separator to decrease temperature and send to the sale gas system (27.83 bar.g pressure, 41.20oC
temperature).
For the liquid line, liquid petroleum consists of condensate and water. Petroleum liquid flows to the high pressure
flash vessel (6.49 bar of pressure, 24.28oC of temperature) which condensate is separated from water. Water flows to
the produced water separator and boiled at a boil-off tank, and subsequently this water is destroyed in the thermal
oxidizer (700oC). Moreover, condensate flows to the low pressure flash vessel and evaporates. Finally, condensate
runs to a condensate tank to storage.Temperature and pressure in each location in GPP are useful for correlating to
the conceptual model for thermodynamic simulation.

Sample Collection
Condensate samples are collected two times in August and October in 2012. These samples are sampling in the
inlet slug catcher, where the natural gas and other fluid mixtures (condensate and water) are separated out. Thus,
petroleum-field water must be removed first from condensate. Besides, natural gas, entirely methane, is also
withdrawn from this process. The natural gas is collected at this point.

Natural Gas and Condensate Analysis

Natural gas and condensate compositions are obtained by the gas chromatography (GC), GC-2010 Plus
(Shimadzu), based on ASTM D1945. Two varieties of solvents for carrier phase are n-heptane and n-hexane and
methyl heptadecanoate (C18H36O2) is a standard solution to define the hydrocarbon series on chromatograms.
Physical properties of condensate are examined in the chemical laboratory, such as density, specific gravity, API
gravity, Reid vapor pressure (RVP) and true vapor pressure (TVP), according to ASTM D5191. In addition, these
parameters are daily measured to maintain the quality of condensate.

Chemical Thermodynamic Simulation

Hydrocarbon reservoir is usual to characterize the phase behavior on the pressure-temperature range indicating
the limits over which the fluid subsists as a single phase and the proportions of petroleum oil and gas in
equilibrium over the two phases. Possibility of chemical reaction in petroleum reservoir can create phase
depending on hydrocarbon composition. This reservoir of the study area is a dry gas reservoir, abundant with
methane gas (Campbell, 1994). The advent of highly methane found by analytical results of the natural gas can
prove this assumption. Chemical reactions based on hydrocarbon types are fabricated by chemical species as
reactants and products. The Peng-Robinson equation of state is suitable for simulating the chemical
thermodynamic model (USGS, 2006). Thus, phase equilibrium obtains the occurrence of condensate and natural
gas under chemical equilibrium at any temperature and pressure change. The Prode Properties, computer
software program, provide the thermodynamic framework designed to afford phase envelopein process simulation.
This program is used in direct integration with Excel software program. The calculated equilibrium points including

1036
critical points, cricondentherm and cricondenbar are displayed in Excel page. Thus, phase diagram and phase
envelope of natural gas mixtures are obtained by the Prode Properties (Prode Properties, 2010).
Prediction of hydrocarbon parameters is mainly temperature and pressure. Hydrocarbon component such as
methane is dominantly a precursor in every chemical reaction. Moreover, other gases related to the genesis within
reservoir and exist in the same environment are combined as a reactant. Analytical results from chemical
laboratory indicate carbon dioxide, hydrogen gas and nitrogen in which reservoir are 10, 8, and 2 mole percent,
respectively. Chemical reactions are conducted by using petroleum components from condensate which is
collected in GPP. Two important conditions to control phase behavior consists of pressure and temperature.
General scheme for these two variables, pressure and temperature, are defined as 200 bar.g and 200oC,
respectively. Not only these conditions are concerned, but geological specifications are considerable as well.
Relationship between chemical equilibrium involved in thermodynamics indicates that they lead to the phase
change. All chemical species are representative for evaluation the possibility of phase envelope. Thus, five
chemical reactions are createdtotally, based on these criteria, as follows:

Condition: Pressure = 200 bar.g, Temperature = 200oC.

Variable chemical species depending on reactant, as indicated in each chemical reaction:

1. Chemical Reaction-1

4CH4 (g) + 2C2H6 (g) + 2CO2 (g) + H2 (g) + N2 (g) = (2CH4 (g) + C7H16 (l) + CO2 (aq) + N2 (aq) + 2H2O (aq))* (1)

2. Chemical Reaction-2

3CH4 (g) + C3H8 (g) + 3CO2 (g) + 4H2 (g) + N2 (g) = (CH4 (g) + C7H16 (l) + CO2 (aq) + N2 (aq) + 4H2O (aq))* (2)

3. Chemical Reaction-3

2CH4 (g) + C4H10 (g) + 3CO2 (g) + 4H2 (g) + N2 (g) = (CH4 (g) + C7H16 (l) + CO2 (aq) + N2 (aq) + 4H2O (aq))* (3)

4. Chemical Reaction-4

CH4 (g) + C5H12 (g) + 3CO2 (g) + 4H2 (g) + N2 (g) = (CH4 (g) + C7H16 (l) + CO2 (aq) + N2 (aq) + 4H2O (aq))* (4)

5. Chemical Reaction-5

CH4 (g) + C6H12 (g) + 2CO2 (g) + 3H2 (g) + N2 (g) = (CH4 (g) + C7H16 (l) + CO2 (aq) + N2 (aq) + 2H2O (aq)).* (5)

Remark: “*” = Optional

RESULTS AND DISCUSSION

Natural Gas and Condensate

Natural gas (and other liquid) flows from the gas well to the gas processing plant (GPP) and meets the first
location at the inlet slug catcher to separate fluid mixture. Liquid (condensate and water) are isolated from natural
gas before flow to the high pressure flash vessel. Gas chromatography is a technique to detect hydrocarbon
composition of the natural gas in the gas line. Its chemical composition in mole fraction is indicated in Table 1.
Some parameters of condensate, for example density, specific gravity, API gravity, Reid vapor pressure (RVP)
and true vapor pressure (TVP) are determined in the chemical laboratory and indicated in Table 2. According to the
results in these two tables, the number 1 and 2 are the sampling date on August 18th, 2012 and October 2nd, 2012,
respectively. Each location, inlet slug catcher, high pressure flash vessel, low pressure flash vessel, produced
water separator, condensate tank, shows the comparison of their properties among them. Condensate samples
are also examined by the gas chromatography. Hydrocarbon compositions in mole percent are indicated in Table
1. Chromatogram of condensate at the inlet slug catcher is depicted in Fig. 1. Information from GPP consists of
temperature and pressure, combination with physical properties and chemical properties measuring in chemical
laboratory. All results are appraised to determine chemical thermodynamic concept.

Invariable Mole Fraction

The results of the thermodynamic simulations display the relationship of pressure as a function of temperature.
These correlations depend on chemical reaction from Eq. (1), as indicated in Fig. 2. Mole fraction of hydrocarbon
is defined only 80 mole% due to the result of GC indicating sum of other gases at 20 mole%. Methane is mainly

1037
fraction of hydrocarbon (60 mole%). Other hydrocarbon gas, such as ethane, propane, butane, pentane and
hexane are defined for 20 mole%, and varies these compositions which obtain in Table 3. Every hydrocarbon
equals to 60 mole% of methane and 20 mole% of other gases.

Table 1.Chemical compositions of natural gas and condensate.

Condensate (1) Condensate (2)
Components (mole%) Natural Gas
August 18th, 2012 October 2nd, 2012
Methane 96.6428 - -
Ethane 0.8981 - -
Propane 0.1095 - -
i-Butane 0.0178 - -
n-Butane 0.0356 - -
i-Pentane 0.0151 - -
n-Pentane 0.0232 - -
Hexane 0.1018 - -
Heptane - 20.168 21.424
Octane - 24.259 23.745
Nonane - 18.236 19.681
Decane - 14.418 13.779
Undecane - 10.573 10.155
Dodecane - 6.147 5.772
Tridecane - 3.218 2.899
Tetradecane - 1.736 1.495
Pentadecane - 0.859 0.716
Hexadecane - 0.386 0.333
CO2 0.5021 - -
N2 1.6537 - -
Water - 0.012 0.009

Intensity

C7

C8

C9

C10

C11

C12
C13
C14 C15
C16

(min)

Fig. 1.Chromatogram of condensate (inlet slug catcher).

The relationship of pressure as a function of temperature displays the critical point (red triangle, as shown in
Fig. 2) which occurs under specific conditions, such as temperature, pressure or composition of hydrocarbon at

1038
which no phase boundaries exist. Dew line (dark blue line) is saturation line appearing in gas phase, but liquid
fraction line (bright blue line) is saturation line taking place in liquid phase. Moreover, cricondenbar (CriP) point is
the maximum pressure, and cricondentherm (CriT) is the maximum temperature at which liquid and vapor may
coexist in equilibrium. Unit of pressure is bar.g which is abbreviated from bar gauge which actually measures the
difference between the fluid pressure and the atmospheric pressure.

Table 2.Physical properties of condensate.

Sampling Location Sampling Time Density (g/cm3) API gravity Specific gravity RVP (psi) TVP (psi)
1 0.76 55.77 0.7556 0.91 1.51
Inlet Slug Catcher
2 0.75 55.52 0.7566 0.75 1.35
1 0.72 57.50 0.7487 0.99 1.60
High Pressure Flash Vessel
2 0.76 58.20 0.7459 0.97 1.58
1 0.75 59.83 0.7396 1.42 1.52
Low Pressure Flash Vessel
2 0.74 60.03 0.7362 1.68 1.95
1 0.73 62.84 0.7281 3.03 3.70
Produced Water Separator
2 0.73 61.28 0.7340 1.99 2.51
1 0.75 57.50 0.7487 1.33 1.96
Condensate Tank
2 0.73 61.59 0.7328 1.64 2.16

Table 3.Carbon ratio and mole fraction of each chemical reaction.

Chemical Reaction Carbon Ratio Mole Fraction (%)
Eq. 1 C1:C2 60:20
Eq. 2 C1:C3 60:20
Eq. 3 C1:C4 60:20
Eq. 4 C1:C5 60:20
Eq. 5 C1:C6 60:20

Temperature (oC)
-160 -140 -120 -100 -80 -60 -40 -20 0
100

Dew line CP 1 90
Liquid fraction
80
CP 1
70
Pressure (bar.g)

60

50

40

30

20

10

0
Parameters at Critical Point
Pressure 90.62 bar.g
Temperature -35.22oC
Cricodentherm -27.17oC
Cricodenbar 92.72 bar.g

Fig.2. Pressure as a function of temperature for methane and ethane, and parameters at critical point.

1039
Variable Mole Fraction
According to chemical reaction for thermodynamic assumption, hydrocarbon types and their concentration (mole
fraction) are varied, as shown in Table 4. The outcomes from thermodynamic simulation show phase envelopes
(Fig. 3-5) where a dew line (solid line) is saturation line appearing in gas phase, but bubble line (dash line) is
saturation line taking place in liquid phase. Each phase diagram reveals the critical point of temperature (Tc) and
pressure (Pc). This point replicates that the liquid and gaseous phases display indistinguishable. Moreover,
cricodentherm (CriT) and cricodenbar (CriP) come from their phases. Relationship of these variables can conclude
in the plot of temperature as a function of pressure.

Table 4.Pressure and temperature of each chemical reaction.

Reaction Carbon Ratio Concentration (mole%) Pc (bar.g) Tc (oC) CriP (bar.g) CriT (oC)
Eq. 1 C1:C2 60:20 90.62 -35.23 92.72 -27.18
50:30 90.89 -20.92 95.39 -14.71
40:40 88.11 -8.460 -1.103 -4.150
30:50 83.61 2.360 -1.103 5.080
20:60 78.25 11.76 -1.103 13.27
Eq. 2 C1:C3 60:20 121.56 -5.340 124.18 16.91
50:30 116.98 20.15 126.20 35.21
40:40 106.48 40.17 121.95 49.57
30:50 94.49 55.60 1.013 61.41
20:60 82.76 68.70 -1.013 71.49
Eq. 3 C1:C4 60:20 142.01 18.20 146.20 50.61
50:30 129.44 50.70 143.93 70.92
40:40 111.91 56.00 134.24 61.41
30:50 94.99 74.40 -1.013 86.32
20:60 79.92 106.67 -1.013 109.00
Eq. 4 C1:C5 60:20 181.44 51.83 190.55 98.63
50:30 154.49 94.27 182.24 120.97
40:40 126.34 122.55 165.23 137.44
30:50 102.50 142.43 145.32 150.38
20:60 83.12 157.03 -1.013 160.96
Eq. 5 C1:C6 60:20 215.87 92.19 238.10 147.22
50:30 185.36 120.45 201.30 167.21
40:40 131.75 171.32 193.01 186.69
30:50 102.98 191.37 164.89 199.34
20:60 80.50 205.91 -1.013 209.48

Equilibrium Constant of Chemical Reaction

The van't Hoff equation in chemical thermodynamics relates to the change of temperature and equilibrium
constant, given the standard enthalpy change (ΔHo) for calculation. The energy associated with a chemical
reaction can be evaluated its possibility happening under this circumstance. Gibbs free energy (referred to ∆Go) is
also the chemical potential that is minimized when a system reaches to equilibrium at constant pressure and
temperature. Its derivative with respect to the reaction complement of the system disappears at the equilibrium
state. Thus, the free energy is a suitable criterion of spontaneity for the genesis of the chemical reaction with
constant pressure and temperature. The free energy of a chemical reaction is calculated by the sum of its enthalpy
(H) along with the temperature (T) and entropy (S). Gibbs free energy, enthalpy and equilibrium constant (K eq) of
each chemical reaction are calculated at temperature of 0, 15, 25 and 200 oC, as shown in Table 5 and
corresponding to Fig. 6. According to Gibbs free energy, the chemical reaction Eq. (1) is non-spontaneous process
because of the positive energy, which implies the reverse reaction. This chemical reaction is not capable of
proceeding in a given direction, as written. It needs to be driven by an outside source of energy. However, the
others are the spontaneous reactions and relevant to the assumptions based on conceptual chemical
thermodynamics. Moreover, enthalpy of reaction indicates all negative numbers as an exothermic reaction.
Therefore, these chemical reactions do not need to add more heat to the system.

1040
 Temperature (oC)
-200 -150 -100 -50 0 50 100
200

180

160

140

Pressure (bar.g)
120

100

80

60

40

20

0
Fig. 3.Effect of composition on phase envelope, methane-ethane binary according to the chemical reaction Eq. (1).
Temperature (oC) Temperature (oC)
-150 -100 -50 0 50 100 -200 -150 -100 -50 0 50 100 150
250 250

200 200

Pressure (bar.g)
Pressure (bar.g)

150 150

100 100

50 50

0 0
(a) (b)
Fig. 4.Effect of composition on phase envelope of binary compositions of (a) methane-propane for the chemical reaction Eq. (2), and (b) methane-
butane for the chemical reaction Eq. (3).

For invariable mole fraction of hydrocarbon, methane and another gas are controlled in the ratio of 60:20.
Temperature as a function of pressure presents the linear correlation, as shown in Fig. 7. Increasing the number of
carbon atoms will gradually intensify the temperature and pressure. Linear relationship of the critical point and
CriT-CriP line is similar to each other (Fig. 8). However, at high temperature and pressure, these two lines are
obviously separated. Equilibrium at critical point expresses the different phases between liquid and gas. On the
other hand, for variable mole fraction of hydrocarbon, the difference of methane affect to the relationship between
temperature and pressure. No linear correlation is found all over the plot. The interval between temperature and
pressure obtain from reducing methane concentration. Moreover, hydrocarbon gas change from ethane to hexane
can cause the different consequences. The lowest value of pressure is -1.013 bar.g which is hardly realistic. The
results of these chemical thermodynamic simulations can predict the possibility of phase equilibria. Hydrocarbon
types and mole fraction of these components influence to the occurrence of phase envelope.

1041
Table 5.Gibbs free energy, enthalpy and equilibrium constant of each chemical reaction.
Reaction Gibbs Free Energy (kJ/mol) Enthalpy (kJ/mol) Equilibrium Constant (Keq)
0oC 15oC 25oC 200oC
Eq. 1 32.555 -217.900 0.93073 0.92838 0.92695 0.91184
Eq. 2 -17.94 -310.228 1.01537 1.01033 1.00726 0.97514
Eq. 3 -75.99 -363.084 1.03752 1.03355 1.03113 1.00571
Eq. 4 -134.565 -417.015 1.0615 1.05795 1.05578 1.03297
Eq. 5 -193.837 -471.633 1.08577 1.08301 1.08133 1.06359
Temperature (oC) Temperature (oC)
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200 250
300 300

250 250

200 200

Pressure (bar.g)

Pressure (bar.g)
30% CH4 150 150

100 100

50 50

0 0
(a) (b)
Fig. 5.Effect of composition on phase envelope of (a) methane-pentane binary for the chemical reaction Eq. (4) and (b) methane-hexane binary for
the chemical reaction Eq. (5).
Gibb free energy
(kJ/mol)
Gibbs free energy (kJ/mol) Enthalpy (kJ/mol)

-250 -200 -150 -100 -50 0 50 -500 -450 -400 -350 -300 -250 -200
1.10
1.10
(a) (b) 1.08
1.08
1.06 1.06
Equilibrium constant

Equilibrium constant
1.04 1.04

1.02 1.02

1.00 1.00

0.98 0.98
0C
0.96 0C 0.96
15C 15C
25C 0.94 25C 0.94
200C 0.92 200C 0.92
0.90 0.90

Fig.6. Equilibrium constant as a function of temperature at 0, 15, 25 and 200oC of (a) Gibbs free energy, and (b) enthalpy.

For equilibrium condition of thermodynamic simulation, thermodynamic calculations based on the chemical
reaction are pursued under exothermic reaction which reveal by enthalpy computation. These enthalpies relate to
the equilibrium constant at different temperatures: 0, 15, 25 and 200oC. The higher temperature can cause the
lower equilibrium constant. This condition indicates that chemical reaction tend to be depleted. Creation of the
chemical reaction depends on the estimation of hydrocarbon type. Reactants and products define possibly of
phase equilibrium which is correlated to Gibbs free energy.

Phase Behavior of Dry Gas Reservoir Fluid

The thermodynamic simulation indicates that the phase diagram on the pressure-temperature construction
t y p i f i e s the behavior of a dry gas reservoir. If the pressure and temperature are reduced from the original
reservoir conditions to standard stock tank conditions (condensate tank), there is no liquid recovery and the
reservoir fluid remains completely in the gaseous phase during the process. The pressure-temperature curve for

1042
hydrocarbon reservoir can be classified in terms of its phase envelope. The typical components of gas production
from petroleum reservoirs can explain in their phase envelopes.

Temperature (oC)
-50 0 50 100 150 200
300

250
y = 0.993x + 126.1
R² = 0.997 C6 H14
y = 0.833x + 110.6 200

Pressure (bar.g)
C5 H12 R² = 0.992

C4 H10 150
C3 H 8
C2 H 4 100

Critical point
50
CriT-CriP

0
Fig. 7. Pressure as a function of temperature for each chemical reaction at critical point.

Temperature (oC)
-50 0 50 100 150 200 250
250
Crirical point
CriT - CriP C6 H14

C5 H12 200

C4 H10
Pressure (bar.g)

150

100
C2 H6

50

-50
Fig. 8. Pressure as a function of temperature at critical point and indicating cricondenbar (CriP) and cricodentherm (CriT) of overall data
corresponding to Fig. 7.

1043
CONCLUSIONS
Hydrocarbon in Sin Phu Horm Gas Field is studied the relationship of chemical species and predicted phase by
thermodynamic simulation. This inquiry is valuable research to comprehend the phase equilibria of natural gas and
condensate. Their thermodynamic concepts base on physical and chemical characteristics in equilibrium condition.
Phase envelope can be evaluated at suitable phase. The condensate samples are collected from the gas processing
plant, Hess (Thailand) Limited, where is sampling on August, 18th and October, 2nd, 2012.
The series of sample collection are followed based on inlet slug catcher, inlet high pressure flash vessel, outlet
high pressure flash vessel, inlet produced water separator, outlet produced water separator, low pressure flash vessel
condensate tank A and condensate tank B. Their physical properties and chemical composition of natural gas and
condensate are examined. The physical properties of condensate are analyzed in the basis of API gravity, density,
Reid vapor pressure (RVP), true vapor pressure (TVP), pressure and temperature. The API gravity of condensate in
each process is presented in the range of 55-63o. Pressures in each process are decreased from inlet slug catcher to
the condensate tank, which the maximum pressure is 747.54 bar.g of the inlet slug catcher. Methane is discovered
approximately 96%mole and other hydrocarbon gases 0.1-1%.
Possibility of chemical reaction depending on chemical species is assumed by equilibrium condition
between natural gas and condensate. Five variable chemical reactions are selected by mainly methane. The
interest reactants are paired between methane and another hydrocarbon combined to carbon dioxide gas,
hydrogen gas and nitrogen gas. The results from experiment can be beneficial for evaluating an assumption to
predict the phase equilibrium depending on the chemical reaction. The Prode Properties is a computer simulation
program which shows characteristics of phase equilibrium of each component. Phase envelope of natural gas show
that temperature is inversely proportional to equilibrium constant. Low equilibrium constant at high temperature
indicates that chemical reaction tends to be move further slowly. If the number of carbon atoms increases from 1 to 6,
the temperature and pressure will increase gradually. Hydrocarbon types and mole fraction of these components
affect to the occasion of phase envelope.

Acknowledgements
I would like to thank Khon Kaen University, Department of Geotechnology for giving me a chance to conduct this
research. Also, I would like particularly to express my appreciation to Hess (Thailand) Limited to provide all samples
and information for this research. This project could not successfully complete without the kindness of the Chemical
Process and Engineering Program, Department of Mechanical and Process Engineering, The Sirindhorn
International Thai-German Graduate School of Engineering (TGGS) at King Mongkut’s University of Technology
North Bangkok (KMUTNB) to support the GC and several facilities for analyzing all samples.

REFERENCES
Alfradique, M. F. and Castier, M. 2007.Calculation of phase equilibrium of natural gases with the Peng-Robinson
and PC-SAFT equations of state. Oil & Gas Science and Technology 62:707-714.
Atilhana, M., S. Aparicio-Martinezb, K. R. Hallc, M. El-Halwagic and G. A. Iglesias-Silva. 2010. On the PρT, PηT
and phase envelope behavior characterization of Qatari type natural gas mixtures. Proceedings of the 2nd Annual
Gas Processing Symposium. Qatar, January 10-14, 2010. 399-407.
Ayala, L. H. 2006. The role of fluid thermodynamic behavior in natural gas handling and transportation.Business
Briefing: Hydrocarbon World 2006. London. UK: Touch Briefings Ltd.
Campbell, J. M. 1994. Gas Conditioning and Processing. 17th Ed. United States of America: Campbell Petroleum
Series.
Freze, R., J. L. Chevalier, A. Peneloux and E. Rauzy. 1983. Vapour-liquid equilibria calculations for normal fluid
systems using a new cubic equation of state. Fluid Phase Equilibria, SPE J., 3:65-80.
Gabitto, J. F. and C. Tsouris. 2010. Physical properties of natural gas. Journal of Thermodynamics 1:1-12.
Michelsen, M. L. 1980. Calculation of phase envelopes and critical points for multicomponent mixtures. Fluid
Phase Equilibria 4:1–10.
Ministry of Energy. 2007. Petroleum field in Thailand. http://www.dmf.go.th/.
Prode Properties. 2010. Phase envelope. http://www.prode.com.
Sattayarak, N. 2005. Petroleum potential of northeast Thailand. International Conference on Geology, Geotechnology
and Mineral Resources of Indochina (GEOINDO’2005). Khon Kaen University. Khon Kaen. Thailand. 21-30.
USGS.2006. http://pubs.usgs.gov/of/2005/1451/equation.html.

1044
 PERFORMANCE ESTIMATION OF GAS TURBINE SYSTEM VIA DEGREE-DAY METHOD
1 2
U. Unver , M. Kılıc
1
Yalova Universitesi, Kazım Karabekir Mah. Rahmi Üstel Cad. No:1, Yalova Turkey
2
Uludag Universitesi, Makine Muhendisligi Bolumu, Görukle Bursa Turkey
umit.unver@yalova.edu.tr, mkilic@uludag.edu.tr
ABSTRACT
Nowadays the focus of the power sector is on renewable energy sources. Even so, in Turkey, the most preferred
power plants are still co-generation power plants, and through co-generation systems, considering installed power,
gas turbine power plants take the largest place. In gas turbine power cycles, both power and efficiency are related
with the atmospheric temperature. In this study, the degree-day method is used to estimate the electric power of a
gas turbine. The installed capacity of the gas turbine is 237 MW. The data are actual running values that are supplied
by Ovaakça Power Plant Administration of Bursa/Turkey. In this paper it is indicated that, for a 45 oC increment of
atmospheric temperature, mass flow through the compressor at a unit time decreases about 85,2 kg. Therefore, the
compression work decreased about 11.45%. In addition, by means of the atmospheric temperature increment, both
the total energy input via natural gas and intake airflow rate were decrease by 17.1% and this result a reduction in net
electric power production.

INTRODUCTION
Energy is the world's one of the most important issue for the last two centuries. Upto recent days, energy demand
was met by fossil fuels. But, because of the negative effects of fossil fuels on the environment, the release of
hazardous gases came into prominence. Therefore, many countries are turning to natural gas, which is a clean
source rather than fuel oil, for easy installation and high efficiency of energy production. Also in our country, in the 90s
when air pollution began to threaten human health, thermal systems were converted to natural gas, to meet domestic
and industrial energy demand. And the conversion was stimulated by government.
The use of high-tech materials and methods in natural gas-fired turbines, which are used in power generation
systems, increased efficiency of energy production. In addition, the emissions of hazardous gases in the flue gases
decreased to controllable limits. However, although quick and easy installation of gas turbine cycles there are also
some disadvantages. One of the most important disadvantages is that the gas turbine systems are highly affected by
the atmospheric conditions. It is indicated that the compressors of gas turbines are designed to pass a constant
volumetric flow (Arrieta and Lora 2001). When the atmospheric temperature raises specific volume increases. Thus,
through a constant volume the mass flow rate decreases and as a result both the useful power that is achieved from
the gas turbine and the thermal power that is achieved from the heat recovery steam generator decrease. (Bassily
2001) was studied the interaction between atmospheric temperature and the compressor work, as well as their effect
on systems performance, also studied the efficiency improvement by cooling of inlet air via evaporative coolers.
Similarly Alhazmy has addressed the issue to the inlet air cooling via the spray cooling and cooling coil methods
(Alhazmy 2001). Gareta and Gil also focused on cooling the inlet air by various methods (Gareta and Gil 2004).
Arpacı and Binark were examined a natural gas fired co-generation cycle and described the algorithm of calculation
of system parameters in details.
All the literature mentioned up to here declare that the increase of the atmospheric temperature has a negative effect
on gas turbine cycles performances. Thus they all agreed that the selection of the place where the power plant will be
installed has significant importance. However, studies which are investigating the gas turbine systems were not
clearly stated that how climate impact should be calculated.
In this paper, we have studied to determine a relationship between the Degree Day Method, which is generally used
to determine the optimum insulation thickness of buildings, and net electric power produced by gas turbine cycles. In
this way, primarily, it will be possible to determine the atmospheric affect on the net electric power production, and
secondarily, to compare different climatic locations with reliable numerical data would be possible.
The degree day method that we used in this study, was described in detail in (Kaynaklı et. al 2012). They studied to
determine the optimum insulation thickness considering solar energy via degree day method.

MATERIAL
The data were obtained from a operating, 501F type gas turbine, which is produced by Mitsubishi Heavy Industries.
At the topping cycle electric power is generated by natural gas combustion. Additionally, exhaust gases about 550 oC,
are sent to Heat Recovery Steam Generator (HRSG) to obtain more electric power from the bottoming cycle, without
using extra fuel.
The natural gas consumption is about 14 kg/s at full load while the atmospheric temperature is 15 oC. Annual average
of Lowest Heating Value (LHV) is calculated about 45000 kJ/kg, from monthly analysis result data of natural gas,
which is given by power plant administration. The combustion system of 701F consists of 20 combustion cells.
Combustion cells consist of 2 stage burners and a by-pass valve. Cooling cycle of turbine consist of rotor cooling

1045
cycles and 4 stationery cooling cycles. Cooling air is obtained from various stages of turbine blades. The compressor
and turbine have single shaft on couple of bearings.

Characteristic data for gas turbine;

Manufacturer : Mitsubishi Heavy Industries / Japan
Type : Axial flow – 701F
Gross Output Power / Efficiency : 239 000 kW / 701 F
NOx emission : 50 mg/m3
Number of blade stage :4
Inlet Temperature : 1350 oC (ISO)
Exhaust Temperature : 558 oC
Revolution per Minute : 3000 rev/min.
Critic Rev. : 1. Critic 930, 2. Critic 2550
Start up time : 1200 sec
Combustion chamber type /num. : Multi cell / 20
Ignition : Spark
Compressor Type : Axial Flow
Stage Number : 17
Compression Ratio : 16
Air Flow (ISO) : 580 kg/s (%100 load, 15 oC)

METHOD
The Degree-day method is a powerful tool to calculate heating and cooling loads of buildings. First a base
temperature is defined then from the difference of mean temperature of each day of the year and the base
temperature, heating and cooling degree-days can be obtained. A wide range of knowledge can be obtained from the
recent studies [7-2]. Generally;

(1)

Equation is used to define degree-day.

Gas turbines are very sensitive cycles vs. atmospheric conditions and so degree-day. Because of this, it is expected
to flow the same volumetric amount when atmospheric temperature changes. In this case; mass flow rate change,
which means the power required to compress air;

(2)

should change. A change in the mass flow rate also affects the energy that is supplied by natural gas, and finally
affects the net electric power production. The net electric power, which is produced by gas turbine cycle, can be
calculated by:

(3)

RESULTS and DISCUSSION

As it is well known, when the air temperature increases, the specific volume of the air also increases. The stages of a
turbo compressor are meant to be constant. Thus, the amount of volume flow per unit time is considered to remain
unchanged. In this case, for a constant rotation, an increment of volume of unit mass flow through stages means a
reduction in the mass flow rate. According to this statement, the variation of specific volume and mass flow rate with
respect to inlet air temperature difference are given in Figure 1. As it is seen from the figure, with 45 oC increase of
temperature, specific volume raised from 0.76518 m3/kg to 0.89264 m3/kg. The increment is about 0.12746 m3/kg.
For small scale cycles, this value can be considered as negligible, but this slight change affects mass flow rate
dramatically. The mass flow rate was about 621.6 kg/s at -3 oC and decreased down to 536,5 kg/s with the 45 oC
decrease of atmospheric temperature. In other words, for a 45 oC increment of atmospheric temperature, mass flow
through the compressor at a unit time decreases about 85,2 kg. If we consider the working fluid as water with 10 m
head, the potential energy of the fluid will be about 8 MW, which means a considerably high potential.
If the enthalpy is defined as the energy level of the fluid, we can consider that the energy level of the fluid increases
with the increasing atmospheric temperature. And that means enthalpy also increases. Naturally, when the inlet air
temperature of the compressor increases, exit air temperature also increases. Which means the enthalpy of air at the
compressor exit increases. In the Figure 2 the enthalpies at inlet, exit of the compressor and the difference between
these are given. At first glance, the increase of the enthalpy at the exit seems to be useful but if we look in to the
Figure 2, when the atmospheric temperature is at -3 oC the enthalpy is about 684,8 kJ/kg while with 45 oC

1046
temperature increment it increases about 56 kJ/kg and reaches to 740,7 kJ/kg. Mean while inlet air enthalpy was
270,1 kJ/kg at -3 oC and for a 45 oC increment it rises up to 425,5 kJ/kg. In other words, the enthalpy difference,
which can be, calculated via Eq. 2 goes up about 10,8 kJ/kg with the increase of atmospheric temperature about 45
o
C. The increase of difference between enthalpies, means a serious energy loss if we take such a big mass flow rate
in to account.

Figure 1 Variation of specific volume and mass flow rate with respect to atmospheric temperature.

Figure 2 Inlet, exit and the difference air enthalpies at the compressor.

In Figure 3 the variation of the energy supplied to the system with combustion and the compression work vs
atmospheric temperature are given. As it is seen in the figure, decrease of the mass flow rate of air reduces the

1047
compression work affirmatively. In other words, with the increasing atmospheric temperature, the compression work
reduces by means of decrease of mass flow rate of air. In the calculations, compressor efficiency is considered 0,65.
The compression work was obtained 396,6 MW at -3 oC. With 45 oC increase of temperature, 45,4 MW of reduction is
observed and compression work decreased to about 351,2 MWs. On the other hand, analyzing the reduction of
compression work alone is specious. Because the reduction of the energy supplied by natural gas is more dramatic.
The energy supplied by natural gas was about 653,8 MW and with the 45 oC increase of temperature, it reduces
about 112 MW and decreases to 542 MW. That means the reduction of the energy that is supplied by natural gas is
more than the reduction of energy that is used for compression work. Which also means that, by 45 oC increase of
atmospheric temperature, the reduction of the total energy entering to the system is more than the reduction of the
compression work. The main cause of this is, in gas turbine cycles, the atmospheric temperature difference not only
affects the inlet air mass flow rate but also affects the mass flow rate of natural gas negatively. The decrease of flow
rates of fuel and air results more reduction on the energy that is supplied by fuel than the reduction of the
compression work. Consequently, the net electric power produced by the gas turbine cycle also reduces.

Figure 3 Variation of compression work and total input energy vs. atmospheric temperature.

So far, the findings show us, atmospheric temperature has a direct impact on gas turbine cycles’ power production.
Recent studies note that the effect of atmospheric temperature must be taken into consideration during the selection
of the power plants location. Of course, proximity to the sources and the market are very important parameters during
the selection of location of a power plant where it will be installed. In parallel to this, with this study, the relationship
between the electric power achieved through the plant and the effect of climate was given via degree day method.
Degree day method is a method that may expose the impact of atmospheric temperature. can become visible. With
the method, significant comparable numerical results can be obtained by using the degree day method for different
climatic zones during the selection of the plants’ location. The degree day curve of Bursa, where the power plant
considered in this study is located, is given in Figure 4. For the same time interval, the electric power produced by the
power plant is given in Figure 5. As can be seen in the figures, there is a significant similarity attracts the attention
between the degree days and the produced power. The causes of similarity were examined in detail so far. As it is
seen from the figure, the net electric power was about 240 MW during the cold days in the winter. During spring
season, when the atmospheric temperature begins to raise, the electric power began to decrease. When the
atmospheric temperature reaches to maximum during summer, the net electric power dropped to the minimum values
about 200 MW. It is also observed that the slopes of the curves are in counter agreement with each other.

1048
Figure 4. Atmospheric temperature difference vs. days of the year in order.

Figure 5. Net electric power production vs. days of the year in order.

CONCLUSION
In this study, the variation of net electric power achieved from gas turbine cycle was analyzed with respect to
atmospheric temperature. Also a similarity between atmospheric temperature and power production were tried to be
exposed. It was seen that, with the increase of temperature, specific volume increased 16%, as a result, the intake air
flow rate decreased by 13.7%. The reduction of intake air and increment of atmospheric temperature were negatively
affected the total energy input via natural gas, and intake air flow rates in parallel. Both the total energy input via

1049
natural gas and intake air flow rate were decrease by 17.1%. The atmospheric temperature increment also caused to
decrease in compression work by 11.45%. However, if we take in to account that the decrease of compression work
is less than the decrease of total energy input, the compression work relatively seem to be ascending in proportion to
the total energy input. As a result, the net electric power decreases with increasing temperature.
In the study, quite a significant relationship was determined between degree day and net electric power curves. It is
seen that the relation between the net electric power production and atmospheric conditions may be defined
quantitatively. In this way, during the selection of location of power plants, forecasting the production would be
possible even if the power plants were not operating. By using this method, not only possible production but also
unavailability may be forecasted quantitatively. Moreover, the effect of different climate zones may be compared, and
an alternative and clear criterion, except from proximity to market and source parameters, may be achieved.

NOMENCLATURE
DD Degree Day
Tb Base Temperature C
To Atmospheric Temperature C
Wcomp Compression work kW
Mass flow rate kg/s
hi Inlet air enthalpy kJ/kg
hi Exit air enthalpy kJ/kg
comp Compressor efficiency
We Net electric power kW
LHV Lower Heating Value kJ/kg

REFERENCES
1. Arrıeta,FRP. and Lora EES., 2004. Influnce of ambient temperature on combined-cycle power plant
performance. Applied Energy.
2. Bassily AM., Effects of evaporative inlet and after cooling on the recuperated gas turbine cycle. Applied
Thermal Engineering, 2001., Vol 21, pp 1875-1890.
3. Alhazmy MM. and Najjar YSH., Augmentation of gas turbine performance using air coolers. Applied Thermal
Engineering, 2004., Vol 24, pp 415-429.
4. Gareta R. and Gil LMRA., Methodology fort he economic evaluation of a gas turbine air cooling systes in a
combined cycle applications. Energy, 2004., Vol 29, pp 1805-1818.
5. Arpacı İ. ve Binark AK., Doğal gazlı kojenerasyon sistemlerinde ekserji analizi. 7th International Combustion
Symposium 17-18 July 2002, Ankara. Pp 368-375.
6. Kaynaklı O., Özdemir S. And Karamangil İM., Güneş ışınımı ve duvar yönü dikkate alınarak optimum ısıl
yalıtım kalınlığının belirlenmesi. Journal of Faculty of Engineering and Architecture of Gazi University, 2012.,
Vol 27, No 2, pp 367-374.
7. Bolattürk A., Determination of optimum insulation thickness for building walls with respect to various fuels
and climate zones in Turkey. Applied Thermal Engineering, 2006., Vol 26, pp.1301-1309.
8. Bolattürk A., Optimum insulation thickness for building walls with respect to cooling and heating degree-
hours in the warmest zone of Turkey. Building and Environment, 2007.,
9. Ozkahraman HT. And Bolattürk A.,The use of tuff Stone cladding in buildings for energy conservation.
Construction end Building Materials, 2006., Vol 20, pp 435-440.
10. Çomaklı K. And Yüksel B., Optimum insulation thickness external walls for energy saving. Applied Thermal
Engineering, 2003., Vol 23, pp 473-479.
11. Dombaycı ÖA., The environmental impact of optimum insulation thickness for external walls of buildings.
Building and Environment, 2007., Vol 42, pp 3855-3859.
12. Dombaycı ÖA., Gölcü M. and Pancar Y., optimization of insulation thickness for external walls using different
energy sources. Applied Energy, 2006., Vol 83, pp 921-928.

1050
COMBINED EFFECT OF VISCOUS DISSIPATION ON THE COUPLING OF CONDUCTION
AND MIXED CONVECTION ALONG A VERTICAL SLENDER HOLLOW CYLINDER
1 2
A. Kaya and O.Aydin
1
Department of Mechanical Engineering, Kahramanmaras Sutcu Imam University, 46100 Kahramanmaras,
TURKEY, e-mail: kaya38@ksu.edu.tr
2
Department of Mechanical Engineering, Karadeniz Technical University, 61080 Trabzon, TURKEY, e-mail:
oaydin@ktu.edu.tr

ABSTRACT
An analysis is presented to investigate the influences of viscous dissipation and Joule heating in the entire
thermo-fluid dynamic field resulting from the coupling of bouncy forced flow with conduction along one side
of a heated/cooled vertical slender hollow cylinder. The governing equations are transformed into
dimensionless non-similar equations by using set of suitable transformations. The non-similar solutions
using the Keller box method are obtained. The wall conduction parameter, the buoyancy parameter, the
viscous dissipation parameter and the magneto-hydrodynamic parameter are the main parameters. For
various values of these parameters the local skin friction and local heat transfer parameters are determined.
The validity of the methodology is checked by comparing the results with these available in the open
literature and a fairly good agreement is observed.

Keywords: vertical slender hollow cylinder, conjugate heat transfer, buoyancy effect, viscous dissipation,
joule heating

INTRODUCTION
The interaction between the conduction inside and the buoyancy forced flow of fluid along a solid surface is
termed as conjugate heat transfer process (Mamun et al., 2008). In many practical applications, such as
heat exchangers, heaters, nuclear reactors, pipe insulation systems, etc., the effect of conduction within the
solid wall is significant and must be taken into account. Hence, the analysis of this type of heat transfer
mechanisms possesses necessary the coupling of the conduction in the solid body and the convection in
the fluid surrounding it (Chang, 2006a; Kaya, 2011a).
The conjugate heat transfer problem, in which the coupled heat transfer processes between the
solid body (conduction mechanisms) and the fluid flow (convection mechanisms) are considered
simultaneously, has been investigated by several researchers. Mamun et al. (2008) investigated the heat
generation effect on natural convection flow along and conduction inside a vertical flat plate. Miyamoto et al.
(1980) studied two-dimensional conjugate heat transfer problems of free convection from a vertical flat plate
with a uniform temperature or a uniform heat flux at the outside surface of the plate. Sparrow and Chyu
(1982) investigated the conjugate problem for a vertical plate fin with various heat transfer coefficients
under forced convection. They assumed that the heat conduction in the fin was to be one-dimensional.
Hossain (1992) studied the effect of viscous and Joule heating on the flow of an electrically conducting fluid
past a semi infinite plate of which temperature varies linearly with the distance from the leading edge and in
the presence of uniformly transverse magnetic field. Na (1995) investigated the effect of wall conduction on
the heat transfer characteristics of the natural convection over a vertical slender hollow circular cylinder.
Wang (1998) studied the thermo-fluid-dynamic field resulting from the coupling of wall conduction with
laminar mixed convection heat transfer of micropolar fluids along a vertical flat plate. Shu and Pop 1999)
investigated the conjugate heat transfer across a vertical finite wall separating two forced and free
convection flows at different temperatures. They assumed that the heat conduction in the wall was only in
the transversal direction. Pop and Na (2000) reported a numerical study of the steady conjugate free
convection over a vertical slender hollow circular cylinder with the inner surface at a constant temperature
and embedded in a porous medium. Vaszi et al. (2001) investigated two-dimensional conjugate free
convection from an inclined flat plate in a semi-infinite porous medium under the boundary layer
approximation. Jilani et al. (2002) investigated the conjugate heat transfer in a heat generating vertical
cylinder washed by a laminar forced convection boundary layer. Amin (2003) analyzed the influences of
both first and second order resistance, due to the solid matrix of non-darcy porous medium, Joule heating
and viscous dissipation on forced convection flow from a horizontal circular cylinder under the action of
transverse magnetic field. Chang (2006a, 2008) presented a numerical analysis of the flow and heat
transfer characteristics of forced and natural convection in a micropolar fluid flowing along a vertical slender
hollow circular cylinder with wall conduction and buoyancy effects. Chang (2006b) presented a numerical
analysis of the flow and heat transfer characteristics of mixed convection in a micropolar fluid flowing along
a vertical flat plate with conduction effects. It is assumed that the heat conduction in the wall was only in the
transversal direction. Jahangeer et al. (2007) investigated conjugate heat transfer problem pertinent to
rectangular fuel element of a nuclear reactor dissipating heat into an upward moving stream of liquid
sodium. Hsiao (2007) studied a radiative and viscous dissipation effects conjugate heat transfer problem of

1051
a second-grade viscoelastic fluid past a stretching sheet. Alim et al. (2008) investigated the influences of
viscous dissipation and Joule heating in the entire thermo-fluid dynamic field resulting from the coupling of
bouncy forced flow with conduction along one side of a heated flat plate. Hsiao and Hsu (2009) studied a
conjugate mixed convection heat transfer problem of a second-grade viscoelastic fluid past a horizontal flat-
plate fin. Kaya (2011a, 2011b) studied steady laminar magneto-hydrodynamic (MHD) mixed convection
heat transfer about a vertical slender hollow cylinder under the effect of wall conduction for pure and porous
medium.
In the present paper, the studies of the conjugate effects of viscous dissipation and Joule heating
on the coupling of conduction with mixed convection along a heated/cooled vertical slender hollow cylinder
have been investigated. The boundary layer equations governing the flow are reduced to local non-similarity
equations which are solved using the implicit finite difference method (Keller box). Variation in the fluid–
solid interfacial temperature distribution, the local skin friction and local heat transfer parameters as well as
the velocity and temperature profiles are presented to highlight the influence of the wall conduction, viscous
dissipation, magneto-hydrodynamic and buoyancy parameters.

ANALYSIS
This study considers the case of a mhd mixed convection flow past a vertical slender hollow circular
cylinder of length L and outer radius ro (L >> ro). Also, this study considered the case of a vertical slender
hollow cylinder such that the heat conduction within the wall is one-dimensional (Chang 2006a, 2006b). The
physical model and coordinate system are shown in Fig. 1.
x

Tw(x)

u
g
T0
L v

B0

r
ri

r0

u∞,T∞

Figure 1. The schematic of the problem

The gravitational acceleration, g, acts in the downward direction. The temperature and velocity of the fluid at
a distance remote from the cylinder are given by T∞ and u∞, respectively. The temperature of the inside
surface of the cylinder is maintained at a constant temperature of T0, whereT0 > T∞. Other than the density
variation, the remaining fluid properties are assumed to be constant. The temperature difference between
the body surface and the surrounding fluid generates a buoyancy force, which results in an upward
convective flow. A magnetic field of constant strength B 0 is applied in the negative r direction at all times. Its
interaction with the electrically conducting working fluid produces a resistive force in the negative x
direction. This force is called the Lorentz force. The magnetic Reynolds number is assumed to be small so
that the induced magnetic field is neglected. Under foregoing assumptions and taking into account the
Boussinesq approximation and the boundary layer approximation, the system of continuity, momentum and
energy equations can be written (Duwairi, 2005; Alim et al., 2008):

1052
   ru    r 

0 (1)
x r
u u  v    u   B02
u      r   g  T  T    u  u  (2)
x r  r  r  r  
T T v 1   T   v   u 2  B02 2
u
x
  r 
r Pr r r  r   c p
   
 r   c
u  u u  (3)
 p

The viscous dissipation and Joule heating term are included in the energy equation (Eq. (3)). Here
u and υ are the velocity components in the x and r direction, respectively, T is the temperature of the fluid, β
is the coefficient of thermal expansion, v is the kinematic viscosity, ρ is the fluid density, g is the
acceleration due to gravity σ is the electrical conductivity of the fluid and B0 is the magnetic flux density.
The boundary conditions are given by the following:
r  r0 ; u    0, T  Tw  x 
r  ; u  u , T  T (4)
where subscripts w and ∞ refer to the wall and the boundary layer edge, respectively. In addition, Tw(x) is
the outer surface temperature of the cylinder, which is not known a priori. One of objective of the current
study is to predict the outer surface temperature of the cylinder, Tw(x). Therefore, an additional governing
equation is required for the slender hollow cylinder based on the simplification that the wall of cylinder
steady transfers its heat to the surrounding fluid. Since, the outer radius of the hollow cylinder, r o, is small
compared to its length, L, the axial conduction term in the heat conduction equation of the cylinder can be
omitted (Pozzi and Lupo, 1988; Na, 1995; Chang, 2006a). The governing equation for the temperature
distribution within the slender hollow circular cylinder is given by Chang (2006a) and Kaya (2011a) as
follows:
1 d  dT 
r   0; 0  x  L; ri  x  r0 (5)
r dr  dr s

The boundary conditions for the wall of cylinder are given:

At r  ri , Ts  T0 , (6a)

dT T  x, r0 
At the interface (r  r0 ), Ts  T  x, r0  : ks  k f (6b)
dr s r f
where ks and kf are the thermal conductivity of the cylinder and the fluid, respectively. The boundary
conditions given in Eq. (6b) state the physical requirements that the temperature and heat flux of the
cylinder must be continuous across the solid–fluid interface. From Eqs. (5) and (6b), the temperature
distribution Tw at the interface is derived as (Chang, 2006a):
Kf  r  T  x, r0 
Tw  x   T  x, r0   r0 ln  0   T0 (7)
K s  ri  r
To seek a solution, the following dimensionless variables are introduced:
x 1  r 2  r02  T  T
 , ,     x, r   r0  vu x  f  ,  ,
12
Re1/2
x  , (8)
r0 2 x  0  r T0  T
where  ( x, y) is the free stream function that satisfies Eq. (1) with u  1 r   r  and
   1 r   x 
In terms of these new variables, the velocity components can be expressed as
r0  vu   f 
12
1
u  u f  ,     f   f   (9)
r x   2  
The transformed momentum and energy equations together with the boundary conditions, Eqs. (2),
(3) and (4), can be written as
1  f  f 
1    f    f   ff   Ri  Mn  f   1    f   f   (10)
2    

1053
 1  1
1         f    Ec 1    f 
2

Pr Pr 2
(11)
  f 
 Mn Ec  f   f   1    f    
   
with the boundary conditions;
f
  0; f  2  0, f   0,   1  p 1 2 

  ; f   1,  0 (12)
kf r 
where p ln  0  Re1 2 is the conjugate heat transfer parameter. It should be noticed that for the
ks  ri 
limiting case of p = 0, the thermal boundary condition in Eq. (12) on the wall becomes isothermal. Intuitively,
if the heat conductivity of the hollow cylinder is very large, its temperature can be expected to be
approximately uniform at T0. Hence, the magnitude of p determines the importance of the wall heat
conduction effect (Chang, 2006a; Kaya, 2011a and b).
The corresponding dimensionless groups that appeared in the governing equations defined as:
 cp v Gr g  T0  T  r03 u r
Pr   , Ri  2 , Gr  , Re   0 , Mn  Ha Re ,
k  Re v 2
v
 B0 r0   
2 2 12 2
u
Ha  ,   2  , Ec  (13)
  Re  c p T0  T 
where Pr is the Prandtl number, Ri is the Richardson number, Gr is the Grashof number, Re is the
Reynolds number, Mn is the magnetic parameter, Ha is the Hartman number, λ is the transverse curvature
parameter and Ec is the Eccert number.
The dimensionless interfacial temperature distribution θw, is defined as follows (Chang, 2006).
Tw  T
w  (14)
T0  T

NUMERICAL SOLUTION
The system of transformed equations under the boundary conditions, Eqs. (10)–(12), have been solved
numerically using the Keller box scheme, which is proved to be an efficient and accurate finite-difference
scheme (Cebeci and Bradshaw, 1977). Readers are referred to Cebeci and Bradshaw (1977) for the details
of the numerical methods. This is a very popular implicit scheme, which demonstrates the ability to solve
systems of differential equations of any order as well as featuring second-order accuracy (which can be
realized with arbitrary non-uniform spacing), allowing very rapid x or ξ variations (Takhar and Beg, 1997;
Kaya, 2011b).
A set of non-linear finite-difference algebraic equations derived are then solved by using the
Newton quazi-linearization method. The same methodology as that followed by Takhar and Beg (1997) is
followed. Therefore, for the finite-difference forms of the equations, the reader is referred to Takhar and Beg
(1997) for the brevity of the article.
In the calculations, a uniform grid of the step size 0.01 in the η-direction and a non-uniform grid in
the ξ-direction with a starting step size 0.001 and an increase of 0.1 times the previous step size were found
to be satisfactory in obtaining sufficient accuracy within a tolerance better than 10 -6 in nearly all cases. The
value of η∞ = 16 is shown to satisfy the velocity to reach the relevant stream velocity (Kaya, 2011b).
In order to verify the accuracy of the present method, the present results were compared with those
of Chang (2006a), Chen and Mucoglu (1975) and Takhar et al.(2000). The comparison is found to be in
good agreement, as shown in Table 1.

1054
 Table 1. Comparision of the local skin friction and local heat transfer parameters with Pr=0.7, Ri=0.0, Mn=0.0, Ec=0.0 and p=0.0.
1 2 Chen and Mucoglu Takhar et. al. Chang (2006a) Present Study
4  vx  (1975) (2000)
  
r0  u  f   ,0     ,0  f   ,0     ,0  f   ,0     ,0  f   ,0     ,0 

0.0 1.3282 0.5854 1.3281 0.5854 1.3280 0.5852 1.3201 0.5846

1.0 1.9172 0.8669 1.9167 0.8666 1.9133 0.8658 1.8934 0.8599
2.0 2.3981 1.0986 2.3975 1.0963 2.3900 1.0940 2.3822 1.0918
3.0 2.8270 1.3021 - - 2.8159 1.2982 2.8098 1.2902
4.0 3.2235 1.4921 - - 3.2187 1.4925 3.2102 1.4898

RESULTS AND DISCUSSION

The aim of this study was to investigate the flow and heat transfer characteristics for the mixed convection
flow along a vertical slender hollow cylinder with wall conduction effect, taking into account the effects of
Ohmic heating and viscous dissipation. The following ranges of the main parameters are considered: Ri
=0.0, 1.0, 5.0 and 10.0; Ec= -0.1, 0.0 and 0.1; Mn=0.0, 0.5, 1.0 and 2.0 and Pr = 1.0. Hence, the numerical
computations were performed with Re = 250 (Chang, 2006a; Kaya, 2011 a and b). The conjugate heat
transfer parameter, p, is ranged 0.0–0.3 to conform to the practical cases. For example, the fluid (for water:
kf = 0.628 W/m K) flows along a slender stainless steel tubes with r o=0.03 m, ri=0.02 m and ks=15 W/m K,
we can find the conjugate heat transfer parameter p=0.27 (Chang, 2006a).
The combined effects of p, Ri, Mn and Ec on the momentum and heat transfer are analyzed. The
Richardson number, Ri represents a measure of the effect of the buoyancy in comparison with that of the
inertia of the external forced or free stream flow on the heat and fluid flow. Outside the mixed convection
region, either the pure forced convection or the free convection analysis can be used to describe accurately
the flow or the temperature field. Forced convection is the dominant mode of transport when Ri→0,
whereas free convection is the dominant mode when Ri→∞. Buoyancy forces can enhance the surface heat
transfer rate when they assist the forced convection (Kaya, 2011b). Viscous dissipation, as an energy
source, severely distorts the temperature profile. Remember positive values of Ec correspond to wall
heating (heat is being supplied across the walls into the fluid) case (T0 > T∞), while the opposite is true for
negative values of Ec (Aydin and Kaya, 2008).
The effect of conjugate heat transfer parameter p on the velocity (a) and temperature (b) profiles
within the boundary layer with Ri=5.0, Ec = - 0.1, Mn=0.5 and Pr = 1.0 is shown in Fig. 2 (a) and (b),
respectively. The increasing of the conjugate heat transfer parameter decreases velocity and temperature
gradients at the wall. A lower wall conductance ks or higher convective cooling effect due to greater kf
increases the value of p as well as causes greater temperature difference between the two surfaces of the
cylinder. The temperature at the solid-fluid interface is reduced since the temperature at the outside surface
of the cylinder is kept constant (Kaya, 2011a and b).

1.6 1.0

p=0.0 Ri=5
1.4
p=0.1 Mn=0.5
0.8 p=0.2 Pr=1
1.2 p=0.3 Ec= - 0.1

1.0 p=0.0 Ri=5 0.6

p=0.1 Mn=0.5
(1,)
f'(1,)

0.8 p=0.2 Pr=1

p=0.3 Ec= - 0.1
0.4
0.6

0.4
0.2
0.2

0.0 0.0
0 2 4 6 8 0 1 2 3 4 5 6

 

(a) (b)
Figure 2. Dimensionless velocity (a) and temperature (b) profiles for different p while Pr=1.0, Ri =5,Mn=0.5, Ec= -
0.1 and ξ=1.0.

The variation of the interfacial temperature, the local skin friction and the local heat transfer
parameters for different values of p with ξ are shown in Fig. 3(a), (b) and (c), respectively where Ri=5.0,
Mn=0.5, Ec = - 0.1 and Pr = 1.0. It can be seen that the temperature of the fluid on the wall increases

1055
 monotonically with ξ for a given value of p (Fig. 3(a)). Comparing with isothermal cylinder (p = 0), an
increase in the conjugate heat transfer parameter, p, causes a reduction in the interfacial temperature. This
is because an increased value of p corresponds to a lower wall conductance k s and promote a greater
surface temperature variations (Fig. 3a). The increases value of p increases the momentum and thermal
boundary layer thickness and therefore the local skin friction and the local heat transfer parameters
decreases as shown in Fig. 3(b) and (c). Also, increasing the streamwise direction ξ increases the
interfacial temperature, the local skin friction and the local heat transfer parameters (Fig.3 (a), (b) and (c)).
Similar results were found in the literature (Chang, 2006a and b; Kaya, 2011a and b).

4.0
1.00
p=0.0 Ri=5
3.5 p=0.1 Mn=0.5
0.95 p=0.2 Pr=1
p=0.3 Ec= - 0.1
3.0
0.90
2.5

f''(,0)
w

0.85
2.0

0.80
1.5
p=0.0 Ri=5
p=0.1 Mn=0.5
0.75 p=0.2 Pr=1 1.0
p=0.3 Ec= - 0.1

0.70 0.5
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

 

(a) (b)
0.9

0.8

0.7

0.6
-'(,0)

0.5

p=0.0 Ri=5
0.4 p=0.1 Mn=0.5
p=0.2 Pr=1
p=0.3 Ec= - 0.1
0.3

0.2
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

(c)
Figure 3. Effect of conjugate heat transfer parameter p on the dimensionless interfacial temperature (a), local skin
friction (b) and local heat transfer (c) parameters against the streamwise direction ξ whilePr=1.0, Ri =5, Mn=0.5
and Ec= - 0.1.

Figure 4 shows the effects of Ec on the dimensionless velocity and temperature profiles for the
cases of isothermal cylinder (p=0) and non-isothermal cylinder (p=0.2). For a positive value of Ec, since the
temperature of the wall is greater than that of the free stream, there will be a heat transfer from the wall to
the fluid. The viscous dissipation will cause to a heat generation inside the fluid, which results in increase in
the temperature distribution and decrease in the velocity distribution in the flow region. This is due to the
fact that heat energy is stored in the fluid due to frictional heating. Due to the increased bulk fluid
temperature, the temperature gradient will decrease. For a negative value of Ec, since To < T∞, the viscous
dissipation will increase the temperature distribution in the flow region more. Finally, this leads to an
increased temperature gradient, as will be shown later, which will result in increased heat transfer values.
As it can be noticed from Fig. 4, the effect of Ec intensifies downstream (Aydin and K2008).

1056
 1.6 1.0
Ri=5 Ec= - 0.1
p=0.0
1.4 Mn=0.5 Ec= 0.0
p=0.2 Ec= - 0.1 Ri=5 p=0.0
Pr=1 Ec= 0.1
Ec= 0.0 Mn=0.5
0.8 p=0.2
1.2 Ec= 0.1 Pr=1

1.0
0.6
f'(1,)

(,h)
0.8

0.4
0.6

0.4
0.2
0.2

0.0 0.0
0 2 4 6 8 0 1 2 3 4 5 6

 
(a) (b)
Figure 4. Dimensionless velocity (a) and temperature (b) profiles for different Ec while Ri=5, Pr=1.0, Mn=0.5 and
ξ=1.0 for isothermal (p=0) and non isothermal plate (p<0).

Figure 5 shows the variation of the interfacial temperature, the local skin friction and the local heat
transfer parameters for different values of Ec with ξ are shown in Fig. 5(a), (b) and (c), respectively where
Ri=5.0, Mn=0.5 and Pr=.0 for isothermal cylinder (p=0) and non-isothermal cylinder (p=0.2).

0.92 4.0
Ec= - 0.1 Ri=5 p=0.0
Ec= - 0.1 Ri=5 Ec= 0.0 Mn=0.5 p=0.2
0.90 Ec= 0.0 Mn=0.5 3.5 Pr=1
Ec= 0.1
Ec= 0.1 Pr=1
p=0.2
0.88 3.0

0.86 2.5
f''(,0)
w

0.84 2.0

0.82 1.5

0.80 1.0

0.78 0.5
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

 
(a) (b)
0.9

Ec= - 0.1 Ri=5 p=0.0

0.8 Ec= 0.0 Mn=0.5
p=0.2
Ec= 0.1 Pr=1

0.7

0.6
-'(,0)

0.5

0.4

0.3

0.2
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

(c)
Figure 5. Effect of Ec on the dimensionless interfacial temperature (a), local skin friction (b) and local heat transfer
(c) parameters against the streamwise direction ξ while Ri =5, Pr=1.0 and Mn=0.5 for isothermal (p=0) and non
isothermal plate (p<0).

1057
 Increasing the Ec decrease the interfacial temperature and the local skin friction whereas increase
the local heat transfer parameter. Also, both the local skin friction and the local heat transfer parameters
decrease for non-isothermal wall case (Fig. 5a-c) .
Figure 6 shows the dimensionless velocity (a) and temperature (b) profiles inside the boundary
layer for different values of the buoyancy parameter Ri for the cases of isothermal cylinder (p=0) and non-
isothermal cylinder (p=0.2). The increasing of Ri increases velocity and temperature gradients at the wall.
And also, increasing the conjugate heat transfer parameter decreases the velocity and temperature profiles
in the boundary layer.
2.0 1.0

Ri=0.0
Ri=0.0 Ec= - 0.1 p=0.0 Ec= - 0.1 p=0.0
Ri=1.0
Ri=1.0 Mn=0.5 Mn=0.5
p=0.2 0.8 Ri=5.0 Pr=1 p=0.2
Ri=5.0 Pr=1
Ri=10.0
1.5 Ri=10.0

0.6

(1,)
f'(1,)

1.0

0.4

0.5
0.2

0.0 0.0
0 2 4 6 8 0 1 2 3 4 5 6

 

(a) (b)

Figure 6. Dimensionless velocity (a) and temperature (b) profiles for different Ri while Ec = - 0.1, Pr=1.0, Mn=0.5
and ξ=1.0 for isothermal (p=0) and non isothermal plate (p<0).

The variation of the dimensionless interfacial temperature distributions, the local skin friction and
the local heat transfer parameters as a function of ξ in the boundary layer are shown in Fig. 7 (a), (b) and
(c), respectively. It can be seen that as the values of ξ increases, the interfacial temperature rises, i.e. the
interfacial temperature increases from the bottom of the cylinder towards the top of the cylinder. Compared
with the limiting case of Ri = 0 (i.e. pure forced convection), an increase in the value of Ri gives rise to a
reduced interfacial temperature since a greater value of Ri indicates a greater buoyancy effect, which
increases the convection cooling effect and hence reduces the wall temperature (Fig. 7(a)). Fig. 7(b)
illustrates the effect of the buoyancy force on the local skin friction factor for p=0 (dashed lines) and 0.2
(solid lines). It is observed that the local skin friction factor increases with the buoyancy effect. The reason
for this is that an increase in the buoyancy effect in mixed convection flow leads to an acceleration of the
fluid flow, which increases the local skin friction factor. Additionally, the higher the value of the buoyancy
effect, the more the sensitivity of the wall conduction effects influences the skin friction factor. In the limiting
case of Ri = 0, i.e. pure forced convection with no buoyancy effects, the conjugate heat transfer parameter,
p, is virtually independent of the skin friction factor because the buoyancy effect generated by the
temperature difference is relatively weaker. In Fig. 7(c), the effect of the buoyancy force on the local heat
transfer parameter is illustrated. It is noted that as the value of the Ri increases, the local heat transfer
parameter also increases, both for the case of an isothermal cylinder (dashed lines) and a non-isothermal
cylinder (solid lines). This is because an increased buoyancy effect generates a greater buoyancy force,
which increases the fluid velocity, and hence raises the local heat transfer parameter (Chang, 2006b; Kaya,
2011b).

1058
 0.95 6
Ri=0.0 Ec= - 0.1
Mn=0.5 Ri=0.0 Ec= - 0.1 p=0.0
Ri=1.0 Ri=1.0
Pr=1 Mn=0.5
Ri=5.0 5 Ri=5.0 p=0.2
p=0.2 Pr=1
Ri=10.0 Ri=10.0
0.90

f''(,0)
0.85
w

2
0.80

0.75 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
 
(a) (b)
1.2
Ri=0.0
Ec= - 0.1 p=0.0
Ri=1.0
Mn=0.5
Ri=5.0 Pr=1 p=0.2
1.0 Ri=10.0

0.8
-'(,0)

0.6

0.4

0.2
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

(c)
Figure 7. Effect of Ri on the dimensionless interfacial temperature (a), local skin friction (b) and local heat transfer
(c) parameters against the streamwise direction ξ while Ec = - 0.1, Pr=1.0 and Mn=0.5 for isothermal (p=0) and
non isothermal plate (p<0).

Figure 8 shows the dimensionless velocity (a) and temperature (b) profiles inside the boundary
layer for different values of the magnetic parameter Mn for the cases of isothermal cylinder (p=0) and non-
isothermal cylinder (p=0.2). The increasing of the magnetic parameter Mn increases velocity and
temperature gradients at the wall. As mentioned above, increasing the conjugate heat transfer parameter p
decreases velocity and temperature gradients at the wall.
The variations of the dimensionless interfacial temperature distributions, the local skin friction and
the local heat transfer parameters as a function of ξ in the boundary layer for different magnetic parameter
are shown in Fig. 9 (a), (b) and (c), respectively. Increasing the magnetic parameter Mn decreases the
interfacial temperature (Fig. 9(a)). The local skin friction and the local heat transfer parameters with different
values of Mn for isothermal cylinder (p=0) and non-isothermal cylinder (p=0.2) are illustrated in Fig. 9(b) and
(c). The magnetic force aiding the flow and increases the local skin friction (i.e. shear stress) and the local
heat transfer (i.e. heat transfer rate) parameters at the wall.

1059
 2.0 1.0
Mn=0.0 Ec= - 0.1 p=0.0
Mn=0.5 Ri=5.0 p=0.2
Mn=1.0 Pr=1.0
0.8
Mn=2.0
1.5

0.6

(1,)
f'(1,)

1.0
0.4
Mn=0.0 p=0.0
Ec= - 0.1
Mn=0.5 Ri=5.0
Mn=1.0 p=0.2
Pr=1.0
Mn=2.0 0.2
0.5

0.0

0.0
0 2 4 6 8 0 1 2 3 4 5 6

 
(a) (b)
Figure 8. Dimensionless velocity (a) and temperature (b) profiles for different Mn while Ec = - 0.1, Pr=1.0, Ri=5
and ξ=1.0 for isothermal (p=0) and non isothermal plate (p<0).
0.90 4.5
Mn=0.0 Ec= - 0.1 Mn=0.0 Ec= - 0.1
p=0.0 p=0.0
Mn=0.5 Ri=5.0 4.0 Mn=0.5 Ri=5.0
Mn=1.0 Pr=1.0 p=0.2 Mn=1.0 Pr=1.0 p=0.2
0.88
Mn=2.0 Mn=2.0
3.5

0.86 3.0
f''(,0)
w

2.5

0.84
2.0

1.5
0.82

1.0

0.80 0.5
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

 

(a) (b)
1.2
Mn=0.0
Ec= - 0.1 p=0.0
Mn=0.5
Ri=5.0
Mn=1.0 Pr=1.0 p=0.2
1.0 Mn=2.0

0.8
-'(,0)

0.6

0.4

0.2
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

(c)
Figure 9. Effect of Mn on the dimensionless interfacial temperature (a), local skin friction (b) and local heat
transfer (c) parameters against the streamwise direction ξ while Ec = - 0.1, Pr=1.0, Ri=5 for isothermal
(p=0) and non isothermal plate (p<0).

1060
CONCLUSIONS
This study has analyzed the influences of viscous dissipation and Joule heating in the entire thermo-fluid
dynamic field resulting from the coupling of bouncy forced flow with conduction along one side of a
heated/cooled vertical slender hollow cylinder. The nonlinear formulation governing equations and their
associated boundary conditions have been obtained and solved using the non-similarity transform and the
finite difference method (Keller box), respectively. The influences of the conjugate heat transfer parameter,
the viscous dissipation parameter, the Richardson number and the magnetic parameter on the solid–liquid
interfacial temperature distribution, the local skin friction and the local heat transfer parameters have been
systematically examined. From the present numerical investigation, the following conclusions can be drawn:
1. An increase in the conjugate heat transfer parameter decreases the velocity and the temperature
gradient and therefore decreases the dimensionless interfacial temperature distribution, the local
skin friction and the local heat transfer parameters.
2. An increase in the magnetic and buoyancy parameters increase the local skin friction and local
heat transfer parameters and decrease the dimensionless interfacial temperature distributions.
Increasing Mn and Ri decreases the velocity and temperature gradients at wall for non-isothermal
cylinder (i.e. p > 0).
3. An increase in the viscous dissipation parameter increase the local skin friction and the
dimensionless interfacial temperature distributions and decrease the local heat transfer parameter.

NOMENCLATURE
cp specific heat of the convective fluid
Ec Eccert number
g gravitational acceleration
Gr Grashof number
f dimensionless stream function
Ha Hartman number
k thermal conductivity
L length of the cylinder
Mn magneto-hydrodynamic parameter
p conjugate heat transfer parameter
Pr Prandtl number
ri, ro inner and outer radii of the hollow cylinder
Re Reynolds number
Ri Richardson number
T temperature
T0 temperature of the inside surface of the cylinder
u, v velocities in x and r directions, respectively
x,r coordinates in axial and radial directions, respectively
Greek symbols
 similarity variable
λ transverse curvature
 dimensionless streamwise coordinate
ρ fluid density
μ dynamic viscosity
υ kinematic viscosity
θ dimensionless temperature profile in Eq. (8)
θw dimensionless interface temperature in Eq.(14)
Subscripts
f condition in the fluid
s condition in the wall of cylinder
w wall
∞ free stream

REFERENCES

Alim, M.A., Alam, Md M., Mamun, A. and Hossain, B., 2008. Combined effect of viscous dissipation and joule
heating on the coupling of conduction and free convection along a vertical flat plate. International
Communications in Heat and Mass Transfer 35: 338–346.
Aydin, O. and Kaya, A., 2008. Non-Darcian Forced Convection Flow of Viscous Dissipating Fluid over a Flat Plate
Embedded in a Porous Medium. Transport in Porous Medium 73: 173–186.
Cebeci T. and Bradshaw, P., 1977. Momentum Transfer in Boundary Layers. Washington, DC: Hemisphere.

1061
Chang, C.L., 2006b. Numerical simulation of micropolar fluid flow along a flat plate with wall conduction and
buoyancy effects. Journal of Physics D: Applied Physics 39: 1132–1140.
Chang, C.L., 2008. Numerical simulation for natural convection of micropolar fluids flow along slender hollow
circular cylinder with wall conduction effect. Communications in Nonlinear Science and Numerical Simulation 13:
624–636.
Chang, L., 2006a. Buoyancy and wall conduction effects on forced convection of micropolar fluid flow along a
vertical slender hollow circular cylinder. International Journal of Heat and Mass Transfer 49: 4932–4942.
Chen, T.S. and Mucoglu, A., 1975. Buoyancy effects on forced convection along a vertical cylinder. ASME-
Journal of Heat Transfer 97: 198–203.
Duwairi, H.M., 2005. Viscous and joule heating effects on forced convection flow from radiate isothermal porous
surfaces. International Journal of Numerical Methods Heat and Fluid Flow 15: 429–440.
El-Amin, M.F., 2003. Combined effect of viscous dissipation and Joule heating on MHD forced convection over a
non isothermal horizontal cylinder embedded in a fluid saturated porous medium. Journal of Magnetism and
Magnetic Materials 263: 337–343.
Hossain, M.A., 1992. Viscous and Joule heating effects on MHD-free convection flow with variable plate
temperature. International Journal of Heat and Mass Transfer 35: 3485–3487.
Hsiao, K.L. and Hsu, C.H., 2009. Conjugate heat transfer of mixed convection for viscoelastic fluid past a
horizontal flat-plate fin. Applied Thermal Engineering 29: 28–36.
Hsiao, K.L., 2007. Conjugate heat transfer of magnetic mixed convection with radiative and viscous dissipation
effects for second-grade viscoelastic fluid past astretching sheet. Applied Thermal Engineering 27: 1895–1903.
Jahangeer, S., Ramis, M.K. and Jilani, G., 2007. Conjugate heat transfer analysis of a heat generating vertical
plate. International Journal of Heat and Mass Transfer 50: 85–93.
Jilani, G., Jayaraj, S. and Ahmadi, M.A., 2002. Conjugate forced convection-conduction heat transfer analysis of a
heat generating vertical cylinder. International Journal of Heat and Mass Transfer 45: 331–341.
Kaya, A., 2011a. The effect of conjugate heat transfer on MHD mixed convection about a vertical slender hollow
cylinder. Communications in Nonlinear Science and Numerical Simulation 16: 1905–1916.
Kaya, A., 2011b. Effects of buoyancy and conjugate heat transfer on non-Darcy mixed convection about a vertical
slender hollow cylinder embedded in a porous medium with high porosity. International Journal of Heat and Mass
Transfer 54: 818–825.
Mamun, A.A., Chowdhury, Z.R., Azim, M.A. and Maleque, M.A., 2008. Conjugate heat transfer for a vertical flat
plate with heat generation effect. Nonlinear Analyses: Modeling and Control 13(2): 213–223.
Miyamoto, M., Sumikawa, J., Akıyohı, T. and Nakamura, T., 1980. Effects of axial heat conduction in a vertical flat
plate on free convection heat transfer. International Journal of Heat and Mass Transfer 23: 1545–1553.
Na, T.Y., 1995.Effect of wall conduction on natural convection over a vertical slender hollow circular cylinder.
Applied Science Resources 54: 39–50.
Pop, I. and Na, Y., 2000. Conjugate free convection over a vertical slender hollow cylinder embedded in a porous
medium. Heat and Mass Transfer 36: 375–379.
Pozzi, A. and Lupo, M., 1988. The coupling of conduction with laminar natural convection along a flat plate.
International Journal of Heat and Mass Transfer 31(9): 1807–1814.
Shu, J.J. and Pop, I., 1999. Thermal interaction between free convection and forced convection along a vertical
conducting wall. Heat and Mass Transfer 35: 33–38.
Sparrow, E.M. and Chyu, M.K., 1982. Conjugated forced convection–conduction analysis of heat transfer in a
plate fin. ASME-Journal of Heat Transfer 104: 204–206.
Takhar, H.S. and Beg, O.A., 1997. Effects of transverse magnetic field, Prandtl number and Reynolds number on
non-darcy mixed convective flow of an incompressible viscous fluid past a porous vertical flat plate in a saturated
porous medium. International Journal of Energy Resource 21: 87–100.
Takhar, H.S., Chamkha, A.J. and Nath, G., 2000. Combined heat and mass transfer along a vertical moving
cylinder with a free stream. Heat and Mass Transfer 36: 237–246.
Vaszi, A.Z., Ingham, D.B., Lesnic, D., Munslow, D. and Pop, I., 2001. Conjugate free convection from a slightly
inclined plate embedded in a porous medium. ZAMM-Zeitschrift fur Angewandte Mathematik und Mechanik 81(7):
465–479.
Wang, T.Y., 1998. The coupling of conduction with mixed convection of micropolar fluids past a vertical flat plate.
International Communications in Heat and Mass Transfer 25(8):1075–1084.

1062
 THREE DIMENSIONAL NUMERICAL ANALYSIS OF THERMAL OUTPUT OF
A STEEL PANEL RADIATOR

1 2 1 3 3
M. Kılıç , G. Sevilgen , M. Mutlu , M.Kaya , and T.Sığırtmaç
1
Uludağ Üniversitesi, Mühendislik Mimarlık Fakültesi, Makine Mühendisliği Bölümü, Görükle Kampüsü, Bursa
2
Uludağ Üniversitesi, Mühendislik Mimarlık Fakültesi, Otomotiv Mühendisliği Bölümü, Görükle Kampüsü, Bursa
3
Coşkunöz A.Ş. Bursa
e-mail: mkilic@uludag.edu.tr , e-mail: gsevilgen@uludag.edu.tr , e-mail: mustafamutlu@uludag.edu.tr,
e-mail: mekaya@coskunoz.com.tr , e-mail: tsigirtmac@coskunoz.com.tr

ABSTRACT
Nowadays, the new developments in properties of insulation materials used in buildings and international
regulations lead to use of more efficient heating systems. On the other hand, the increased energy efficiency
of modern buildings makes necessary the use of low temperature heating systems. Thus, we need to design
a new generation of energy efficient radiators which have low capacity of water volume and low weight
compared to the conventional radiators. Moreover, the standards defined for testing procedures of thermal
output of steel panel radiators have been changed depending on the objectives of energy efficiency. In this
content, detailed analysis of thermal output of a panel radiator should be performed to meet the needs for
these objectives. Three dimensional Computational Fluid Dynamics(CFD) is a useful tool for thermal output
analysis of a steel panel radiator. In this study, three dimensional numerical calculations of thermal output of
a steel panel radiator, in accordance with TS EN442, was performed by using CFD analysis based on finite
volume method. The numerical calculations were performed under steady-state conditions. The results
obtained from the numerical simulations were in good agreement with the experimental data available in the
literature.

Keywords: Panel radiator, thermal output, CFD (Computational Fluid Dynamics)

INTRODUCTION
Steel panel radiators are the most common used elements for heating systems of indoor environments
such as buildings and offices. This heating element was chosen for energy efficiency and simple compact
design for many years in the world. On the other hand, design configurations should be performed due to
new developments in insulation materials, new international regulations and testing procedures and the
consumer demands for thermal comfort level. Moreover, the displacement, location, surface area and
operating temperature of the panel radiator(s), heat transfer characteristics of the wall surfaces and surface
area of the window(s) have great impacts on the thermal comfort and energy consumption [1]. Thermal
sensation of human body is dependent on local heat transfer characteristics of the human body surfaces [2].
For this reason, consideration of local heat transfer characteristics of human body surface is highly important
for numerical calculations with sufficient accuracy and acceptable results. A combined numerical simulation
was performed modeling for airflow, thermal radiation, heat and moisture transfer with thermo physiological
properties between a standing human body and its surroundings in a adiabatic room [3]. Considering the
energy efficiency, studies show that low temperature heating may improve indoor air quality as well as the
thermal comfort conditions [4,5]. Experimental studies show that people living in houses with low
temperature heating systems are more satisfied with indoor climate than people living in houses with high
temperature heating systems [6,7]. Previous researches for thermal comfort used to be carried out by means
of experiments. But, performing experiments with consideration of the all parameters reveals complexity and
difficulty. However, researchers have also used combined numerical simulations based on CFD method with
experimental studies [8,15). Analysis of complex HVAC systems based on numerical calculations with
sufficient accuracy and acceptable results is now possible for HVAC researchers due to improved computer
technology and CFD techniques [16,17]. All considerations described above, it can be said that thermal
output of a steel panel radiator is the main important parameter for design configurations and heating
performance. For this purpose, detailed analysis of thermal output of a steel panel radiator should be
performed to meet the needs for these objectives. In this study, three dimensional steady state numerical
calculations of thermal output of a steel panel radiator, in accordance with TS EN442, was performed by
using CFD method.

1063
METHOD
Today there are many computer software packages for CFD analysis. In this study Ansys Fluent software is
used for three dimensional flow field and heat transfer analysis. Fluent software solves continuum, energy
and transport equations numerically with natural convection effects [18]. Sectional view of steel panel
radiators and the CAD Model of the steel panel radiator used in the numerical study are shown in Fig.1 and
Fig.2. In the numerical simulations, we used original dimensions and geometry of a steel panel radiator
without any simplifications in geometrical detail and we chose type-10 steel panel radiator, according to TS
EN442, and the type of fittings was TBSE (Top-Bottom-Same-End) connection.

Figure-1 Sectional view of steel panel radiator

Constructions of steel panel radiators include panels, convectors (fins), grilles and T connections in general.
In the numerical calculations, type-10 radiator model (Fig.2 (b)) which includes one panel without any
extended surfaces (convectors) was used. The dimension of steel panel radiator is shown in Table-1. Water
flow volume (Fig.3) extraction was achieved from original CAD data of steel panel radiator to get precise
velocity and temperature fields of real radiator.

Table-1 The Dimensions of Steel Panel Radiator(s) used in this study

Type Connection type Length(mm) Height(mm) Width(mm)
Type-10 TBSE 2000 mm 600 mm 13 mm

The boundary surfaces defined in the computational model are shown in Fig.3. In numerical calculations
mesh generation is very crucial for getting accurate predicted results and reducing computing time. The
mesh structure of the surfaces of the radiator includes triangular elements. The computational domain
consists of about 2.300.000 volume cells (Fig.4, Table-2).

Table-2 Mesh structure and boundary conditions

Mesh structure
Element type Tetrahedral
Number of elements 2.300.000
Number of nodes 450.000

1064
 Convection

Conduction

T-fittings Radiation

Convectors

Panel

(a)

Water inlet

Water outlet

(b)
Figure-2 The detailed CAD model of the steel panel radiator (a) Type-22, (b) Type-10

Heat is transferred from warm water inside a radiator to air and room surfaces surrounding the radiator. In
the numerical simulations, we calculated thermal output of radiator from Eq.1 and we just employed on the

-1 -1
parameters shown in the left side of the Eq.1. The parameters on the left side, m(kg/s) , cp (W s kg K ),
Tinlet and Toutlet are the mass flow of water inside the radiator, specific heat capacity of water and temperature
difference between inlet and outlet water, respectively. The parameters on the right side, U (W/m K), A(m ),
2 2

Tln are the total heat transfer coefficient, total surface area of the radiator and logarithmic temperature
difference between radiator surface and indoor air.

  m c p (Tinlet  Toutlet )  UATln (1)

1065
 Water inlet

Water outlet

Wall surfaces

Figure-3 Water flow volume extraction from the original CAD data of steel panel radiator

Figure-4 Mesh Structure of computational domain for water flow volume

In the solution process, numerical calculations were performed under steady-state conditions. Water flow
was assumed as laminar and Simple C algorithm was chosen for velocity-pressure coupling and cell based
option was set for flow field computations (Table-3). The first order schemes were employed in the
discretization of the all governing equations at the beginning of the numerical simulations and then the
second order scheme was chosen depends on convergence criteria and getting precise results. The
−4
convergence criterions were assumed when the normalized residuals of flow equations are less than 10
−6
and the energy equations are less than 10 .

1066
Table-3 Solver settings and turbulence model used in the numerical simulations
Solver settings and turbulence model
Time Steady-state
Solver Pressure Based/Coupled
Gradient option Least squares cell-based
Formulation First-order implicit scheme
Equations Combined simulation of flow and energy
Flow Laminar

Table-4 Three different case obtained from experimental results

Cases Tinlet (°C) Mass-flow rate (kg/s) Tair (°C)
Case-I 85.7 0.027719 20.4
Case-II 74.9 0.027822 20.5
Case-III 52.4 0.027972 19.8

Table-5 Boundary conditions used in numerical simulations

Boundary conditions

Surface Boundary-type Properties

Tinlet=Texp (obtained from experiments)
Water inlet Mass flow inlet Mass-flow rate:(obtained from experiments)
Water outlet Pressure outlet Atmospheric conditions
hm is computed from Eq.2-3,
Panel surfaces and Ambient temperature Tair (obtained from experiments)
T-fittings connections Wall thickness 0.0012 mm
Material Steel

Three different cases were considered according to standard test conditions of thermal output of a steel
panel radiator. In all three cases, the value of water inlet temperatures were set to 85.7°C, 74.9°C, 52.4°C
and the value of entering mass-flow rates were set to 0.027719 kg/s, 0.027822 kg/s, 0.027972 kg/s,
respectively. Thus, the leaving temperature of water was computed to calculate the thermal output of radiator
in all cases. The reference air temperature values were obtained from the experimental results and they
were set to 20.4°C, 20.5°C, 19.8°C, respectively (Table-4). The boundary conditions used for all cases in the
numerical simulations were shown in Table-5. At the water outlet surfaces, atmospheric boundary conditions
were used for calculating the leaving temperature of water. Thickness of the panel surfaces were set to
0.0012 m and the material type of the panel radiator was steel. At the panel surfaces, boundary condition
was set which include convection, and shell conduction zone. At the outer surfaces of the panel radiator
natural convection occurs thus, we used UDF (User Defined Function) in Fluent for computation of heat
transfer coefficient (Eq.1) at the outer surfaces of the panel radiator surfaces. UDF is a function that you
program that can dynamically loaded with the Fluent solver to enhance the standard features of the code.
More descriptions and detailed information can be obtained from reference [19]. We used correlation for Nu
(Nusselt) number as described in Eq.2 [20], where, Ra, Pr and Gr are the dimensionless numbers as they
known Rayleigh, Prandtl and Grashof numbers respectively. hm (W/m K) is mean heat transfer coefficient, kf
2

(W/mK) is the thermal conductivity of fluid(water) and L (m) is the length of the vertical plate.

hm L
Nu L  (2)
kf
2
 0.387 Ra L
1/ 6

Nu L  0.825  8 / 27 
; Ra L  GrL Pr (3)
 
1  (0.492 / Pr)9 /16 

1067
RESULTS AND DISCUSSIONS
Computed values of thermal output (W) and temperature (°C) of leaving water from radiator are shown in
Table-6 and the temperature distributions obtained from the numerical simulations are shown in Fig.5-7.

Table-6 Three different case obtained from experimental results

Toutlet (°C) Toutlet (°C) Q (W) Q (W) Difference
Cases Tinlet (°C)
(CFD) (Exp.) (CFD) (Exp.) (%)
Case-I 85.70 72.79 73.40 1501.20 1431 5
Case-II 74.90 64.68 65.20 1191.04 1131 5
Case-III 52.40 47.08 47.40 622.12 583 7

In case-I, thermal output was computed about 1500 W and the difference in percentage between
experimental and CFD data was 5%. From the numerical results shown in Table-6, we can say that the
predicted results both temperature and thermal output are in good agreement with the experimental data.

Figure-5 Temperature contours of panel radiator surface (Case-I)

Figure-6 Temperature contours of panel radiator surface (Case-II)

1068
Figure-7 Temperature contours of panel radiator surface (Case-III)

Figure-8 Temperature fields by radiator 10 X 500 x 2000 mm (single-side top connection) [8]

Figure-9 The locations of the lines described for temperature results

1069
The cooler and warmer area of the whole radiator surfaces were at the bottom and the top of the radiator,
respectively. The maximum predicted temperature was achieved near the water inlet surfaces for all cases.
A temperature field obtained by Vavricka R.and Basta J. is shown in Fgiure-8. They used thermography for
mapping temperature fields of radiator surface. This radiator used in their study was nearly about our steel
panel radiator in view of dimensions and conditions. The temperature fields obtained from their experimental
studies have same trends with our numerical results shown in Fig.5-7. Three different lines were defined to
get comparative results on the surface of the panel radiator (Fig.9.).

From the results shown in Fig. 10-12, all three cases have same trends and higher temperature gradients
were computed between bottom and top of the radiator surfaces. The maximum temperature gradients were
obtained in the zone which has 0.1 m height from the radiator bottom. The temperature values along the
radiator surface were not remarkably altered from the half part of the radiator surface in all three cases.
Another important result is that the temperature difference between inlet and outlet water temperature,
decreases depends on heat transfer coefficient.
T(°C)

Height of the radiator (m)

Figure-10 Temperature values obtained from CFD results for Case-I

T(°C)

Height of the radiator (m)

Figure-11 Temperature values obtained from CFD results for Case-II

1070
 T(°C)

Height of the radiator (m)

Figure-12 Temperature values obtained from CFD results for Case-III

CONCLUSION

In this study, three dimensional steady-state numerical calculations of thermal output of a steel panel
radiator, in accordance with TS EN442, was performed by using CFD method. In the numerical simulations,
original dimensions and geometry of a steel panel radiator was used without any simplifications in
geometrical detail. Three different cases were considered according to standard test conditions of thermal
output of a steel panel radiator. The water inlet temperatures, mass flow rates and ambient temperatures
were set to different values obtained from experimental results in all three cases. The results show that
higher temperature gradients occurred at the radiator surface and the temperature values along the radiator
surface were not remarkably altered from the half part of the radiator surface in all three cases. The lowest
temperature difference between inlet and outlet surfaces was obtained for the lowest inlet temperature.
Another important result was that the convective heat transfer coefficient gets an important role on the
thermal output of steel panel radiators. Numerical results show that CFD is a useful tool for predicting
thermal output of a panel radiator and temperature distributions of panel radiator surfaces. The results
obtained from the numerical simulations were in good agreement with the experimental data.

ACKNOWLEDGEMENT

The authors wish to thank the COSKUNOZ A.Ş. supporting for this research.

REFERENCES

[1] Sevilgen, G. and M. Kilic. 2011. Numerical analysis of air flow, heat transfer, moisture transport and
thermal comfort in a room heated by two-panel radiators, Energy and Buildings, vol. 43, no.1, January 2011,
pp. 137–146.

[2] Murakami, S., S. Kato, J. Zeng. 2000. Combined simulation of airflow, radiation and moisture transport for
heat release from a human body, Building and Environment 35 489-500.

[3] Kilic, M. and G. Sevilgen. 2008. Modelling airflow, heat transfer and moisture transport around a standing
human body by computational fluid dynamics, International Communications in Heat and Mass Transfer,
Vol.35(9), pp.1159–1164.

[4] Myhren, J.A. and S. Holmberg. 2008. Flow patterns and thermal comfort in a room with panel, floor and
wall heating, Energy and Buildings 40, 524–536.

[5] ASHRAE Handbook–Fundamentals, chapter 8, 1997. Atlanta: American Society of Heating, Refrigeration
and Air-Conditioning Engineers.

1071
[6] Hutter E. 1991. Comparison of different heat emitters in respect of thermal comfort and energy
consumption, in: Proceedings of the International Centre for Heat and Mass Transfer, Heat and Mass
Transfer in Building Materials and Structures, pp. 753–769.

[7] Myhren J.A. and S. Holmberg. 2009. Design considerations with ventilation-radiators: Comparisons to
traditional two-panel radiators, Energy and Buildings 41, 92–100.

[8] Vavricka R.and J. Basta. Temperature fields of radiators, Departmen of Environmental Czech Technical
University in Prague, Faculty of Mechanical Engineering Technicka 4 ,166 07, Prague 6 , Czech Republic.

[9] Gritzki R., A. Perschk, M. Rösler, W. Richter and TU Dresden. 2007. Modeling of Heating Systems and
Radiators in Combined Simulations, Proceedings of Clima 2007 WellBeing Indoors, Germany.

[10] Beck S.M.B., S.C. Grinsted, S.G. Blakey, K. Worden. 2004. A novel design for panel radiators, Applied
Thermal Engineering 24, 1291–1300.

[11] Akın D. 2007. Computer Aided Design of Thermal Systems. November, İzmir.

[12] Pillutla G., R. Mishra, S. Barrans. 2008. Determination of convective heat transfer over a standalone
water filled radiator using thermal imaging, School of Computing and Engineering Researchers’ Conference,
University of Huddersfield, November.

[13] Bangert K. 2010. Radiator heat transfer augmentation by changes to wall surface roughness and
emissivity, Mini-project report, University of Sheffield, EPSRC Pioneering research and skills.

[14] Aydar E. 2009.The estimation of the thermal efficiency of panel type radiators by CFD analysis and an
alternative design research to increase efficiency, Yüksek Lisans Tezi, Marmara Üniversitesi, Makine
Mühendisliği Bölümü.

[15] Erdoğmuş A B. 2011.Simulation of the Heater Test room Defined by EN 442 Standard and virtual
Testing of Different Type of Heaters, Graduate School of Engineering and Sciences of İzmir Institute of
Technology, İZMİR.

[16] Akın D. 2007. Computer Aided Design of Thermal Systems, Graduate School of Natural and Applied
Sciences of Dokuz Eylül University, Degree of Master of Science in Mechanical Engineering, Energy
Program.

[17] Myhren J.A. 2011. Potential of Ventilation Radiators: Performance assessment by numerical, analytical
and experimental investigations, Royal Institute of Technology, Department of Civil and Architectural
Engineering, School of Architecture and the Built Environment, Fluid and Climate Technology, Haninge-
Stockholm, Sweden.

[18] Fluent 6. 2001. user’s guide, Lebanon, NH, USA, Fluent Inc.

[19] Fluent 6.2 .2003. UDF Manual, Lebanon, NH, USA, Fluent Inc, 2003.

[20] Churchill, S.W. and H.H.S. Chu. 1975. Correlating equations for laminar and turbulent free convection
from a vertical plate, Int.J. Heat Mass Transfer, 18:1323-1329.

1072
 NUMERICAL SIMULATION OF COMPRESSION STROKE USING SMOOTHED PARTICLE
HYDRODYNAMICS
a a a b
T. El-Gamma1 , E. E. Khalil , H. Haridy , E. Abo-Serie
a
Dep. of Mechanical Power Engineering, Cairo University, Giza, Egypt
b
Mechatronics Department, Mevlana University, Konya, Turkey
tarekelgammal@eng.cu.edu.eg, khalile1@asme.org, hatem_kayed@eng.cu.edu.eg, eaboserie@mevlana.edu.tr

ABSTRACT
In a direct application for the promising numerical technique, Smoothed Particle Hydrodynamics (SPH), a simulation is
conducted for compression stroke of air inside a flat piston-cylinder pattern. Flow velocities, pressures, densities and
temperatures are calculated and represented during the whole stroke at different piston positions inside the cylinder.
The simulation targets to emphasize the competitive nature of SPH against the conventional numerical methods like
Finite Difference (F.D.), Finite Volume (F.V.) and Finite Element (F.E.) in complex deformable geometries.
Investigations about the value of the smoothing length (h) and optimizing the ghost particles position, interaction, and
updating have been undergone. The results show a significant reduction in computational time without scarifying the
accuracy.

Key words: Compression, Mesh-less, Smoothing length, Virtual particles

I. INTRODUCTION
Smoothing Particle Hydrodynamics (SPH) is one of the recently developed mesh-less methods which have acquired
the reliability in the last decade as a numerical solution approach. SPH is a Lagrangian adaptive method that does not
need a mesh for solving the discretized domain. Using this method, the physical domain is divided into unrelated
particles that have physical properties (i.e. volume, mass, velocity, pressure…etc) of the domain. The field function
(Temperature, pressure, velocity….etc.) is calculated for a discrete particle from the field functions of the all
surrounding particles weighted by the smoothing function (W ij). The smoothing (weighting) kernel function takes into
account the percentage effect of the neighboring particles within a predefined radius called smoothing length (h). Many
common kernel functions use a dependent factor (K) to multiply with the smoothing length (h) (Liu et al., 2003).

After being developed for simulating the cosmological phenomenon (Lucy, 1977), SPH has been used in solving fluid
and solid mechanics problems. The adaptive nature of SPH makes it more suitable for simulations involving fluid free
surfaces, wave discontinuity (Sod, 1978), solidification (Monaghan et al., 2005) multiphase flow (Tartakovsky et al.,
2009) and highly deformable geometries which are difficult and require large memory and long computational time due
to the continuous change in the computational mesh in space and time to fit the domain leading to inaccuracy and
sometimes system collapse.

Pressure wave propagation is of great interest to be studied. High attention is directed for those which is developed by
impacts and reflected by solid walls. An important example is the adiabatic compression process for a gas to increase
its pressure for storage (Air compressors) of further processing (Internal combustion engines). Differential (i.e.
Lagrangian) analysis is required in this case with high effect and accurate method.

In this paper, a simulation is done for compression stroke inside a cylinder to track the behavior of air particles and
measure the fluid and flow properties at different points during the stroke till the end. The verification is applied to the
isentropic relations depending on the compression ratio and the initial and boundary conditions. The smoothing length
(h) was varied to minimize the error and the virtual particles was laid and carefully treated to optimize the solution.

II. PROBLEM DEFINITION

The case represents a compression stroke in internal combustion engine using a flat piston inside a cylinder
(D=0.1285 m, Lstroke=1.2D = 0.15842 m).The stroke starts from rest (velocity u=0) at the top dead center (TDC) to stop
and finish at the bottom dead center (BDC). The cylinder other specifications were as follows: compression ratio (rc=
6), rotational speed (N=1000 rpm).
The medium under investigation is air and it’s dealt as perfect gas of constant specific heat under constant volume
(Cv= 717.5 J/kg) and isentropic index (γ=1.4).The initial pressure and temperature are (Pi= 1*105 Pa, Ti= 300 K).

1073
 Finally the cylinder piston arrangement is considered adiabatic but not isentropic. It should be noted that the
simulation was 1-D (The direction of piston motion)
The governing equations describing the change in density (ρ), pressure (p), horizontal velocity component (u), and
temperature (T) are:

Continuity equation:

(1)

Linear momentum conservation:

(2)

Energy conservation:

(3)

Equation of state:

(4)

Speed of sound:

(5)

Note the air is considered inviscid (i.e. no shear term in linear momentum equation). Artificial viscosity is added
later to account for particle over penetration. The piston is governed by equations of motion and adiabatic boundary
conditions (Fazio et al., 2010), such that the properties of air particles attached to the piston are calculated by:

Piston equations of motion:

(6)

(7)

Piston momentum equation

(8)

Adiabatic density-pressure gradients

(9)

Where (up) is the piston velocity in (m/s), (ap) is piston acceleration (m/s2), (cp) is the sound speed of air particles at
the piston conditions (m/s), (ω ) is the angular velocity of the crank shaft (rad/s), (t) is the time from the start till
the required moment (s), and ( ) is the connecting rod length in (m) and it is taken equal to Lstroke.
The cylinder head side is fixed. This applies a zero velocity particle with constant zero gradients for the other
properties (density, pressure, temperature) at all times.
The validation of this case already exists in the isentropic solution for a compression process of ratio (r c) for an air
charge of initial pressure and temperature (Pi and Ti). This is shown n the following relations:

1074
 (10)

(11)

pf and Tf are the final pressure and temperature after the compression stroke.

III. METHODOLOGY
i. Governing equations using SPH
SPH formulation starts from the integral representation of field function using a limited-zone, kernel function
W(xij,h). Then we can write the function approximation in SPH formulation as:

(12)

Where <f(xi)> is the kernel approximation of the scalar field f(x) at particle (i).
For two or three dimensions, the kernel function W (xij,h) can be replaced by W (rij,h) such that (rij) is the vectorial
distance between (i) and (j).
The gradient (∂f(xi)/ ∂x) and Laplacian (∂2f(xi)/ ∂x2) of the field function are evaluated in integral representation
(Rook et al., 2007) to be:

(13)

(14)

The field functions in (12), (13) and (14) are discretized in space, using the particles representing the domain, so
the integrations are approximated by summations of functions weighted by the kernel function (Rook et al., 2007).

(15)
ρ

(16)
ρ

(17)
ρ

SPH formulation mentioned above can be substituted into the governing equations (1-3). Accordingly, the SPH
representation of 1-D continuity, momentum and energy can be written as follows:

(18)

(19)

(20)

Note that ( ) is the artificial viscosity term which will be discussed later.

The compact cubic spline is kernel function has been used in this study (Monaghan et al., 1985). The formulations
of the cubic spline kernel function (W ij) and its spatial derivative can be shown in (21) and (22).

1075
 (21)

(22)

Where (R=xij/h) is the ratio between the magnitude of the distance between two neighboring particles (xij) and the
smoothing length (h). In Eq. (22), the coefficient (a1) is dependent on the smoothing length (h) and has values of (1/h).

IV. NUMERICAL ANALYSIS

i. Domain Particles Distribution
Using Matlab®, the cylinder domain is discretized into (Nx) equally spaced particles which are classified into (Nx-2)
interior particles with2 boundary particles (one at each side).
The initial conditions of (ρ, P, T, u) are set to all the particles. Particle differential length will be calculated from its
spacing in x coordinate (i.e. m/ρ = dx). The cylinder starts from rest (i.e. ui=0) at pressure and temperature mentioned
above with adiabatic boundaries (q"=0).

ii. Virtual (ghost) particles

The virtual or ghost particles (Allahdadi et al., 1993 and Randles et al., 1996) are generated in every time step as
an image for the medium particles near the boundary (i.e. lie within the kernel domain of the wall/boundary particle) as
in Fig. 1.
The virtual particle gains its physical properties from its real particle and the boundary condition of the wall. The
existence of artificially particles plays an important role in completing the kernels to give accurate physical properties
for all particles. Also they facilitate accurately representing the boundary conditions as in F.D. especially for Neumann
boundary condition.
The virtual particles were applied to the two boundaries, the moving piston and the cylinder head, and they are
increasing every time step due to compactness of particles.

iii. Boundary forces

If any fluid particle approaches the special limited zone of the boundary, it is exerted by a repulsive force from the wall
particle (along the line of centers) to prevent the penetration. This force is calculated in the same way as the molecular
force of Lennard-Jones form. The repulsive force on particle (i) is generated according to this relation (Liu et al., 2003):

(23)

Where is the repulsive force on particle (i) in the (k)th dimension. (r0) is the cut-off distance which determines the
effective zone of the wall repulsive force. (r0) is usually taken a fraction from the initial spacing. (D) is a problem
dependant factor.
iv. Artificial viscosity
The artificial viscosity approach (Monaghan, 1994) is proposed to prevent particles inter-penetration by adding
viscous terms in the momentum and energy equations to overcome the drawbacks arise from SPH kernel in solving
the pressure waves along the particles. The proposed artificial terms are:

ρ (24)

Where in most reviews (Liu et al., 2003) but in this case they were taken equal to 0.1 and zero
respectively to minimize their detrimental effect of disturbing the particles order, while other parameters are calculated
from:

1076
 (25)

(26)

ρ ρ
ρ (27)

(28)

(29)

v. Time Marching
The property (i.e. density, temperature) at the current time step is calculated using the value of the previous time
step and its rate of change.

At t = dt (i.e. n=1)

(30)

Where are evaluated by initial values T0 from the governing energy equation. For n > 1:

(31)

Where are evaluated by Tn-1/2from the energy equation.

The time step has been chosen to satisfy the Courant-Levy condition for mechanisms involving thermo-fluids
(Monaghan, 1992 and Liu et al., 2003).

(32)

In this case, time step was considered as (dt = 1 x 10-5 s) which is very convenient to the above mentioned
condition that deals with small time steps.

V. RESULTS
i. Optimum smoothing length
Due to the variety of properties in this case, the optimized smoothing length (hopt) was based on reducing the
maximum percentage error found at the stroke end and it was always dedicated to the pressure values. Note that the
percentage error is calculated from this equation:

–
% error = (33)

Fig. 2 shows the permanent increase of hopt value with the increase in number of discretizing particles (Nx). This
trend is best fitted with proper polynomial equation of 3rd order with negligible coefficient (nearly it is 2nd order). Other
lower and higher polynomial orders aren’t better than this fit.

1077
Figure 1: variation of optimized smoothing length factor

Other important variable to be taken into consideration was the calculation time to achieve the acceptable
accuracy. This is demonstrated in Figs. 3 and 4 where the maximum percentage error and the numerical time of
calculation are plotted against (Nx). It’s worth noting that the plotting values were considered in case of (hopt) of each
(Nx).

Percentage error for different particles numbers

Absolute maximum percentage error (%)

0,08
0,07
0,06
0,05
0,04
0,03
0,02
0,01
0
41 51 61 71

Number of Particles Nx

Figure 2: The effect of increasing of particles number to the computational percentage error

1078
 calculation time for different particles numbers

4,5
computational time (sec)

4
3,5
3
2,5
2
1,5
1
0,5
0
41 51 61 71
Number of Particles Nx

Figure 3: The effect of increasing of particles number to the computational time consumed

From Figs. 3 and 4, the chosen number of particles was (Nx=41) with optimized factor of smoothing length
(hopt=1.5). This is because of the acceptable limits of error, time (0.015% and 1.37 sec respectively) and well
organized final distribution of particles.
After the choice of (Nx) and (hopt), the results produced by the code are presented for any time during the stroke to
show the development of properties under effect of pressure waves and the motion of particles with their internal
velocities. The plots are illustrated in lines of property variation with different air particles according to the position
inside the cylinder.

ii. Transient period

The following plots (Fig. 4,5,6) are showing the percentage difference of particles properties (pressure, density and
temperature) to the isentropic values at three times of compression (t = 0.01 s, 0.02 s and 0.03 s) . It’s obvious that the
properties near piston head at the first period (t =0.01s) are having a lower value than the isentropic. This means that
the pressure wave was coming from the other part of the cylinder raising its properties. Meanwhile, at t =0.02 and 0.03
s, the cylinder stroke highly decreased, so the wave is reflected between both sides in very short distance and time.
Consequently, the reflected pressure wave is added with the compression applied by piston causing an exceptional
rise near the piston head.
In Fig. 7, the particles velocities are compared to the mean piston speed (mps) at the three different times to show
the variation in place and time. The highest gradient is achieved at t = 0.02 s as the piston reaches its maximum at t=
0.015s in half the stroke then decrease to stop at t= 0.03s at the end of stoke.

1079
Figure 4: Pressure difference from isentropic values in (%)

Figure 5: Density difference from isentropic values in (%)

1080
Figure 6: Temperature difference from isentropic values in (%)

Figure 7: particles velocities in percentage of mean piston speed

1081
VI. CONCLUSION
From the simulation, the property field evolves during compression stroke depending on the pressure waves sent
by the moving piston. These waves are reflected at the boundaries with continuous shrinking in stroke length causing
a high rate of properties development.
Optimizing the smoothing length in SPH achieves the merits of high accuracy and less computation time. Any
increase or decrease in the smoothing length will change the kernel weighting function (W ij) in (shape and value) and
disorients it from the actual property profile.
The optimized (h) is not a constant value at every circumstance. It varies with the discretizing particles number (Nx)
and other parameters like initial gradients. In this case, (h opt) changes in a weak 3rd order polynomial relation (almost
2nd order polynomial relation) with (Nx).
The proper choice of the number of virtual particles and the boundary treatment are very important in achieving the
accuracy at the boundary and avoiding system collapse. Meanwhile the addition of artificial viscosity terms helped in
preserving the motion from improper behavior of some particles due to inconvenient approximation of kernel function.
SPH ability of calculating the properties at any differential particle independent on predefined connectivity or
distribution with smooth results makes it superior in applications of material deformation and the generated pressure
waves with its reflections. This was well obvious in this case study of compression stroke for air charge.

VII. NOMENCLATURE
II Shear viscosity coefficient
II Bulk viscosity coefficient
ρ Density, kg/ m3
φ Non-singularity coefficient,
Artificial viscosity term
a Piston acceleration, m/s2
a1 Dimension coefficient of smoothing kernel function
C Specific heat, kJ/ kg. K
D Engine Diameter, m
L Engine stroke length, m
P Pressure, Pa
q Heat rate, W
RF Repulsive force per particle mass (m/s2)
T Temperature, K
u Velocity, m/s
X Position
SUBSCRIPT
i Initial, interested particle
f final
j Neighbor particle
opt. optimum
p piston
SUPERSCRIPT
" flux

VIII. REFERENCES
[1] Liu, G. R. and M. B. Liu. 2003. Smoothed particle hydrodynamics: a meshfree particle method. Singapore: World
Scientific Publishing Co. ch. 1. pp. 18.
[2] Lucy L. B. 1977. A numerical approach to the testing of the fission hypothesis, Astro. J., vol. 82, pp. 1013–1024.
[3] Sod G. A. 1978. A survey of several finite difference methods for systems of nonlinear hyperbolic conservation
laws. Journal of computational physics. vol. 27. pp. 1–31.
[4] Monaghan J. J., H. E. Hupper, and M. G. Worster. 2005. Solidification using smoothed particle hydrodynamics.
Journal of Computational Physics. vol. 206. pp. 684–70.
[5] Tartakovsky A. M., K. F. Ferris, and P. Meakin. 2009. Lagrangian particle model for multiphase flows. Computer
Physics Communications.vol. 180. pp. 1874–1881.
[6] Fazio R. and G. Russo 2010. Central Schemes and Second Order Boundary Conditions for 1D Interface and
Piston Problems in Lagrangian Coordinates. Commun. Comput. Phys. vol. 8: 797-822.

1082
[7] Rook R., M.Yildiz, and S. Dost 2007. Modeling transient heat transfer using sph and implicit time integration.
Numerical Heat Transfer, Part B: Fundamentals: An International Journal of Computation and Methodology. vol.
51, no.1. pp. 1-23.
[8] Monaghan J. J. and J. C. Lattanzio 1985. A refined particle method for astrophysical problems. Astronomy and
Astrophysics Journal, vol. 149. pp. 135-143.
[9] Allahdadi F. A., T. C. Carney, J. R. Hipp, L. D. Libersky, A. G. Petschek. 1993. High strain Lagrangian
hydrodynamics: a three-dimensional SPH code for dynamic material response. Journal of Computational Physics.
vol. 109. pp. 67-75.
[10] Randles P. W. and L. D. Libersky. 1996. Smoothed particle hydrodynamics: some recent improvements and
applications. Computer Methods in Applied Mechanics and Engineering. vol. 138. pp. 375-408.
[11] Monaghan J. J. 1994.Simulating free surface flow with SPH. Journal of Computational Physics, vol. 110: 399-406.
[12] Monaghan J. J. 1992. Smoothed Particle Hydrodynamics. Annu. Rev.” Astron. Astrophys. vol. 30: 543-574.

1083
 HYDRODYNAMIC ASPECTS OF OSCILLATING FLOW THROUGH POROUS MEDIA
CONSISTING OF STEEL SPHERES

M. T. Pamuk and M. Ozdemir

Piri Reis University, Department of Naval Engineering, Istasyon Mah. Tuzla, Istanbul, Turkey
Istanbul Technical University, Mechanical Engineering Faculty, Beyoglu Istanbul,Turkey
email: turgaypamuk@hotmail.com, ozdemirmu4@itu.edu.tr

ABSTRACT
Hydrodynamic aspects of oscillating flow through porous media are investigated within the scope of this
experimental work where water is used as the flow medium. Friction factors are found reducing the data of
hydrodynamic experiments for oscillating flows. The experimentally determined maximum friction factors for
oscillating flow are represented as correlations. They are then compared to those obtained by the other
researchers. The results can readily be used in the design of cryogenic coolers and heat exchangers where
pressure drop thus hydraulic losses are of primary importance.

Keywords: Porous medium, oscillating flow, friction factor, permeability, inertial coefficient

INTRODUCTION
The porous media have been used widely in many engineering fields such as cryogenic coolers, solid matrix heat
exchangers, cooling of electronic equipment and regenerators in order to enhance heat transfer. However, due to
the complex structures of porous media, pressure loss thus fluid pumping power has always been very important.
Henri Darcy, a French hydrologist who designed a water filter for the city’s drinking water system established the
first theory on the hydrodynamic aspects of fluid flow through porous media about 150 years ago. Another historic
work was accomplished by Ergun (1952) who calculated experimentally the coefficients of the constitutive
equations required by the continuum modeling such as permeability and inertial coefficients of porous media.
These experimental coefficients can be found widely in the literature according to the types of porous media.

Özdemir (1996), investigated the hydrodynamic aspects of porous media consisting of wire screen meshes under
steady flow conditions. His work suggests a new method to calculate the porosity using REV. He used Ergun’s
equations to define the permeability and other parameters of the media used in his experiments.

Dukhan et. al (2006) conducted steady-state unidirectional pressure-drop measurements in wind-tunnel for airflow
through nine compressed and uncompressed isotropic open-cell aluminum foam samples, having different
porosities and pore densities. They found that the compressed foam produced significantly higher pressure drop,
which increased with increasing Darcian velocity following the quadratic Forchheimer equation. Both compressed
foam and uncompressed foam were correlated using an Ergun-like equation with the correlation.

Riberio et al. (2010) tried to establish equations to represent wall effects in densely packed porous media
subjected to steady flow. According to their findings, average porosity increases as D/dp decreases. This feature
should be taken into consideration especially when particle sizes are big when compared to that of the channel
through which the flow takes place.

Aside from these developments in steady fluid flow through porous media, the oscillating fluid flow also has found
many engineering applications such as internal combustion engines, Stirling engines, cryogenic coolers and other
periodical processes in thermal and chemical systems.

Zhao and Cheng (1996) investigated experimentally oscillatory pressure drops through a woven-screen packed
column. They presented correlations for maximum pressure drop factor and cycle-averaged pressure drop factor
in the kinetic Reynolds number range of 0.001 to 0.13 and in dimensionless fluid displacement range of 614.73 to
2827.56, under the condition of sinusoidal motion of air. They found that the values of cycle-averaged pressure
drop of oscillatory flow were several times higher than that of steady flow.

Ju et al. (1998) conducted experiments to determine the oscillating flow characteristics for a regenerator consisting
of stacks of wire meshes in a pulse tube cryogenic cooler. The oscillating flow characteristics appear not only as
pressure drops but also as phase lags, as gas is used as the flow medium throughout the experiments. It is found
that the value of the cycle-averaged pressure drop of the oscillating flow in the regenerator is two to three times
higher than that of a steady flow at the same Reynolds numbers based on the cross-sectional mean velocity.

1084
Hsu et al. (1999) studied experimentally pressure-velocity correlations of steady and oscillating flows in
regenerators made of wire screens. They indicate that, for oscillating flows, the velocity responses quite linearly to
the pressure gradient when the piston amplitude is small. This suggests that Darcy's law is valid for small
amplitude oscillating flows. When the piston amplitude becomes large, the response and therefore the correlation
of pressure-drop and velocity in the regenerator become nonlinear.

Nam and Jeong (2004) analyzed a regenerator under oscillating flow. In their model, they used an additional
parameter named as Breathing factor for the precise estimation of the phase angle. They obtained empirical
correlations of the friction factor and the Breathing factor for the screen regenerators.

Ju and Shen (2006) studied comparatively the oscillating flow characteristics of cryogenic cooler regenerators at
low temperatures. They proposed universal friction factors for various temperature conditions to be used in
regenerator designs.

Cha (2007) studied experimentally and numerically hydrodynamic parameters of micro porous media for steady
and oscillatory flow on oscillating flow in eight different micro porous media. His findings show that pressure drop
in oscillatory flow is not necessarily lower than that of steady flow which actually depends on packing material and
the oscillating frequency.

Shen and Ju (2008) summarized typical experimental results and correlations on the friction factor of regenerators,
at different operating frequencies, at room and cryogenic temperatures. The comparison of these friction factor
data were presented to clarify the reason for their difference. Finally, a new correlation of friction factor for
oscillating flow regenerator, in terms of two non-dimensional parameters, was presented.

Jin and Leong (2008), and Leong and Jin (2006) have conducted an experimental study regarding steady and
oscillating flows through open cell aluminum foams. Considering various porosities and permeabilities, they
conclude that flow resistance increases with form coefficient and decreases with the increasing permeability for a
given porosity. Form drag is the primary reason for pressure loss by increasing flow velocity. They presented
correlations of friction factors as Zhao and Cheng. They also showed that the pressure loss is increased both with
increasing Ao and kinetic Reynolds number Re

Cheadle et al. (2011) studied packed sphere regenerator friction factors under oscillating flow conditions using
CFD Analysis. They used Valensi number which is defined as kinetic Reynolds number in other publications, to
take into account the oscillation frequency. They point out that friction factor is independent of Valensi number, for
Va < 10,and is therefore not effected by oscillating flow in this range. For Va > 10, however, oscillating flow effects
are evident.

Pamuk and Özdemir (2012) studied experimentally oscillatory and steady flows of water through two different
porous media consisting of mono-sized stainless steel balls. They provided expressions for friction factors for both
steady and oscillating flows.

In this paper, findings of Pamuk and Özdemir (2012) have been compared to those of Zhao and Cheng (1996)
and Leong and Jin (2006) who used wire screen meshes and aluminum foam, respectively. It has thus been
possible to assess these three most common packing materials in terms of pressure drop under oscillating flow
conditions.

EXPERIMENTAL SETUP
Water is used in this study as the working fluid unlike the majority of the experimental studies in literature. The air
dissolved or encapsulated by water is a major problem when it comes to pressure and velocity measurement. The
compressibility of water is considerably effected due to air content, thus the flow rates of half cycles are different.
In order to overcome this problem the whole system was vacuumed then filled with water.

A schematic diagram of the experimental setup is shown in Fig. 1. Main components of the experimental setup are
an oscillation generator (a double-acting cylinder), an electrical motor-reducer to drive it, and a flywheel with an
adjustable crank-arm. The remaining components are related to piping.

Data collected by means of Keithley 2700 data acquisition system from pressure sensors installed at the inlets of
the porous medium are sent to a PC to be processed. An inductive type proximity sensor is installed to sense a
metallic target mounted on the periphery of the flywheel such that the location of the target corresponds to the
maximum piston movement (full stroke). Since pressure and location signals are both collected by the data
acquisition system, it is possible to synchronize them due to the fact they are in the same phase, as opposed to

1085
when gas is used as the working fluid where significant phase lags may occur depending on the oscillation
frequency.

The test chamber is made from a stainless steel pipe. Porous media are formed by filling the test chamber with
mono sized stainless steel balls of 1 and 3 mm in diameter. The test chamber is shaken to ensure that no
excessive voids are left. Wire meshes are installed at both ends of the test chamber in order to prevent the balls
from scattering.

Fig. 1. Experimental setup: 1.Test Section (Porous Medium), 2.PE Pipe, 3.Pipe of 32-mm in diameter, 4.Keithley 2700, 5.PC 6. Oscillation
Generator, 7. Motor reducer, 8. Flywheel, 9. Crank Arm, 10. Inductive Proximity Sensor, 11. Air purging device. 12. Separating valves for steady
and oscillating flows.

An uncertainty analysis based on the method described by Figliola and Beasley (2006) is performed. The
uncertainties of angular frequency, maximum fluid displacement xmax, dimensions, the amplitude of area averaged
fluid velocity and pressure gradient ( P / L ) are estimated to be 0.43%, 0.51%, 1.95%, 0.23% and 1.76%
respectively.

EXPERIMENTAL RESULTS AND DISCUSSION

Oscillating flow experiments are conducted for 3 different flow amplitudes and 8 different flow frequencies.
Maximum fluid displacements are chosen according to the displacement of the piston as 130 mm, 170 mm and
195 mm. In this case, the maximum flow displacements (xmax) at the entrance of the porous channel are
calculated from the ratio of cross-sectional areas of piston and the entrance of porous channel as 72.63 mm, 94.97
mm and 108.94 mm. For each amplitude, line frequency of 50 Hz is changed from 5 Hz to 15 Hz by an A/C motor
drive. As a result of this, flow frequency changes in the range of 0.115-0.345 Hz depending on the 6.9-20.7 rpm
speed of motor reducer which normally operates at 69.7 rpm at 50 Hz of line frequency.

The displacement of the piston is taken as zero at the rear position inside the cylinder and it becomes maximum
value which is equal to the diameter of the flywheel at the forward position. The piston displacement will be equal
to the fluid displacement due to the fact that the fluid is incompressible. Hence, at the entrance of the porous
medium, the fluid displacement xm varies with angular frequency () and time (t) according to
x max
x m (t )  (1  Cost ) (1)
2
where x max  2 RA p / A . Here, R, Ap and A are flywheel radius, cross sectional areas of double acting cylinder
and test chamber, respectively. The cross-sectional mean fluid velocity in the channel is

1086
u m (t )  u max Sint (2)
where u max  x max / 2

Fig. 2. Temporal variation of pressure gradient for various oscillation frequencies

(xmax=94.97 mm, first medium)

In Fig. 2 it is seen that the slope of the pressure gradient is subjected to an abrupt change near the oscillation axis.
The rate of pressure gradient is low within this region where flow velocity is very low, but it is considerably higher
for the other region. This abrupt change may mean that the flow regime inside the porous medium changes. Four
distinct flow regimes for steady flow through porous medium are determined with Reynolds number defined as
Re p   (u m /  ) d /  . These regimes are Darcy flow regime (Rep<1), inertial flow regime (1<Rep<150), unsteady
laminar flow regime (150<Rep<300) and unsteady-and chaotic flow regime (Rep>300) (Kaviany 1995, page 48-49).
The instant average velocity of oscillating flow changes 0 to 0.12 m/s, so that Reynolds number increases up to
approximately 300.

The temporal variation of pressure gradient with respect to cycle angle and the maximum flow displacement is
given in Fig.3 for a fixed frequency (=0.233 Hz). As seen in the figure, amplitude of the pressure gradient
increases as the amplitude of the flow displacement increases. The abrupt change near the oscillation axis can be
observed here also.

Fig. 3. Variation of pressure gradient respect to various flow displacements for fixed frequency
(=0.233 Hz, first medium)

1087
Because the mean velocity and pressure gradient vary with time in oscillating flows, friction factor defined above
also varies with time. For this reason, friction factor is defined in the literature with amplitudes of mean velocity
and pressure gradient as shown in the following.
2 DPmax
f max  (3)
2
u max L
Variation of friction factor obtained from the experimental data with respect to Remax for the first porous medium is
shown in Fig. 4. The correlation equation in the form of Eq. 4 is obtained by curve fitting of the experimental data.
3083998
f max   1882 (4)
Remax
It is seen that this kind of correlation equation is in reasonable agreement with experimental data. Eq. 5 expresses
the behavior of the maximum friction factor in oscillating flow through porous medium, in a good manner.

5000

experiment
4000
correlation

3000
fmax

2000

first medium
1000
fmax =3083998/Remax +1882

0
0 1000 2000 3000 4000 5000 6000 7000
Remax

Fig. 4. Maximum friction factor versus Remax (first medium).

Fig. 5 shows the variation of maximum friction factor with respect to Remax for the second porous medium. And the
correlation equation obtained by curve fitting of the experimental data is
532936
f max   612.1 (5)
Remax

1200

experiment
1000
correlation
800
fmax

600

400
second medium
200
fmax =532936/Remax +612.1

0
0 1000 2000 3000 4000 5000 6000 7000
Remax

Fig. 5. Maximum friction factor versus Remax (second medium).

1088
As seen in Fig. 4 and Fig. 5, the maximum friction factor for the first medium with the porosity of 0.369 is about
four times higher than that of the second medium with the porosity of 0.3912. Increasing porosity reduces the
maximum friction factor considerably as expected. These correlation equations of maximum friction factors for
oscillating flows through porous media have a maximum relative error of 9% in comparison with experimental data.

The above correlations have then been compared to those of Zhao and Cheng (1996) and Leong and Jin (2006)
who used wire screen meshes and aluminum foam, respectively. Fig. 6 shows all 6 correlations together. Their
correlations normally given in terms of pore hydraulic diameter are here re-calculated in terms of channel-diameter
based Reynolds numbers in order to compare them. For the selected set of porous media, it is obvious that the
aluminum foams that have a porosity of 0.90 have the lowest friction factor. However friction factors of both wire
screen meshes and steel balls have the same order of magnitudes.

Fig. 6. Comparison of maximum friction factors for various packing materials.

CONCLUSION
Friction factor representing the pressure drop through porous media subjected to oscillating flow decreases
asymptotically independent of the porous media. However, the porosity differences and variation of permeability
due to complex geometries within the porous media has key importance in terms of the magnitude of the friction
factors, thus the pressure loss through these media. It is therefore important to decide the type of the medium
when it comes to comparing the cost of constructing the porous medium to energy cost due to pressure loss.
However, heat transfer features of these different types of media have not been mentioned here and assumed to
be comparable to each other.

NOMENCLATURE
A cross-sectional area of the test chamber
Ap cross-sectional area of double acting cylinder
Ao non-dimensional displacement defined as Ao = xmax/D
d ball diameter
D inner diameter of test chamber
Dh hydraulic diameter, same as inner diameter of test chamber
f temporal friction factor defined in Eq. (8)
fmax maximum friction factor defined in Eq. (9)
F inertial coefficient

1089
K permeability
L length of the porous medium
ΔP pressure difference
ΔPmax amplitude of pressure difference
P non-dimensional pressure parameter defined in Eq. (1)
R radius of flywheel
ReD Reynolds number defined in Eq. (1)
Remax Reynolds Number defined as Remax  Ao Re / 2
2
Re ω Kinetic Reynolds Number ( D /  )
t time
um cross-sectional mean fluid velocity
umax amplitude of mean fluid velocity
xm temporal fluid displacement at the inlet of the test chamber
xmax maximum fluid displacement at the inlet of the test chamber
2
Va Valensi Number ( D /  )
Greek letters
 porosity
 fluid density
 angular frequency, rad/s
 frequency, 1/s
 dynamic viscosity

REFERENCES
Ergun, S. , 1952, Fluid Flow Through Packed Columns, Chemical Engineering Progress, Vol. 48, No.2, pp. 89-94

Özdemir, M., 1996, Tel Örgü Katmanlarından Oluşan Gözenekli Ortamda Zorlanmış Isı Geçişi, Ph.D. Thesis,
Istanbul Technical University, May 1996

Dukhan N., Feliciano, R., Hernandez, A., 2006, Air Flow Through Compressed and Uncompressed Aluminum
Foam: Measurements and Correlations, Transactions of the ASME, 1004 / Vol. 128, SEPTEMBER 2006

Riberio, A.B., Neto, P., Pinho, C., 2010, Mean Porosity and Pressure Drop Measurements in Packed Beds of
Mono sized Spheres: Side Wall Effects, International Review of Chemical Engineering, Vol.2, N.1, January 2011

Zhao, T.S. and Cheng, P., 1996, Oscillatory Pressure Drops through a Woven Screen Packed Column Subjected
to Cyclic Flow, Cryogenics 36 (1996) 333-341

Ju, Y., Jiang, Y., Zhou, Y., 1998, Experimental study of the oscillating flow characteristics for a regenerator in a
pulse tube cryogenic cooler, Cryogenics 1998 Volume 38, Number 6

Hsu, C., Fu, H., Cheng, P., 1999, On Pressure-Velocity Correlation of Steady and Oscillating Flows in
Regenerators Made of Wire Screens, Transactions of the ASME, Vol. 121, MARCH 1999

Ju, Y. and Shen, Q., 2009, Comparative study of oscillating flow characteristics of cryogenic cooler regenerator at
low temperatures, Front. Energy Power Eng. China 2009, 3(1): 80–84

Cha, J.J., 2007, Hydrodynamic Parameters Of Micro Porous Media For Steady And Oscillatory Flow: Application
To Cryogenic cooler Regenerators, A Dissertation Presented to the Academic Faculty, Georgia Institute of
Technology, August 2007

Shen, Q.Q. and Ju, Y.L., 2008, A New Correlation Of Friction Factor For Oscillating Flow Regenerator Operating
At High Frequencies, CP 985, Advances in Cryogenic Engineering, Transactions of the Cryogenic Engineering
Conference-CEC, Vol. 53

Jin L.W. and Leong K.C., (2008), Pressure drop and friction factor of steady and oscillating flows in open-cell
porous media, Transport in Porous Media (2008) 72:37–52 DOI 10.1007/s11242-007-9134-3, Published online:
15 May 2007 © Springer

1090
Leong, K.C. and Jin, L.W., 2006, Effect of oscillatory frequency on heat transfer in metal foam heat sinks of
various pore densities, International Journal of Heat and Mass Transfer 49 (2006) 671–681

Cheadle, M.J., Nellis, G.F., Klein, S.A., 2011, Regenerator Friction Factor and Nusselt Number Information
Derived from CFD Analysis, Cryogenic coolers 16, Edited by S.D. Miller and R.G. Ross Jr., International Cryogenic
cooler Conference, Inc, Boulder, CO, 2011

Pamuk, M.T. and Özdemir, M., 2012, Friction factor, permeability and inertial coefficient of oscillating flow through
porous media of packed balls, Experimental Thermal and Fluid Science 38 (2012) 134–139

Figliola R.S. and Beasley D.E. (2006), Theory and Design for Mechanical Measurements, John Wiley

1091
 AN EXPERIMENTAL STUDY ON ENHANCEMENT OF HEAT TRANSFER IN A SOLAR AIR
HEATER COLLECTOR BY USING POROUS MEDIUM
a,b c d
G. Maraba , M. Mobedi , B. Ozerdem

a
Energy Engineering Program, Izmir Institute of Technology, Izmir, Turkey
b
R&D Center, Indesit Company A.Ş., Manisa, Turkey
c
Department of Mechanical Engineering, Izmir Institute of Technology, Izmir, Turkey
d
Department of Energy Systems Engineering, Bahcesehir University, Istanbul,Turkey
gizem.maraba@indesit.com, moghtadamobedi@iyte.edu.tr, mehmetbaris.ozerdem@bahcesehir.edu.tr

ABSTRACT
The aim of this study is to enhance the heat transfer of a solar air heater by using porous material.
Depending on literature survey, the design parameters were determined and, a solar air heater experimental
set-up was constructed. In order to obtain the optimum design parameters and maximum efficiency of solar
air heater, many experiments have been carried out including clear and porous material assisted channels.
For this purpose, different configurations and designs were investigated. Temperature data along the
collector, average velocity of air inside the channel, pressure drop due to the porous material, fan and
electrical heater power consumptions were measured. The effect of porous material on the heat transfer and
collector efficiency was discussed based on these measurements. After all results examined, the use of
porous material inside a solar air heater had a positive effect on the system, at the expense of the pressure
drop increase.

INTRODUCTION
Solar air heaters are simple devices that utilize incident solar radiation to obtain clean energy for a wide
usage. The solar air heater device intercepts solar radiation, converts this radiation to the heat in air and
delivers the air for use. The main components of a solar air heater are an absorber plate, one or more
channels for the flowing air, insulation for the bottom and lateral sides of the solar collector and one or more
transparent covers. The use of a blower is optional for the air supply.

Solar air heaters may be divided into two main categories. The first category is related to the air channel flow
configuration. The various configurations may be expressed in four sub-titles under this category. The sub-
categories are; single flow single pass (Fig.1), double flow single pass (Fig.2), single flow double pass
(Fig.3), and single flow recycled double pass (Fig.4). The second category is related to the air channel
design. The air channel design affects the system efficiency significantly. For that reason different design
configurations can be used in solar collectors. The second category can also be expressed in three sub-
categories such as; flat plate (Fig.5), extended surface assisted (Fig.6), porous media assisted (Fig.7). Other
studies exist that cannot be explained under any of these categories. These studies are categorized as
special designs.

Studies considering above mentioned categories have been conducted by many researchers. Heat transfer
and fluid characteristics of a solar air heater whose duct was packed with wire mesh screen matrices
investigated by Varshney et al.(1998). Hachemi,(1999) compared the thermal performances of solar air
heater collectors with selective and nonselective absorber plates. Midilli et al., (1999) investigated the drying
periods of mushrooms and pollens of honey by using an experimental drying set-up connected to a solar air
collector. Yeh et al., (2002) investigated collector efficiency of a double flow solar air heater with fins
attached over and under the absorbing plate. Moummi et al., (2004) analyzed heat transfer characteristics of
a solar air collector with rows of rectangular plate fins inserted perpendicular to the flow, theoretically and
experimentally. Ozturk et al., (2004) investigated the thermal performance of a solar air heater with a packed
flow channel, experimentally. Sahu et al., (2005) investigated the heat transfer coefficient of a solar air
heater with 90° broken transverse ribs on the absorber plate, experimentally. Potdukhe et al., (2008)
developed a solar dryer including a drying chamber and a solar air heater with thermal storage. Prasad et
al., (2009) presented the details of heat transfer and friction characteristics of a packed bed solar air heater
as a function of bed geometry and operating parameters. Karmare et al.,(2009) investigated the effect of
roughness of rib grits on the thermohydraulic performance of solar air heaters, experimentally. Kumar et al.,
(2009) investigated the heat transfer and friction characteristics of a solar air heater which was artificially
roughened with W-shaped ribs, experimentally. Moheseni-Languri et al. (2009) investigated the energy and

1092
exergy analysis of a solar air heater, in order to find the optimum mass flow rate. Hans et al., (2010)
investigated the effect of multiple V-ribs on the heat transfer coefficient and friction factor in an artificially
roughened solar air heater duct. Pakdaman et al., (2011) evaluated the performance of a natural convection
solar air heater with a rectangular finned absorber plate.

Fig. 1. A schematic view of single flow single pass. Fig.2. A schematic view of double flow single pass.

Fig.3. A schematic view of a single flow double pass. Fig.4. A schematic view of a single flow recycled double
pass

Fig.5. A schematic view of a flat plate solar air heater

Fig.6. A schematic view of a fin assisted absorber plate. Fig.7. A schematic view of a porous assisted solar air heater.

1093
.
In this study, the enhancement of heat transfer in a solar air heater by using porous medium is investigated.
First, the measurements are conducted while the flow channel of the experimental set-up is kept clear.
Then, a porous material is placed and the effect of this material on the collector efficiency and heat transfer
is investigated.

EXPERIMENTAL SET-UP
A solar heater was constructed from galvanized steel as an experimental set-up. Dimensions of the solar
collector were 1500 x 500 x 50 mm . Since it was a laboratory simulation model, underside of the collector
was considered as an absorber plate. The electrical heater fixed to the bottom of the absorber plate of the
collector was supplied a maximum power of 15 kW. The system was insulated from all sides to minimize
heat losses to surroundings. Air was supplied to the system with help of a controlled radial fan. Solar
radiation was simulated by a jacket type electrical heater placed under the solar air collector. The
temperature measurements were conducted by using k-type thermocouples. The data taken by the
thermocouples were transferred to a computer with help of a data logger. The flow straightening duct was
used in order to establish a fully developed flow. In addition, a preheater was located after the radial fan, as
well. Therefore, air temperature was maintained at a fixed value before it entered the collector. Fig.8 shows
photographic views of the experimental set-up.

Fig.8. Photograph of the experimental set-up.

A wire mesh like porous material with 50 x 50 mm square dimensions was placed in the collector after the
reference experiments were completed with a clear solar air collector channel to enable comparison of
results with and without porous material. The collector was tested with different porosities to reach the
optimum heat transfer and collector efficiency. The porous material is illustrated in Fig. 9 below.

Fig.9. The employed porous material having high porosity

1094
In this study, an ELİMKO E-380 model data logger was used. The data logger was connected to a computer
and a software program was installed to store the temperature values. For each experiment the data were
taken at 5 minutes time intervals. The velocity measurement was performed using a device named TSI 8386-
M Velocity Calc Plus meter. For measurement accuracy, 5 velocity measurement holes were opened on the
top side of the flow straightening duct of the solar air heater system. A control panel of the solar air heater
system was constructed to manipulate flow rate of the fan, power of the preheater and panel electrical heater
which simulated solar radiation. There was an on-off switch on the control panel box that provided the main
electricity supply to the system and the control panel. Location of thermocouples used to measure the air
temperature is shown in Fig.10 below.

Fig.10. Location of thermocouples.

MATERIAL AND METHOD

First, the reference experiments of clear channel were performed. Then, layer assisted experiments starting
from 1 layer to 7 layers were conducted. The calculations used during these experiments are shown below.

Heat Transfer Calculations

The heat gained by the hot air is calculated as shown in Eq.(1).
Qgain m
 C p (Tmean out  Tmean in ) (1)

In order to calculate the heat losses, the temperatures of each side of the collector were measured and heat
loss for each side was calculated individually. For lateral side calculations, Eqs. (2) and (3) are used
according to laminar and turbulent cases, respectively.

0.67 Ra1L/ 4
Nu  0.68  (2)
1  (0.492/ Pr) 9 / 16 4 / 9

2

 0.387 Ra1L/ 6 

Nu  0.825  8 / 27 
(3)

 
1  (0.492/ Pr)9 /16  


For top side calculations Eqs. (4) and (5) are used:

4 1/4 7
Nu  0.54 Ra1/ 4 (10 <Ra <10 ) (4)

7 1/4 11
Nu  0.15 Ra1/ 3 (10 <Ra <10 ) (5)

1095
For bottom surface calculation Eq.6 is used:

5 1/4 10
Nu  0.27 Ra1/ 4 (10 <Ra <10 ) (6)

For all of these calculations Rayleigh number is calculated as follows:

g (Ts  T ) L3
Ra  (7)

By using Nusselt numbers given above, heat transfer coefficients for each surface were calculated.

Air Flow Rate Calculations

During the experiments the velocity values were taken from the holes that were opened above the flow
straightening duct. Five holes were opened at the inlet section while as the rest were opened at the exit
section of the system. An average velocity value is calculated for each experiment according to Eq.(9),
below:

 (Vi A)
Vmean  (8)
Across section
The mass flow rate value for each experiment is calculated using Eq.(10), below:
   Across section Vmean
m (9)

Mean Temperature Calculations

The inlet and exit temperature data were collected from the holes which the velocity data were also taken. A
total number of 30 temperature values were collected from each five holes for inlet and exit temperatures. By
using Eq.(11) given below, mean temperature values are calculated for inlet and exit sides.

 Vmean TdA
Tmean  (10)
m

RESULTS AND DISCUSSION

Change of Velocity Profile

The velocity data were obtained for each 8 experiments (clear and starting from 1- layer to 7- layer assisted).
The mean velocity and mass flow rate for each 8 experiments are 0.189 m/s and 0.0055 kg, respectively.
Based on the collected data, the velocity profiles for clear, 4-layer porous assisted (semi-filled), and 7- layer
porous assisted (fulfilled) channels are plotted and shown in Figures from 11 to 13. Therefore, the effect of
porous layers on the velocity distribution in the channel is observed. As it can be seen from the plots clearly,
the porous layers ruin the velocity distribution in the channel. In Fig. 11, air flows freely in the clear channel.
It has the minimum velocity values at the top and at the bottom of the collector whereas reaches its
maximum velocity value at the centerline of the collector as expected. In Fig.12, the channel is semi-filled
with porous material. As a result of those porous layers which fill the half of the channel, the velocity value is
decreased. On the other hand, the second half of the channel which is not filled with porous layers has a
similar velocity profile with the clear channel. And finally, in Fig.13, the channel is fully filled with porous
layers. These porous layers composed a resistance to the flow in the channel. So the velocity of air is
decreased a lot.

Change of Absorber Surface Temperature

The temperature data were collected longitudinally and transversely for each 8 experiments (clear and layer
assisted ones). The inlet and outlet temperature values are plotted and given for clear, 4-layer porous
assisted (semi-filled) and 7-layer porous assisted (fulfilled) channels. The average inlet temperatures for
these 3 cases are 28.5°, 28.6° and 28.5°C while as the average outlet temperatures are 55.34°, 56.5° and
56.67°C, respectively. The temperature differences for three cases are very close to each other since
constant heat flux is applied to the bottom of the channel. Moreover, the energy balance between inlet and

1096
outlet, and the imposed heat is established. It is observed that the net heat transfer to the surface is almost
equal to the heat transferred to the air.

Fig.11. The velocity profile for clear channel.

Fig.12. The velocity profile for 4- layer porous assisted (semi-filled) channel.

Fig.13. The velocity profile for 7- layer porous assisted (fulfilled) channel.

The absorber temperatures were measured from five different points. It can be seen clearly from these data
that, the absorber plate temperature decreases as the porous layer number increases. The maximum
absorber plate temperature value for clear channel is 116.8°C. This thermocouple value decreases to 82.3°C
by adding 7 layers of porous material inside the collector. All values for these five thermocouples are
decreased from clear channel to full filled channel. When the flow channel is clear, the absorber plate
temperature is at the highest value. For the channel with clear fluid, a thermal resistance occurs at the
bottom of the channel and this prevents heat transfer from the surface to the air. By inserting of porous layer,

1097
the thermal resistance is disturbed. Moreover, heat conduction in the transverse of the channel increases.
Porous media provides also mixing of the fluids in the channel. That’s why heat transfer coefficient from
surface to fluid increases and, as a result, surface temperature reduces.

Fig.14. The change of absorber plate temperature according to porous layers.

CONCLUSIONS
The enhancement of heat transfer in a solar air heater collector by using porous medium was investigated in
this study. The heat balance between the net imposed heat transfer to the plate and heat transferred to the
air was satisfied. The absorber surface temperature is decreased as the porous layers added inside the flow
channel. The porous media made the temperature distribution inside the channel more uniform. When the
temperature distributed uniformly inside the channel, the absorber temperature is decreased and
consequently heat transfer to the surroundings is reduced. This causes the increase of solar air heater
performance.

NOMENCLATURE
A area, m2
cp specific heat ,kJ/kgC
g gravitational acceleration, m/s2
L characteristic length, m
m mass, kg
Q heat, kW
T temperature, C
V velocity, m/s

Greek Letters
α thermal diffusity, m2/s
β thermal expansion, 1/K
˅ kinematic viscosity, m2/s
 density, kg/m3

Subscripts
s surface
∞ free stream
i element

Superscripts
.
rate

1098
REFERENCES
Hachemi, A. 1999. Comperative study on the thermal performances of solar air heater collectors with
selective and nonselective absorber-plate. Renewable Energy 17: 103-112.
Hans. V.S., R.P. Saini and J.S. Saini.2010. Heat transfer and friction factor correlations for a solar air heater
duct roughened artificially with multiple V-ribs. Solar Energy 84:898-911.
Karmare, S.V. and A.N. Tikekar.2009. Experimental investigation of optimum thermahydraulic performance
of solar air heaters with metal rib grits roughness. Solar Energy 83: 6-13.
Kumar, A., J.L. Bhagoria and R.M. Sarviya..2009. Heat transfer and friction correlations for artificially
roughened solar air heater duct with discrete W-shaped ribs. Energy Conservation and Management
50:2106-2117.
Midilli, A., H. Olgun and T. Ayhan.1999. Experimental studies on mushroom and pollen drying. International
Journal of Energy Research 23:1143-1152.
Mohsen-Languri, E., H. Taherian, R. Masoodi and J.R. Reisel.2009. An energy and exergy study of a solar
thermal air collector. Thermal Science 13(1): 205-216
Moummi, N., S. Youcef-Ali, A. Moummi and J.Y. Desmons .2004. Energy analysis of solar air collector with
rows of fins. Renewable Energy 29: 2053-2064.
Ozturk, H.H. and Y. Demirel. 2004. Exergy-based performance analysis of packed-bed solar air heaters.
Internatioanl Journal of Energy Research 28:423-432.
Pakdaman, M. F., A. Lashkari, H. B. Tabrizi and R. Hosseini .2011. Performance evaluation of a natural-
convection solar air-heater with a rectangular-finned absorber plate. Energy Conservation and Management
52: 1215-1225.
Potdukhe, P.A. and S.B. Thombre. 2008. Development of a new type of solar dryer. International Journal of
Energy Research 32:765-782.
Prasad, S.B., J.S. Saini and K.M. Singh. 2009. Investigation of heat transfer and friction characteristics of
packed bed solar air heater using wire mesh as packing material. Solar Energy 83:773-783.
Sahu, M.M. and J.L. Bhagoria .2005. Augmentation of heat transfer coefficient by using 90° broken
transverse ribs on absorber plate of solar air heater. Renewable Energy 30: 2057-2073.
Varshney, I. and J.S. Saini.1998. Heat transfer and friction factor correlations for rectangular solar air heater
duct packed with wire mesh screen matrices. Solar Energy 62(4):255-262.
Yeh, H.M., C.D.Ho and J.Z. Hou .2002. Collector efficiency of double-flow solar air heater with fins attached.
Energy 27: 715-727.

1099
 NOVEL FIN DESIGN FOR THE PANEL TYPE RADIATORS USING CFD
1 2
E. Aydar , I.Ekmekci

1
TUBITAK Marmara Research Center Gebze Campus Energy Institute,
41470 Kocaeli, E-mails: emir.aydar@mam.gov.tr
2
Marmara University, Engineering Faculty, Mechanical Engineering Department, Goztepe Campus,
34722 Istanbul, E-mails: iekmekci@marmara.edu.tr

ABSTRACT
There are many panel radiator manufacturers in Turkey. Their panel radiator designs are very similar
and radiators that have been manufactured by various producers have very similar thermal efficiency
values and specific heating per unit weight of the radiator.
In this study, CFD analysis of existing panel radiators were made with a commercial CFD software of
STAR-CCM+ with top-bottom-opposite-end connection method in three-dimensional space. Panel-
convector-convector-panel, Type-22-600x500 steel panel radiator was used in this numerical study.
In the content of this analysis, new configurations at fin design were performed to increase the rate of
heat transfer of the panel radiator. Four fin designs were drawn and examined. Since triangular
profile fin design contains less material and has higher heat flux, it is more efficient than the other
profiles, and therefore is more suitable for applications requiring minimum weight.

INTRODUCTION
Panel radiators are the most widely-used central-heating emitters to heat most homes and offices in
Europe. There is a high demand for panel radiators due to their compact design and less place
requirements. 80% of the heat output from radiators is natural convection, 20% of the heat output
from radiators is radiation. Although radiators are known as radiator, most of their output is by natural
convection (Beck et al., 2001 and Beck et al., 2004). Since panel radiators are elegant design, light,
cheap and occupy less place, they are in common use at homes and offices. Radiators are the
combination of water circulation channels and high convectors that is directly welded onto these
panels. All panel radiators release a combination of radiant and convective heat into a room as hot
water flows through them.
There is insufficient 3-D numerical study about panel radiator in literature. (Lu et al., 1997) showed
the airflow and temperature field in a heated room by using a CFD package. This study includes
using CFD to predict a three dimensional airflow field and temperature distribution in a full-scale
enclosure with a radiator; the predicted airflow pattern and temperature field will be validated by the
measured data carried out by Lebrun and Marret and using CFD to predict the smoke particle
distribution in the room. (Zhai et al., 2002 and Zhai et al., 2005) are interested in coupling an energy
simulation (ES) program with a computational fulid dynamics (CFD) program. Static and dynamic
coupling strategies have been implemented to bridge the gap between ES and CFD programs to
reduce the computing costs while achieving accurate results but conserve the accuracy and details
of the computed results.
The aim of this study is that there are many panel radiator manufacturers in Turkey, but their panel
radiator designs are very similar and due to this similar design, radiators that have been
manufactured by various producers have very similar values of thermal efficiency and specific
heating for unit weight of the radiator. In our research, CFD analysis of existing panel radiator will be
made with a commercial CFD software of STAR-CCM+ with top-bottom-opposite-end connection
method using forced convection model in three-dimensional space. After that, four fin designs will be
created in order to obtain more efficient panel radiator with containing less material with respect to
original radiator fin design.

MATHEMATICAL MODELING
When inlet water temperature and mass flow rate were fixed, optimum air-side heat transfer
coefficient was obtained and used for the other numerical analysis. Mass flow rate was computed
from energy balance equation as given below.
. .
Q  m c p  T (1)
.
899  m 4189.8  (75  65) (2)
.
m  0.0214 kg / s (3)

1100
All conditions and properties are defined via the STAR-CCM+ GUI. This study incorporates both
multi-region and conjugate heat transfer. Outlet temperature was obtained according to the fixed inlet
temperature and mass flow inlet. In this numerical study, rate of heat transfer of the panel radiator
was obtained by the help of the forced convection model. Rate of heat transfer of the whole panel
type radiator is 899 W/m obtained from experimental result.
Rate of heat transfer of the one module, which was calculated from the rate of heat transfer of the
whole panel type radiator times the module length as follows.

Q'  Q  L (4)
Q  899  0.0017
'
(5)
Q  15.28 W
'
(6)

Mass flow rate is found from the energy balance.

. .
Q  m c p  T (7)
.
15.28  m 4189.8  (75  65) (8)
.
m  0.00036 kg / s (9)

Under most practical conditions, the flow in a tube is laminar for Re<2300, fully turbulent for
Re>10000, and transitional in between. But it should be kept in mind that in many cases the flow
becomes fully turbulent for Re>4000 (Cengel, 2006). In order to define nature of the flow in a circular
tube, Reynolds number was computed as follows.
.
4 m
Re  (10)
  D
4  0.00036
Re  (11)
3.14  0.017  0.0004
Re  67.20 (12)

which is less than the critical Reynolds number of 2300, and the flow regime of a circular tube is
laminar.

COMPUTATIONAL METHODOLOGY
This numerical study is based on a domestic radiator which is in accordance with Standard TS
EN442. Half of the geometry was used for numerical study because of the symmetrical structure of
the panel radiator geometry, and then the drawn geometry was imported to STAR-CCM+.
The panel radiator has been drawn by means of CAD software according to original measurements
of the PCCP (panel-convector-convector-panel) arrangement of the panel radiator. PCCP, Type-22
steel panel radiator is used in this numerical study. Heat transfer rate of the steel panel radiator has
been measured according to EN 442 in an accredited laboratory (EN 422–2, 1997). Connection type
of the steel panel radiator used in this study is top-bottom-opposite-end. The outer panel is made of a
shaped plate with horizontal and vertical depressions. In order to increase heating performance,
some types of radiators are provided with a convector plate, welded to the vertical waterways of the
panel. The panel is made of two stamped steel sheets welded together with a seam weld on the
perimeter and with a spot weld where the depressions are. Since the panel radiator is symmetrical, a
half of the geometry can be used for numerical study. A 500-mm-long, 600-mm-high and 105-mm-
wide is the dimensions of the half of the panel radiator. A 150-mm-high, 473.5-mm-long is the
dimensions of the convector. First four and last four convectors are smaller than middle ones to
prevent the difficulties at the installation of T-Junctions. Panels and convectors are made of steel
sheet. Water channels are 1.1 mm thick, the panel radiator has 25 mm pitch. The convectors are 0.5
mm thick, and have a height of 37 mm from the base. The width of one panel 12 mm, respectively.
Four fin designs were performed to enhance heat transfer rate by using less material usage. It is
aimed to manufacture a more efficient panel radiator. Four different fin designs were performed by

1101
the help of CAD software. Fin material, height, width, length weren’t changed. Cross-sectional area
of the fin was only changed. Original and new fin designs are shown in Figure 1, 2, 3.

Fig. 1. Original fin and new fin design-1 Fig. 2. New fin design-2 and triangular profile fin design

Fin design-3 is different from the other ones. Original design and the other fin ones have two contact
faces but Figure 3 shows that fin design-3 have three contact faces between panel and convector.

Fig. 3. New fin design-3

Mesh sizes of the panel radiator geometry are listed in Table 1.

Table 1. Mesh sizes of the panel radiator geometry

Number of Prism Layers Number of Prism Layers

2

Prism Layer Stretching Prism Layer Stretching

1.2
Prism Layer Thickness Absolute Size 1 mm

Surface Curvature # Pts/circle 36

Surface Size Absolute Size 4 mm

Surface Size Absolute Size 2 mm

The panel radiator with TBOE connection was meshed using tetrahedral volume meshing model and
also surface remesher and prism layer mesher were selected. Figure 4, 5, 6 show the mesh type and
the mesh density of outer side of the original radiator fin design and novel fin designs.

Fig. 4. Computational meshes of the panel radiator with original fin and new design-1

1102
Fig. 5. Computational meshes of the panel radiator with new design-2 and triangular fin

Fig. 6. Computational meshes of the panel radiator with new design-3

The computational meshes with high resolution of approximately 1,800,000-1,900,000 for fluid region
and 3,000,000-4,400,000 cells for solid region were created for the flow analysis of the panel
radiator. The numerical simulations were carried out using the finite volume CFD-based code STAR-
CCM+ with parallel processing to accelerating the calculations. The computations were performed on
a 3.00 GHz CPU and 16GB of RAM workstation.
Steel panel radiator (AISI 316) used in this study was computed at inlet and outlet temperatures
respectively, 75 ºC and 65 ºC according to EN 442 (Incropera et al., 1993). In this numerical study,
inlet temperature and mass flow rate will be determined and outlet temperature will be computed in
order to obtain optimum air-side convective heat transfer coefficient according to boundary
conditions. Inlet water temperature was specified as 75 ºC according to EN 442 (EN 422–2, 1997).
Water was selected as liquid material. Material properties of water was computed as the average of
inlet and outlet water temperatures. The tables of thermophysical properties of saturated water was
used to compute material properties of water. Stainless steel (AISI 316) was selected for solid
material. Material properties of stainless steel were determined according to the database of
commercial CFD code of STAR-CCM+. Material properties of stainless steel (AISI 316) and water
are listed in Table 2. Boundary conditions of the panel radiator geometry are also listed in Table 3.

Table 2. Material properties

Material Type Density(kg/m³) Dynamic Viscosity (Pa- Specific Heat Thermal Conductivity
s) (J/kgK) (W/mK)
Water 977.7 0.0004 4189.8 0.6
Stainless Steel (AISI 316) 8238 468 13.4

Table 3. Boundary conditions of the panel radiator geometry

Boundary Type Physics Values Inputs
Inlet Mass Flow Inlet Mass Flow Rate 0.0214 kg/s
Total Temperature 75 ºC
Direction x-direction
Outlet Flow-Split Outlet Split Ratio 1
Direction x-direction
stSteel
Four T-Fittings Wall Thermal Specification Adiabatic
Panel Surface Wall Thermal Specification Convection
Ambient Temperature 20 ºC

Heat Transfer Coefficient 6.5 W/m²K

Thermal Specification Convection
Convectors Surface Wall Ambient Temperature 20 ºC
Heat Transfer Coefficient 6.5 W/m²K

1103
RESULTS AND DISCUSSION
The view of the scalar temperature distribution of fluid flow, original radiator fin design at air-side
convective heat transfer coefficient of 6.5 W/m²K are given in Figure 7, 8.

Fig 7. Scalar temperature distribution of fluid flow Fig 8. Scalar temperature distribution of original fin design

Figure 7, 8 show that there are high temperature gradients are readily seen at the middle-bottom
region of panel and fluid flow regions. This high temperature gradients are caused by distinct values
of mass flow rates at vertical ducts.

EVALUATION OF NUMERICAL RESULT

Different air-side convective heat transfer coefficients were tested so as to compute outlet water
temperature of 65 °C that is related to panel radiator in accordance with TS EN442. Inlet water
temperature of 75 °C is fix for all study. The numerical results of convective heat transfer coefficients
of 5 W/m2K, 6.5 W/m2K, 8 W/m2K, 10 W/m2K were compared each other. Numerical results at
different air-side heat transfer coefficients of the panel radiator with TBOE connection are listed in
Table 4.

Table 4. Numerical results at different air-side heat transfer coefficients

Specifications Symbol Values
Heat Transfer Coefficient (W/m2K) h 5 6.5 8 10
Mass Flow Inlet (kg/h) 
77.2 77.2 77.2 77.2
m
Inlet Water Temperature (°C) T1 75 75 75 75
Outlet Water Temperature (°C) T2 66.5 64.9 63.6 62
Room Temperature (°C) Tr 20 20 20 20
Heat transfer rate (W) (Based on Enthalpy) 
Qh 755 894.2 1015.8 1157.1

Heat transfer rate (W) (The Heat Output


from the Radiator) 753.7 892.1 1015.4 1156
Q tot

Convective heat transfer coefficient of air-side is directly proportional to the thermal output of the
radiator. Normally the larger air-side heat transfer coefficent, the higher the rate of heat transfer from
the panel radiator. Table 4 should offer a good compromise between heat transfer rates of numerical
study and experimental study whose heat transfer rate is 899 W, when the air-side heat transfer
coefficent is 6.5 W/m2K. Therefore, the air-side heat transfer coefficent of 6.5 W/m2K was used in all
numerical study.
The heat transferred to the air by means of forced convection model, which was calculated from the
difference in enthalpy times the mass flow rate between the inlet and outlet of the water within the
radiator as following.

. .
Qh  m (h1  h2 ) (13)

For the fix inlet temperature of 75 °C, fix mass flow rate of 0.0107 kg/s and convective heat transfer
coefficient of 6.5 W/m2K, outlet temperature of 64.9 °C was obtained.

1104
EFFECTS OF ORIGINAL FIN DESIGN OF THE PANEL RADIATOR ON FIN EFFICIENCY

In the limiting case of zero thermal resistance or infinite thermal conductivity, the temperature of the
fin will be uniform at the base value of Tb . The heat transfer from the fin will be maximum in this case
and can be expressed as
.
Q fin,max  hAfin (Tb  T ) (14)

In reality, however, the temperature of the fin will drop along the fin, and thus the heat transfer from
the fin will be less. To account for the effect of this decrease in temperature on heat transfer, fin
efficiency is defined as

.
Q fin
 fin  . (15)
Q fin,max


where Afin is the total surface area of the fin. Q fin means that actual heat transfer rate from the fin

and Q fin ,max that ideal heat transfer rate from the fin if the entire fin were at base temperature. Each
fin was dissolved as shown in Fig. 9. Actual heat transfer rate from the fin and ideal heat transfer rate
from the fin were computed for each fin and they are listed in Table 5.

Fig. 9. Split each fin of the panel radiator

After values had been computed, fin efficiency for each fin of the original radiator fin design was
computed by using Eq. 15. Fin efficiency results of each fin of the original radiator fin design are
listed in Table 10.

EFFECTS OF NEW FIN DESIGN-1 ON FIN EFFICIENCY

The view of the scalar temperature distribution of fin design-1 at air-side convective heat transfer
coefficient of 6.5 W/m²K is given in Fig. 10.

Fig. 10. Scalar temperature distribution of fin design-1 Figure. 11. Scalar temperature distribution of fin design-2

1105
The values used in order to compute fin efficieny by means of Eq. 15 are listed in Table 6 for
design1.

Table 6 The values of each fin design-1 to compute fin efficiency

Heat transfer rate Heat transfer Base temperature Ideal heat transfer rate from
Specifications from the convector surface area (m²) (ºC) the convector (W)
(W)
Fin 1 16.289 0.1083 69.016 34.504
Fin 2 15.494 0.1081 66.866 32.930
Fin 3 14.321 0.1081 63.180 30.340
Fin 4 13.155 0.1081 59.517 27.766
Fin 5 13.469 0.1194 56.292 28.166
Fin 6 12.690 0.1195 54.258 26.609
Fin 7 12.292 0.1194 53.135 25.716
Fin 8 12.287 0.1194 52.932 25.558
Fin 9 12.568 0.1194 53.853 26.273
Fin 10 13.108 0.1195 55.516 27.587
Fin 11 13.719 0.1194 57.167 28.869
Fin 12 14.384 0.1195 59.153 30.412
Fin 13 15.062 0.1194 61.114 31.908
Fin 14 15.671 0.1194 62.764 33.189
Fin 15 16.201 0.1194 64.312 34.390
Fin 16 15.085 0.1082 65.602 32.071
Fin 17 15.329 0.1081 66.527 32.692
Fin 18 15.456 0.1081 66.852 32.920
Fin 19 15.388 0.1082 66.473 32.684
After values had been computed, fin efficiency for each fin of fin design-1 was computed by using Eq.
15. Fin efficiency results of each fin of fin design-1 are listed in Table 10.

EFFECTS OF NEW FIN DESIGN-2 ON FIN EFFICIENCY

The view of the scalar temperature distribution of fin design-2 at air-side convective heat transfer
coefficient of 6.5 W/m²K is given in Fig. 11.

The values used in order to compute fin efficieny by means of Eq. 15 are listed in Table 7. for design- 2.

Table 7. The values of each fin design-2 to compute fin efficiency

Heat transfer rate Heat transfer Base temperature Ideal heat transfer rate from
Specifications from the convector surface area (m²) (ºC) the convector (W)
(W)
Fin 1 15.339 0.0853 68.933 27.130
Fin 2 14.657 0.0851 67.071 26.037
Fin 3 13.626 0.0851 63.740 24.194
Fin 4 12.591 0.0851 60.469 22.385
Fin 5 12.943 0.0939 57.422 22.840
Fin 6 12.188 0.0939 55.386 21.597
Fin 7 11.909 0.0940 54.372 21.001
Fin 8 11.926 0.0939 54.473 21.040
Fin 9 12.191 0.0939 55.275 21.530
Fin 10 12.644 0.0939 56.543 22.304
Fin 11 13.192 0.0939 58.228 23.332
Fin 12 13.797 0.0939 60.083 24.464
Fin 13 14.396 0.0939 61.802 25.513
Fin 14 14.903 0.0939 63.370 26.470
Fin 15 15.429 0.0939 64.752 27.314
Fin 16 14.296 0.0851 65.950 25.417
Fin 17 14.575 0.0851 66.827 25.902
Fin 18 14.667 0.0851 67.122 26.065
Fin 19 14.565 0.0851 66.772 25.871

1106
After values had been computed, fin efficiency for each fin of fin design-2 was computed by using Eq.
15. Fin efficiency results of each fin of fin design-2 are listed in Table 10.

EFFECTS OF TRIANGULAR PROFILE FIN DESİGN ON FIN EFFICIENCY

The view of the scalar temperature distribution of triangular profile fin design at air-side convective
heat transfer coefficient of 6.5 W/m²K is given in Figure 12.

Fig 12. Scalar temperature distribution of triangular profile fin design Fig 13. Scalar temperature distribution of convector for Design-4

The values used in order to compute fin efficieny by means of Eq. 15 are listed in Table 8 for
triangular profile fin design.

Table 8. The values of each triangular profile fin design to compute fin efficiency
Heat transfer rate from the Heat transfer surface Base Ideal heat transfer rate from
Specifications
convector (W) area (m²) temperature(ºC) the convector (W)
Fin 1 14.941 0.0767 68.874 24.366
Fin 2 14.254 0.0765 66.805 23.273
Fin 3 13.179 0.0765 63.275 21.518
Fin 4 12.274 0.0765 60.362 20.070
Fin 5 13.088 0.0861 57.811 21.160
Fin 6 12.486 0.0861 56.102 20.204
Fin 7 12.243 0.0861 55.385 19.803
Fin 8 12.243 0.0861 55.350 19.783
Fin 9 12.462 0.0861 56.166 20.240
Fin 10 12.878 0.0861 57.320 20.886
Fin 11 13.378 0.0861 58.857 21.746
Fin 12 13.933 0.0861 60.528 22.681
Fin 13 14.448 0.0861 62.086 23.553
Fin 14 14.958 0.0861 63.580 24.389
Fin 15 15.359 0.0859 64.935 25.089
Fin 16 14.047 0.0765 66.073 22.909
Fin 17 14.303 0.0765 66.985 23.363
Fin 18 14.417 0.0764 67.321 23.499
Fin 19 14.120 0.0757 67.160 23.205

After values had been computed, fin efficiency for each fin of triangular profile fin design was
computed by using Eq. 15. Fin efficiency results of each fin of triangular profile fin design are listed in
Table 10.

EFFECTS OF NEW FIN DESIGN-3 ON FIN EFFICIENCY

The view of the scalar temperature distribution of fin design-3 at air-side convective heat transfer
coefficient of 6.5 W/m²K is given in Figure 13.

1107
The values used in order to compute fin efficieny by means of Eq. 15 are listed in Table 9 for design-3.

Table 9 The values of each fin design-3 to compute fin efficiency

Heat transfer rate Heat transfer Base temperature Ideal heat transfer rate from
Specifications from the convector surface area (m²) (ºC) the convector (W)
(W)
Fin 1 27.213 0.1510 67.765 46.881
Fin 2 25.585 0.1508 64.959 44.068
Fin 3 22.985 0.1508 60.328 39.529
Fin 4 20.120 0.1507 55.323 34.600
Fin 5 19.838 0.1664 51.152 33.694
Fin 6 18.344 0.1664 48.791 31.140
Fin 7 17.793 0.1665 47.868 30.160
Fin 8 17.965 0.1665 48.119 30.431
Fin 9 18.526 0.1664 49.072 31.444
Fin 10 19.508 0.1664 50.586 33.081
Fin 11 20.765 0.1665 52.736 35.428
Fin 12 22.176 0.1665 55.054 37.937
Fin 13 23.616 0.1665 57.354 40.426
Fin 14 24.915 0.1665 59.497 42.745
Fin 15 26.155 0.1664 61.423 44.803
Fin 16 24.598 0.1508 63.156 42.301
Fin 17 25.314 0.1508 64.461 43.580
Fin 18 25.648 0.1508 64.987 44.096
Fin 19 25.398 0.1508 64.729 43.843

Fin efficiency results of new fin designs according to the Eq. 15 are listed in Table 10 for each fin.
They were compared with original radiator fin design.

Table 10 Fin efficiency comparison between original radiator fin design and new fin designs
Fin Efficiency ( fin ) (%)

Specifications

Fin 1 58.47 47.20 56.53 61.31 58.04

Fin 2 58.79 47.05 56.29 61.24 58.05
Fin 3 58.86 47.20 56.31 61.24 58.14
Fin 4 58.84 47.37 56.24 61.15 58.15
Fin 5 59.60 47.82 56.66 61.85 58.87
Fin 6 59.94 47.69 56.43 61.79 58.90
Fin 7 59.55 47.79 56.70 61.82 58.99
Fin 8 59.97 48.07 56.68 61.88 59.03
Fin 9 59.58 47.83 56.62 61.57 58.91
Fin 10 59.63 47.51 56.68 61.65 58.97
Fin 11 59.61 47.52 56.54 61.51 58.61
Fin 12 59.69 47.29 56.39 61.43 58.45
Fin 13 59.54 47.20 56.42 61.34 58.41
Fin 14 59.75 47.21 56.30 61.33 58.28
Fin 15 59.56 47.10 56.48 61.21 58.37
Fin 16 59.61 47.03 56.24 61.31 58.14
Fin 17 59.71 46.88 56.26 61.22 58.08
Fin 18 59.85 46.95 56.27 61.35 58.16
Fin 19 59.71 47.08 56.29 60.84 57.92

1108
These fin efficiency results can be easily seen in graph as shown in Figure 14.

Fig. 14. Fin efficiency comparison between original radiator fin design and new fin designs

In Figure 14, it can be seen that arithmetic average of triangular profile fin design is 61.42%,
arithmetic average of original radiator fin design is 59.48%, arithmetic average of design-3 is 58.44%,
arithmetic average of fin design-2 is 56.43%, arithmetic average of fin design-1 is 47.34%. The most
increase at the fin efficiency results was obtained for triangular profile fin design.

The specifications of numerical results of original radiator fin design and new fin designs are listed
together in Table 11.

Table 11. Numerical results of new fin designs according to original radiator fin design

Specifications

Area Magnitude of Fin (m2) 1.616 2.180 1.715 1.559 3.038

Area Magnitude Increase Ratio Relative to
Original Radiator 1 1.349 1.061 0.964 1.879
Heat Transfer Rate of Panel (W) 190.028 187.865 189.633 190.921 174.955
Panel Decrease Ratio Relative to Original
Radiator Fin Design 1 0.988 0.997 1.004 0.920

Heat Transfer Rate of Fin (W) 256.020 272.044 259.912 259.090 426.538
Fin Increase Ratio Relative to Original
Radiator Fin Design 1 1.0625 1.015 1.011 1.666

Outlet Water Temperature (°C) 64.9 64.7 64.8 64.8 61.4

Enthalpy Difference (J/kg) 41673 42336 42136 42200 56338
Heat Transfer Rate (W) (Based on Eq.13)
894.172 908.398 904.136 905.479 1208.837

Increase Ratio for Heat Transfer Rate

Relative to Original Radiator Fin Design 1 1.015 1.011 1.012 1.351

Heat Flux (W/m²) 553.324 416.696 527.193 580.807 397.905

Decrease Ratio for Heat Flux 1 0.753 0.952 1.049 0.719

1109
The area magnitude of fin design-1 is %35 greater than original radiator fin design. The increase at
heat transfer rate from the radiator based on Eq.13 is %1.5. Heat flux is 24.7% smaller than original
radiator fin design. 6.1% increase at area magnitude of fin design-2 according to original radiator fin
design was obtained but %1.1 increase at heat transfer rate from the radiator based on Eq.13 was
obtained. Heat flux is 4.8% smaller than original radiator fin design. The area magnitude of triangular
profile fin design is %3.6 smaller than original radiator fin design. The increase at heat transfer rate
from the radiator based on Eq.13 is %1.2. Heat flux is 4.9% greater than original radiator fin design.
The area magnitude of fin design-3 is %87.9 greater than original radiator fin design. The increase at
heat transfer rate from the radiator based on Eq.13 is %35.1. Heat flux is 28.1% smaller than original
radiator fin design.

CONCLUSIONS
New fin designs were created in order to enhance the rate of heat transfer of the panel radiator with
containing less material. It is aimed to manufacture a more efficient panel radiator. Different air-side
convective heat transfer coefficients were tested so as to compute outlet water temperature of 65 °C
that is related to panel radiator in accordance with TS EN442. 6.5 W/m²K was optimum value fort his
panel radiator and used in these all numerical studies. The contact interface type was used to join
together to permit conjugate heat transfer between a fulid and solid region. Heat transfer rate of the
panel radiator was specified by the help of a commercial CFD code of STAR-CCM+ with top-bottom-
opposite-end connection. Four convector designs were drawn so as to compare their fin efficiencies
and heat transfer rates with respect to the original panel radiator. Since the increase at heat transfer
rate from the radiator based on Eq.13 is 1.2%, in case of 3.6% decrease at the area magnitude of
convector, and heat flux is 4.9% greater than original radiator fin design, most efficient fin among
these ones is triangular profile fin design. Arithmetic average of triangular profile fin design efficiency
is 3.151% greater than original radiator. Figure 14 proves most increase at fin efficiency was
obtained for triangular profile fin design, and thus triangular profile fin design is more efficient, has a
higher heat transfer rate and a lower cost material than original radiator fin design.

NOMENCLATURE
A total area [mm2]
Afin total surface area of the fin [m2]
cp specific heat [kj/kgK]
EN European standard
H panel radiator height [mm]
ha air side heat transfer coefficient [W/m2K]
hi enthalpy at inlet side [kj/kg]
ho enthalpy at outlet side [kj/kg]
k thermal conductivity [W/mK]
L module length of the radiator [mm]
.
m mass flow rate [kg/s]
.
Q heat transfer rate [W]

Q fin actual heat transfer rate from the fin [W]

Q fin,max ideal heat transfer rate from the fin [W]
.
Qh heat transfer based on enthalpy [W]
Q '
rate of heat transfer of the one module [W]
Re Reynolds number
TBOE flow and return top and bottom at opposite ends
TS Turkish standard
Tb base temperature [K]
Ti inlet temperature [K]
To outlet temperature [K]
T∞ ambienttemperature [K]
V fluid inlet velocity [m/s]

1110
Greek Symbols

∆T temperature difference between inlet and outlet side [K]

µ dynamic viscosity [Pa.s, kg/ms]
ρ density [kg/m3]
 fin fin efficiency

Subscripts

i inlet condition
o outlet condition

REFERENCES

Arslanturk, C. and Ozguc, A. F., Optimization of a central-heating radiator, Applied Energy, 83,
1190-1197, 2006.

Beck, S. B. M., Blakey, S. G., and Chung M. C., The effect of wall emissivity on radiator heat output,
Building Services Engineering Research & Technology, 12, 185-194, 2001.

Beck, S. M. B., Grinsted S. C., Blakey S. G., et al., A novel design for panel radiators. Applied
Thermal Engineering, 24, 1291-1300, 2004.

Cengel, Y. A., Heat and Mass Transfer (Third Ed.), Mcgraw Hill Book Companies, Inc, 2006.

Chen, Q., Comfort and energy consumptions analysis in buildings with radiant panels, Energy and
Buildings, 14, 287-297, 1990.

EN 422–2, Radiators and Convectors, part 2, BSI, 1997.

Harris, D. J., Use of metallic foils as radiation barriers to reduce heat losses from Buildings, Applied
Energy, 52, 331-339, 1995.

Holman J.P., Heat Transfer (Seventh Ed.), McGraw-Hill Book Companies, New York, 1992.
Incropera F.P., DeWitt D.P., Introduction to Heat Transfer (Third Ed.), Wiley, 1993

Lu, W., Howarth, A. T., and Jeary, A. P., 1997. Prediction of airflow and temperature field in a room
with convective heat source, Building and Environment, 32, 541- 550, 1997.

Peach S., Radiators and other convectors. J Inst. Heating Ventilation Eng.; 39, 239–53, 1972.

Ward I.C., Domestic radiators: performance at lower mass flow rates and lower temperature
differentials than those specified in standard performance tests, Building Serv. Eng. Res. Technol.,
12, 87–94, 1991.

Zhai, Z., Chen Q., Haves, P., et al., On approaches to couple energy simulation and computational
fluid dynamics programs. Building and Environment. 37, 857-864, 2002.

Zhai, Z. J. and Chen, Q. Y., Performance of coupled building energy and CFD simulations. Energy
and Buildings, 37, 33-344, 2005.

1111
 EFFECT OF RADIATION-CONDUCTION INTERACTION ON MIXED CONVECTION FROM
A SLENDER CYLINDER
1 2
A. Kaya and O. Aydin
1
Department of Mechanical Engineering, Kahramanmaras Sutcu Imam University, 46100 Kahramanmaras, TURKEY, e-
mail: kaya38@ksu.edu.tr
2
Department of Mechanical Engineering, Karadeniz Technical University, 61080 Trabzon, TURKEY, e-mail:
oaydin@ktu.edu.tr

ABSTRACT
The problem of steady laminar mixed convection heat transfer about a slender cylinder is studied numerically,
taking into account the radiation-conduction effect. The resulting governing equations are transformed into the
non-similar boundary layer equations and solved using the Keller box method. The velocity and temperature
profiles as well as the local skin friction and the local heat transfer parameters are determined for different values
of the governing parameters, mainly the Richardson number, the Planck number (i.e. radiation-conduction
parameter) and the surface temperature ratio. For some specific values of the governing parameters, the results
agree very well with those available in the literature. Generally, it is determined that the local heat transfer
coefficient increase, increasing the Richardson number, Ri (i.e. the mixed convection parameter) and Planck
number (i.e. radiation-conduction parameter) Rd or decreasing the surface temperature ratio θw.

Keywords: slender cylinder, buoyancy effect, Radiation, surface temperature ratio

INTRODUCTION
Mixed convection flow along vertical cylinders is important in situations encountered in the areas of geothermal
power generation and drilling operation when free stream velocity and induced buoyancy velocity are of
comparable order and received much less attention. Chen and Mucoglu (1975) analyzed the buoyancy and
transverse curvature effects on forced convection of Newtonian fluid flow along an isothermal vertical cylinder
using the local non-similarity method. The same problem for a uniform surface heat flux case was conducted by
Mucoglu and Chen (1976). Lee et al. (1987) studied the problem of mixed convection along a vertical cylinder with
uniform surface heat flux for the entire mixed convection regime, ranging from pure forced convection to pure free
convection by employing the buoyancy and curvature parameters. Hossain and Alim (1997) studied a free
convection flow of an optically dense viscous incompressible fluid along a vertical thin circular cylinder with effect
of radiation. Hossain et al. (1999) studied mixed convection flow of an optically dense viscous incompressible fluid
along a horizontal circular cylinder with the effect of radiation when the surface temperature is uniform. Takhar and
Nath (2000) studied the mixed convection flow over a continuous moving vertical slender cylinder under the
combined buoyancy effect of thermal and mass diffusion. Kumari and Nath (2004) studied the effects of localized
cooling/heating and injection/suction on the mixed convection flow on a thin vertical cylinder. Datta et al. (2006)
obtained the non-similar solution of a steady laminar forced convection boundary layer flow over a horizontal
slender cylinder including the effect of non-uniform slot injection (suction). Roy et al. (2007) developed general
analysis for the influence of non-uniform double slot injection (suction) on the steady non-similar incompressible
laminar boundary layer flow over a slender cylinder. Anwar et al. (2008) numerically studied the steady mixed
convection boundary layer flow of a viscoelastic fluid over a horizontal circular cylinder in a stream flowing
vertically upwards for both cases of heated and cooled cylinders. Singh et al. (2008) studied unsteady mixed
convection flow over a rotating vertical slender cylinder under the combined effects of buoyancy force and thermal
diffusion with injection/suction where the slender cylinder is inline with the flow. The effect of surface curvature is
also taken into account, especially for the applications such as wire and fiber drawing, where accurate predictions
are desired. Molla et al. (2009) investigated natural convection boundary layer laminar flow from a horizontal
circular cylinder with uniform heat flux in presence of heat generation.
In the present paper the effect of radiation-conduction on mixed convection flow over a slender cylinder with
uniform surface temperature has been analyzed. The boundary layer equations governing the flow are reduced to
local non-similarity equations which are solved using the implicit finite difference method (Keller box). Numerical
results for the velocity and temperature profiles as well as local skin friction and local heat transfer parameters are
presented.

ANALYSIS
In this analysis, a two-dimensional steady laminar mixed convective flow of a viscous incompressible fluid along a
slender cylinder of length L and outer radius ro (L >> ro) is considered. The physical model and coordinate system
are shown in Fig. 1. The gravitational acceleration, g, acts in the downward direction. The temperature and velocity
at a distance remote from the cylinder are given by T∞ and u∞, respectively and the body has a uniform
temperature Tw. The radiative heat flux in the x-direction is considered negligible in comparison with that in the r-

1112
direction. Under foregoing assumptions and taking into account the Boussinesq approximation and the boundary
layer approximation, the system of continuity, momentum and energy equations can be written:

Tw

g
2r0
x, u

r r, υ
r=r0

u∞, T∞

Figure 1. The schematic of the problem

(1)

(2)

(3)

Here u and υ are the velocity components in the x and r direction, respectively, T is the temperature of the fluid, β
is the coefficient of thermal expansion, v is the kinematic viscosity, ρ is the fluid density, g is the acceleration due
to gravity.
The quantity qr on the right-hand side of equation (3) represents the radiative heat flux in the r direction. For
simplicity and comparison, the radiative heat flux term in the energy equation is analyzed by utilizing the
Rosseland diffusion approximation (Sparrow and Cess, 1961) for an optically thick boundary layer as follows:
4 T 4 qr 16   3 T 
qr   and  T  (4)
3 R r y 3 R r  y 
where  is the Stefan–Boltzmann constant, R is the Rosseland mean absorption coefficient. This
approximation is valid at points optically far from the bounding surface, and is good only for intensive absorption,
that is, for an optically thick boundary layer (Hossain and Pop, 1997).
The appropriate boundary conditions for the velocity and temperature of this problem are:
r  r0 ; u    0, T  Tw
r  ; u  u , T  T (5)
To seek a solution, the following dimensionless variables are introduced:
12
 4  vx   r 2  r02   u 
12
T  T
     ,    , 
 r0   u   4r0   vx  Tw  T
2
r
  x, r   r0  vu x  f  ,  ,  1    ,
12
(6)
r02
where  ( x, y) is the free stream function that satisfies Eq. (1) with u  1 r   r  and    1 r   x 

1113
In terms of these new variables, the velocity components can be expressed as
r  vu   f 
12
1
u  u f  ,   0     f   f    (7)
2 2r  x    
The transformed momentum and energy equations together with the boundary conditions, Eqs. (2), (3) and (4),
can be written as
(8)

1  1     f 
1         f    1      w  1  1     f     
3
(9)
Pr Pr 3Pr Rd     
with the boundary conditions;
f
  0; f   0, f   0,  1

  ; f   2,  0 (10)
The corresponding dimensionless groups that appeared in the governing equations defined as:
 cp v Gr g  Tw  T  r03 u r0 3 R k T
Pr   , Ri  , Gr  , Re  , Rd  , w  w (11)
k  Re 2v 2
v 4 T3
T
where Pr is the Prandtl number, Ri is the Richardson number, Gr is the Grashof number, Re is the Reynolds
number, Rd is the Planck number (radiation–conduction parameter), θw is the surface temperature ratio to the
ambient fluid.
In the above system of equations, the radiation conduction parameter is absent from the mixed convection heat
transfer problem when Rd→∞. It should be mentioned that the optically thick approximation should be valid for
relatively low values of the radiation-conduction parameter, Rd. According to Ali et al. (1984), some values of Rd for
different gases are: (1) Rd=10-30: carbon dioxide (100-650oF) with corresponding Prandtl number range 0.76-0.6;
(2) Rd=30-200: ammonia vapor (120-400oF) with corresponding Prandtl number range 0.88-0.84; (3) Rd=30-200:
water vapor (220-900oF) with corresponding Prandtl number 1.
The system of transformed equations together with the boundary conditions, Eqs. 8–10, have been solved
numerically using the Keller box scheme, an efficient and accurate finite-difference scheme, similar to that
described in Cebeci and Bradshaw (1977). For the sake of brevity, details of the numerical method are not
described, referring the reader to Cebeci and Bradshaw (1977).
In order to verify the accuracy of the present method, the present results were compared with those of Chen and
Mucoglu (1975) and Chang (2006). The comparison is found to be in good agreement, as shown in Table 1.

Table 1. Comparision of the local heat transfer rate and local skin friction with Pr=0.7, Rd→∞, θw=0.0 and Ri=0.0.
Chen and Mucoglu
Chang (2006) Present results
ξ (1975)
f   ,0  f   ,0  f   ,0 
0.0 1.3282 0.5854 1.3280 0.5852 1.3201 0.5846
1.0 1.9172 0.8669 1.9133 0.8658 1.8934 0.8599
2.0 2.3981 1.0986 2.3900 1.0940 2.3822 1.0918
3.0 2.8270 1.3021 2.8159 1.2982 2.8098 1.2902
4.0 3.2235 1.4921 3.2187 1.4925 3.2102 1.4898

RESULTS AND DISCUSSION

In this study, the effect radiation-conduction on mixed convection is examined. The following ranges of the main
parameters are considered: the mixed convection parameter Ri =0.0, 1, 5 and 10; Pr = 1.0, the radiation-
conduction parameter Rd = 30, 50, 100, and 150 and the surface temperature ratio θw= 1.5, 1.7, 2.0 and 2.3. The
combined effects of Ri, Rd and θw on the momentum and heat transfer are analyzed.
Figure 2 shows the effect of the buoyancy parameter Ri on the dimensionless velocity and temperature profiles
while Figure 3 shows the effect of the mixed convection parameter Ri on the local skin friction and the local heat
transfer parameters. The Richardson number, Ri represents a measure of the effect of the buoyancy in
comparison with that of the inertia of the external forced or free stream flow on the heat and fluid flow. Outside the
mixed convection region, either the pure forced convection or the free convection analysis can be used to describe
accurately the flow or the temperature field. Forced convection is the dominant mode of transport when Ri→0,

1114
 whereas free convection is the dominant mode when Ri→∞. Buoyancy forces can enhance the surface heat
4 transfer rate when they assist the forced convection (Aydin and
1.0Kaya, 2008).

Ri=0 Pr=1
Ri=0 Pr=1 Ri=1 Rd=30
Ri=1 Rd=30 0.8 Ri=5  =1.7
3 Ri=5 w=1.7 Ri=10
w

Ri=10

0.6
f'(1,)

(1,)
2

0.4

1
0.2

0 0.0
0 2 4 6 8 0 1 2 3 4 5 6

 

(a) (b)

Figure 2. Dimensionless velocity (a) and temperature (b) profiles for different Ri while Pr=1.0, Rd=30, θw=1.7 and ξ=1.

25 1.8
Ri=0 Pr=1
Ri=0 Pr=1 1.6 Ri=1 Rd=30
Ri=1 Rd=30 Ri=5 w=1.7
20 Ri=5 w=1.7 1.4 Ri=10
Ri=10

1.2
15
1.0
-'(,0)
f''(,0)

0.8
10
0.6

0.4
5

0.2

0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0

 
(a) (b)
Figure 3. Numerical values of local skin friction (a) and local heat transfer (b) against the streamwise distance ξ for different Ri while Pr=1.0, Rd=30
and θw=1.7.
Figure 4 shows the dimensionless velocity and temperature profiles inside the boundary layer for different values
of the radiation parameter Rd. The increasing of conduction-radiation parameter increases temperature gradients
near the wall (Fig. 5a), which increases heat transfer rates, this is due to the fact that radiation effects increase
temperatures of ionized gases and the absence of radiation defines small temperatures (Fig. 5b).

3.0 1.0

Rd=30 Pr=1
2.5 Rd=50 Ri=5
0.8 w=1.7
Rd=100
Rd=150
2.0
0.6
(1,)
f'(1,)

1.5
Rd=30
Rd=50 0.4
1.0 Rd=100
Rd=150
(a) (b)
0.2
0.5 Pr=1
Ri=5
w=1.7

0.0 0.0
0 2 4 6 8 0 1 2 3 4 5 6

 
1115
 Figure 4. Dimensionless velocity (a) and temperature (b) profiles for different R d while Pr=1.0, Ri=5, θw=1.7 and ξ=1.

16 1.6
Rd=30 Pr=1
1.4 Ri=5
14 Rd=50
w=1.7
Rd=100
12 1.2 Rd=150

10 1.0

-'(,0)
f''(,0)

8 0.8

6 0.6
Rd=30 Pr=1
4 Rd=50 Ri=5 0.4
w=1.7
Rd=100
2 Rd=150 0.2

0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0

 

(a) (b)
Figure 5. Numerical values of local skin friction (a) and local heat transfer (b) against the streamwise distance ξ for
different Rd while Pr=1.0, Ri=5 and θw=1.7.

The effect of surface temperature parameter θw on the velocity and temperature profiles is shown in Fig. 6.
Increasing the surface temperature ratio increases dimensionless velocity profile (Fig. 6a), which is shown to
increase the local skin friction parameter (Fig. 7a). Moreover increasing the temperature ratio also increased the
temperatures inside the boundary layer (Fig. 6b) and consequently decreases the local heat transfer parameter
(Fig. 7b).
3.0 1.0

w=1.5 Pr=1
2.5 w=1.7 Ri=5
0.8 Rd=30
w=2.0
w=2.3
2.0
0.6
(1,)
f'(1,)

1.5
w=1.5 Pr=1
w=1.7 Ri=5 0.4
Rd=30
1.0 w=2.0
w=2.3
0.2
0.5

0.0 0.0
0 2 4 6 8 0 1 2 3 4 5 6

 
(a) (b)
Figure 6. Dimensionless velocity (a) and temperature (b) profiles for different θw while Pr=1.0, Ri=5, Rd =30 and ξ=1.
16 1.6
w=1.5 Pr=1
14 1.4 Ri=5
w=1.7 Rd=30
w=2.0
12 1.2 w=2.3

10 1.0
f''(,0)

-'(,0)

8
0.8

6
0.6
w=1.5 Pr=1
4 w=1.7 Ri=5
Rd=30 0.4
w=2.0
2 w=2.3
0.2

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0



1116
 (a) (b)
Figure 7. Numerical values of local skin friction (a) and local heat transfer (b) against the streamwise distance ξ for different θw while Pr=1.0, Ri=5
and Rd=30.

CONCLUSIONS
In this article, we have studied numerically the effects mixed convection parameter Ri, radiation-conduction
parameter Rd and surface temperature ratio θw on a steady mixed convective flows about an inclined slender
cylinder. A transformed set of non-similar equations have been solved using the Keller box scheme. From the
present numerical investigation, the following conclusions can be drawn:
1. An increase in the mixed convection parameter Ri increases the local skin friction and the local heat
transfer parameters.
2. An increase in the radiation parameter Rd decreases the local skin friction parameter and increases the
local heat transfer parameter.
3. An increase in the surface temperature parameter θw increases the local skin friction parameter and
decreases the local heat transfer parameter.

NOMENCLATURE
cp specific heat of the convective fluid
Gr Grashof number
f dimensionless stream function
Pr Prandtl number
qr the component of radiative flux in r direction
Re Reynolds number
Ri Richardson number
Rd Planck number
T temperature
u, v velocities in x and r directions, respectively
x,r coordinates in axial and radial directions, respectively
Greek symbols
R Rosseland mean absorption coefficient
 similarity variable
 transverse curvature
 Stefan–Boltzmann constant
ρ fluid density
μ dynamic viscosity
υ kinematic viscosity
θ dimensionless temperature profile
θw surface temperature ratio
Subscripts
w wall
∞ free stream

REFERENCES

Ali, M.M., Chen, T.S. and Armaly, B.F. 1984. Natural convection–radiation interaction in boundary-layer flow over
horizontal surfaces. AIAA Journal 22: 1797–1803.

Anwar, I., Amin, N.and Pop I. 2008. Mixed convection boundary layer flow of a viscoelastic fluid over a horizontal
circular cylinder. International Journal of Non-Linear Mechanics 43: 814 – 821.

Aydin, O. and Kaya, A. 2008. Radiation effect on MHD mixed convection flow about a permeable vertical plate.
Heat and Mass Transfer 45: 239–246.

Cebeci, T., Bradshaw, P. 1977. Momentum Transfer in Boundary layers. Washington D.C.: Hemisphere.

Chang, C.L. 2006. Buoyancy and wall conduction effects on forced convection of micropolar fluid flow along a
vertical slender hollow circular cylinder. International Journal of Heat and Mass Transfer 49: 4932–4942.

1117
Chen, T.S. and Mucoglu, A. 1975. Buoyancy effects on forced convection along a vertical cylinder. ASME Journal
of Heat Transfer 97: 198–203.

Datta, P., Anilkumar, D., Roy, S. and Mahanti, N.C. 2006. Effect of non-uniform slot injection (suction) on a forced
flow over a slender cylinder. International Journal of Heat and Mass Transfer 49: 2366–2371.

Hossain, M.A. and Alim, M.A. 1997. Natural convection-radiation interaction on boundary layer flow along a thin
vertical cylinder. Heat and Mass Transfer 32: 515-520.

Hossain, M. A., Kutubuddin, M. and Pop I. 1999. Radiation-conduction interaction on mixed convection from a
horizontal circular cylinder. Heat and Mass Transfer 35: 307-314.

Hossain, M.A. and Pop, I. 1997. Radiation effect on Darcy free convection flow along an inclined surface placed in
porous media. Heat and Mass Transfer 32(4): 223–227.

Kumari, M. and Nath G. 2004. Mixed convection boundary layer flow over a thin vertical cylinder with localized
injection/suction and cooling/heating. International Journal of Heat and Mass Transfer 47: 969–976.

Lee, S.L., Chen, T.S. and Armaly B.F. 1987. Mixed convection along vertical cylinders and needles with uniform
surface heat flux. ASME- Journal of Heat Transfer 109: 711–716.

Molla, Md. M., Paul, S. C. and Hossain, Md. A. 2009. Natural convection flow from a horizontal circular cylinder
with uniform heat flux in presence of heat generation. Applied Mathematical Modeling 33: 3226–3236.

Mucoglu, A. and Chen, T.S. 1976. Buoyancy effects on forced convection along a vertical cylinder with uniform
surface heat flux. ASME Journal of Heat Transfer 98: 523–525.
Roy, S., Datta, P., Ravindran, R. and Momoniat E. 2007. Non-uniform double slot injection (suction) on a forced
flow over a slender cylinder. International Journal of Heat and Mass Transfer 50: 3190–3194.

Singh, P.J., Roy, S. and Pop, I. 2008. Unsteady mixed convection from a rotating vertical slender cylinder in an
axial flow. International Journal of Heat and Mass Transfer 51: 1423–1430.

Sparrow, E.M. and Cess, R.D. 1961. Free convection with blowing or suction. ASME - Journal of Heat Transfer 83:
387–396.

Takhar, H.S. and Nath G. 2000. Self-similar solution of the unsteady flow in the stagnation point region of a
rotating sphere with a magnetic field. Heat and Mass Transfer 36: 89-96.

1118
 CONCEPTUAL DESIGN OF SMALL WIND TURBINES WITH MORPHING BLADE

Thanapon Chuamvarasart*, Chawin Chantharasenawong, Anak Khantachawana

Department of Mechanical Engineering, Faculty of Engineering
King Mongkut’s University of Technology of Technology Thonburi, 126 Pracha-Uthit Rd., Bangmod, Thungkhru,
Bangkok, Thailand 10140
thanapon.ch@mail.kmutt.ac.th

ABSTRACT
The major disadvantage of the static blades in stall control wind turbines is the reduced power production
performance at high wind speeds beyond the design speed. There is another point to consider. The stall control
turbine may be designed to work efficiently at high wind speeds but these designs would suffer from the high cut-in
wind speeds penalty which results in the low power output at low to moderate winds. Hence, this research aims to
achieve this by introducing a morphing blade capability so that the stall control turbines can excel in their
traditionally weak areas. This work uses shape memory alloy (SMA) as actuators to enable wind turbine blade
morphing capability to improve the performance the turbine by extending its operation range. A morphing
mechanism prototype with shape memory alloy wire actuator is presented and verified that it is capable of its
intended functions under actual working conditions. An aerodynamic performance evaluation is presented to show
that the airfoils can improve the stall control performance.

INTRODUCTION
Horizontal axis wind turbines (HAWTs) can generally be classified into two types (1) pitch controlled and (2) stall
controlled (M. Ragheb., 2013). In simple terms, the pitch control system refers to the movable (in rotation) turbine
blades which enhance the turbine performance by continually optimizing the angle of attack between the blade
and the relative flow to maximize the rotor torque. The pitch control system is only found in large HAWTs, usually
with rated power production higher than 250kW. This is because of the size limitation of the pitch control actuation
system. Hydraulics or electric motors are usually used to provide torque for the blade longitudinal rotation. These
mechanical systems are large and cannot physically fit into the wind turbine rotor hub. Therefore, it is very
common that smaller HAWTs (usually under 250kW) only feature stall controlled rotors. Stall controlled HAWTs are
wind turbines with fixed blades. The blade twist and chord length variations have been optimized for operation at a
specific design condition, usually at the rated wind speed where it produces the maximum power. The stall control
wind turbines clearly have advantages over the pitch control turbines when it comes to construction simplicity due
to fewer moving parts. Maintenance costs are also much lower as a result. However, the major disadvantage of the
unmovable blades is the reduced power production performance at high wind speeds beyond the design speed.
This phenomenon is illustrated in Figure 1 where the pitch controlled turbine power output peaks at the designed
wind speed while the stall controlled wind turbines are able to produce the maximum power over a range of high
wind speeds.

Fig.1. Power curve comparison shows that the stall control turbine loses the opportunity to capture the high energy
during strong winds. These curves have been normalized by its respective maximum power rating because only
their characteristics are being compared. [LiTEN, Gamesa and Nordex]

The pitch controlled wind turbines with the rotatable blade mechanism are able to harvest the high power
production at high wind speeds, i.e. higher than the design wind speeds up to the cut-out wind speeds. There is
another point to consider. The stall control turbine may be designed to work efficiently at high wind speeds but
these designs would suffer from the high cut-in wind speeds penalty which results in the turbine not producing any
power at low to moderate winds. The high cut-in speed is a major problem for countries with generally low average

1119
wind speed such as Thailand. These result in a need for improving both post stall and below cut-in performances
of a stall control turbine. Hence morphing mechanism can be implemented to solve this problem.

Morphing structure is one of the methods to improve aerodynamics characteristics of small aircraft wings which
can be readily applied to wind turbine blade. There are currently several construction techniques. (Peel et al.,
2009) proposed the development of morphing wing using composite skin with pneumatic rubber muscle as
actuators. (Weaver et al., 2008) studied the possibility of using bi-stable laminated composite structures for
morphing an airfoil section. Other bi-stable structures have been suggested by (Bowen et al., 2008) whose
actuator design makes use of piezoelectric bi-stable carbon fibre laminate. (Thill et al., 2008) introduced several
types of morphing structures and skins.

The researchers are focusing on using shape memory alloy (SMA) wires as an actuator. The material exhibits a
Shape Memory Effect which enables it to return from a deformed shape to its original shape when its temperature
reaches the phase transformation temperature. This is usually significantly higher than the room temperature so its
Shape Memory Effect can be triggered using an electric current to generate the internal heat.

Initially several mechanism designs were considered and studied, such as revolver design and door-lock design,
as well as several actuator systems such as electric motors, hydraulics and spring mechanisms. It was eventually
concluded that many designs would be able to perform the task of morphing the airfoil shape efficiently. However,
due to the limited space inside the blade, many of these ideas had to be abandoned because of their inflexible
arrangements and large space requirement. Shape memory alloy is chosen because of two main reasons:
compact size and gapless blade surface. The entire actuation unit is compact when compared to traditional
actuator mechanisms such as electric motors, hydraulics and pneumatics as described by (Hartl et al., 2007) who
suggests that a single SMA component can replace a complicated electromechanical or hydraulic actuator. SMAs
enable gapless blade surface during morphing. This is a very attractive feature that cannot be achieved through
traditional aircraft wing flapping mechanism designs.

This work aims to propose a conceptual design of small HAWT blades with the use smart materials as actuators to
enable wind turbine blade morphing capability to improve the performance the turbine by extending its operation
range. The focuses are (1) wind turbine power production improvement and (2) actuator designs capable of
working under actual operating conditions.

METHODOLOGY

Airfoil Design and XFOIL testing

This work is based on the Clark Y airfoil (Lee et al., 2012) suggested that this airfoil is suitable for use on a small
HAWT blades because of its ‘flat’ lower surface which makes it easy to install the morphing mechanism inside.
The airfoil has a maximum thickness of 11.7% and a 3.5% camber.

The airfoil trailing edge will deflect downwards in the morphed state. This is to generate an extra lift with a drag
penalty. It is expected that the extra lift generation is beneficial to the wind turbine power generation performance by
considering its lift coefficient and higher lift-to-drag ratio, given that the morphed state is triggered at appropriate
operating conditions. The blade with morphed airfoil generates higher lift, resulting in an increase in rotor torque under
the same free stream velocity, hence an increase in power production. [M. Ragheb. 2013] The optimization for
morphing actuation is, however, not covered in this work and the next section will only focus on the airfoil
aerodynamic characteristics calculation by XFOIL [Hao Xuan et al., 2008].

The base airfoil and an example of a morphed airfoil with a 5-degree trailing edge flap are shown in Figures 2a and
2b, respectively. Calculations will be made using operating condition parameters replicating an actual environment of
a generic 1-kW HAWT at rated wind speed of 8 m/s.

Fig.2a. Illustration of the base airfoil geometry (Clark Y) with 11.7% maximum thickness and 3.5% camber

1120
Fig.2b. Illustration of the morphed airfoil geometry (Clark Y) with 5-degree trailing edge flap downward deflection

Morphing Mechanism Design

A prototype wind turbine blade section with 30cm chord length and 40cm span has been fabricated as shown in
Figure 3. Its skin is made from 1mm thick flexible clear PVC sheet and the spar and rib are constructed with 5mm
thick clear acrylic using a combination of fasteners and adhesives. The leading edge and trailing edge are covered
with a layer of green flexible acrylic adhesive to avoid cracks due to repeated movement. A steel rod is installed at
the blade section quarter-chord position to provide a mounting point when the unit is tested in the wind tunnel.

The gapless morph that this prototype is attempting is to deflect its trailing edge downward. The central part of the
blade section houses the morphing actuation unit. A full description of the SMA actuator design is given in the
following section. The trailing edge hinge is located at 80% chord on the internal lower surface. There are
overlapping skins on the upper surface to make it a gapless unit when the trailing edge flap has deflected. See
Figure 4a and 4b for illustration.

Fig.3. A photograph of a 40cm span morphing wind turbine blade section

The actuators are powered by SMA, which requires electricity to initiate movements. This is powered by a
standard external 12V DC power supply. The electricity input is controlled manually at this stage. The power
supply unit is not shown in the figure.

Fig.4a. Illustration of blade section internal construction in the base state

1121
Fig.4b. Illustration of blade section internal construction in the morphed state (downward trailing edge flap
deflection)

Morphing airfoil mechanism

This section describes the flap actuation mechanism, which consists of three main steps namely (1) flap actuator
initiation (2) lock (3) return to original state.
1. Flap actuator initiation
The power supply is manually switched on and an electric current passes through the SMA coils, which contract
due to its original shape being in a tightly wound coil state. The tension then creates a moment about the trailing
edge flap hinge resulting it to deflect downward.
Table 1. Properties of SMA coils and spring in the actuator system
Bias
SMA wire SMA coil Stiffness Stiffness No. of
No. of Number spring
diameter diameter at 75c at 25c bias
coils of units stiffness
springs
(mm) (mm) (N/m) (N/m) (N/m)

Flap 0.85 8 8 131.39 35.749 2 81.75 2

actuation 0.85 6 8 200.93 66.8 4 189.87 2
Returning
trailing
edge flap 0.85 8 8 131.39 35.749 2 73.76 1
to original
position

Fig.5. Top and side view of mechanism without flap.

1122
Fig.6. Top and side view of mechanism with flap.

2. Lock
When the trailing edge deflection has reached the desired angle, a steel rod will slide into the lock position under
the influence of a locking mechanism bias spring. This feature is created to eliminate the trailing edge flap
deflection angle inaccuracy caused by the SMA hysteresis characteristic. [K.Yamauchi et al., 2011]

Fig.7.Before and after the locking mechanism actuation to bring the trailing edge flap deflection down
Figures 7 shows the locking mechanism before the trailing edge flap deflection actuation takes place, the force
from the spring pushes the rods to touch the surface of the slider stick in slot. The stick then slides to the right
hand side until get the position that attaches the rod, the rod is pushed into contact with the surface of the slot
groove. On the other side when stop input power and leave SMA to cool down, slider stick will move back to the
left but at the end of the shaft will be earmarked by the rod stick to lock it at that position for the desired flap angle.

1123
3. Returning to the original state
A pair of SMA coils is triggered by electricity to pull the rod back. Once the rod is out of the grooved slot, the series
of bias springs will pull the trailing edge back to the original position.

Fig.8.Before and after the locking mechanism actuation to bring the trailing edge flap deflection up
Figures 8 shows the mechanism while unlocking with input power to SMA of unlocks mechanism, which will result
in a pull rod stick back and the slider stick moves back to the left which make the mechanism back to its original
position that did not change the shape of the airfoil.

Morphing Mechanism Test

1124
Fig.9. the entire blade section is mounted in an open-jet wind tunnel to test its morphing mechanism.
Figure 9 Show the entire blade section is mounted in an open-jet wind tunnel to test its morphing mechanism. The
wind tunnel speed is set at 15 m/s to replicate the actual relative wind speed that the airfoil is likely to encounter if
it was used on a 1-kW HAWT. A protractor has been installed to measure the trailing edge deflection angle and the
time taken for the deflection to complete is also recorded.

RESULTS AND DISCUSSION

XFOIL testing results

The airfoil lift coefficient, CL is to be considered because it directly affects the total rotor torque. Another point that
should be considered is the value of the drag coefficient, CD must be kept relatively low at desired lift generating
angle of attack. Therefore, a high value of the lift-to-drag ratio, CL/CD is desirable. With an equal emphasis on both
lift generation and the lift-to-drag ratio the value of CL2/CD will be used as an airfoil performance indicator. Fig.10.
compares CL2/CD between conventional airfoil ( = 0) and the morphed airfoil ( = 5) which were tested at Re =
300000.The result shows that morphed airfoil with a trailing edge downward flap of 5 have higher CL2/CD than
conventional airfoil. Therefore the wind turbine blade with morphed airfoil generates higher lift, resulting in an
increase in rotor torque under the same free stream velocity, hence an increase in power production

Fig.10. the comparison of CL2/CD between conventional airfoil ( = 0) and morphing airfoil ( = 5).

1125
Precision testing of morphing airfoil

Fig.11.The actuator response time frequency distributions in both trailing edge deflection directions obtained from 10
actuation cycles

It is found that the morphing mechanism successfully worked. It is able to generate sufficient moment around the
trailing edge hinge so that the trailing edge flap is deflected to a desired angle with insignificant inaccuracy of 5%. The
time taken for the actuator to complete the deflection ranges between 15-17 seconds, while the time take for the
system to return to the original state requires only 6-7 seconds. These results are obtained from an actual experiment
which was repeated over 10 times.

CONCLUSIONS
This present work involved the concepts of morphing blade design by using an SMA wire as an actuator install at
internal structure. A prototype was made and tested using a series of assignments to determine its actuator system
accuracy.

Acknowledgement
This research was supported by The Asahi Glass Foundation (AGF).

NOMENCLATURE
Re Reynolds number
CL Lift coefficient
CD Drag coefficient
CL/ CD Lift to drag coefficient
 Angle of attack
 Trailing edge flap angle

REFERENCES
Woo-Ram Kang, Eun-Ho Kim, Min-Soo Jeong and In Lee. 2012. Morphing Wing Mechanism Using an SMA Wire
Actuator. Int’l J. of Aeronautical & Space Sci 58–63.

C. Thill, J. Etches, I. Bond, K. Potter and P. Weaver. 2008. Morphing skins. THE AERONAUTICAL JOURNAL.

Cezar G. Diaconu, Paul M. Weaver, Filippo Mattioni. 2008. Concepts for morphing airfoil sections using bi-stable
laminated composite structures. Thin-Walled Structures 46 689–701.

Hartl, DJ & Lagoudas, DC. 2007. Proceedings of the Institution of Mechanical Engineers. Journal of Aerospace
Engineering 221:535-552.

P. Giddings, C.R. Bowen, R. Butler, H.A. Kim. 2008. Characterisation of actuation properties of piezoelectric bi-
stable carbon-fibre laminates. Composites: 697–703.

Larry D. Peel, James Mejia, Ben Narvaez, Kyle Thompson, Madhuri Lingala. 2009. Development of a Simple
Morphing Wing Using Elastomeric Composites as Skins and Actuators. Journal of Mechanical Design 131:1-8.

1126
K.Yamauchi, I.Ohkata, K.Tsuchiya and S.Miyazaki. Shape Memory and Superelastic Alloys: Technologies and
Applications (Series in Metals and Surface Engineering). Woodhead Publishing

Hao Xuan, Zhang Weimin, Liu Xiao and Liu Jieping. 2008. Aerodynamic and Aeroacoustic Optimization of
WindTurbine Blade by Genetic Algorithm. 46th AIAA Aerospace Sciences Meeting and Exhibit AIAA 2008-1331.

M. Ragheb. 2013. Lecture Note of NPRE 478 Wind Power Systems. Department of Nuclear, Plasma and
Radiological Engineering University of Illinois at Urbana-Champaign.

Nordex S77 1.5 mw Specifications Booklet (2013)

Nordex N100 2.5 mw Specifications Booklet (2013)

Gamesa G90-2.0MW Specifications Booklet (2013)

LiTEN Specifications Booklet (2013)

1127
 EVALUATION OF THERMAL CHARACTERISTICS OF A BOREHOLE THERMAL ENERGY
STORAGE SYSTEM
a b
Önder Kizilkan , Ibrahim Dincer
a
Department of Energy Systems Engineering, Faculty of Technology, Süleyman Demirel University, Isparta, Turkey
b
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology (UOIT), Oshawa, ON, Canada
onderkizilkan@sdu.edu.tr, Ibrahim.Dincer@uoit.ca

ABSTRACT
Energy storage technologies are usually a strategic and necessary component for the efficient utilization of renewable
energy sources and energy conservation. Their use is important to overcome energetic and environmental issues.
Thermal energy storage (TES) serves at least three different purposes: (i) energy conservation and substitution (by
using natural energy sources and waste energy), (ii) energy peak shifting (from more expensive daytime to less
expensive nighttime rates), and (iii) electricity conservation (by operating efficient devices at full load instead of part
load to reduce peak power demands and increase efficiency of electricity use). Numerous TES systems have
received attention for practical applications. Underground thermal energy storage systems may be divided into two
groups: (i) closed storage systems, so-called: borehole TES, in which a heat transport fluid (water in most cases) is
pumped through heat exchangers in the ground, and (ii) open systems where groundwater is pumped out of the
ground and then injected into the ground using wells (aquifer TES) or in underground caverns. In this study, the
thermal resistance characteristics of borehole heat exchangers of borehole TES at University of Ontario Institute of
Technology (UOIT) are investigated. In this regard, the ground thermal properties of Oshawa are studied, and a
comparison of different methods for determining such thermal resistances is made for practical applications. Finally, a
thermal response test mode for BTES is illustrated.

Keywords: Thermal Resistance, Thermal Response Test, Ground Source Heat Pump, Borehole Thermal Energy
Storage.

INTRODUCTION
Energy savings is a major factor to be taken into consideration, because of the rapid decrease in the fossil fuel
sources. For energy saving and reducing the emissions, it is required to improve the energy efficiency of systems.
For this aim, the renewable energy technology is in continuous development. Heat pumps for space heating and
cooling are an interesting example. To improve their performances, low temperature in heating and high
temperature in cooling operation is required. On the other hand, a suitable source or sink temperature is
necessary. In general, external air is used but in this case the energy performance is affected by climatic
variations. Instead of this, if available and allowed, ground water use becomes a good option under constant
temperature conditions. Another possibility is to use the ground as a heat source or sink, since its temperature is
almost constant all over the year, thus allowing smaller temperature variations than external air during both heating
and cooling seasons (De Carli et al., 2010).

Borehole thermal energy storage systems (BTES) systems use the ground as a heat source or sink for space
conditioning in residential and commercial buildings. In last decades, ground source heat pump (GSHP) systems
have been used increasingly around the world, because they are among the cleanest and most energy efficient
air-conditioning systems for commercial and residential buildings. A typical GSHP system mainly consists of a heat
pump unit with a BTES. BTES systems are coupled with a group of borehole heat exchangers (BHE). BHEs are
devised for the extraction or injection of thermal energy from or into the ground. However, the thermal performance
of BHEs is site-specific, due to the differences of the underground thermal properties, sizes and configuration of
BHEs, and backfill materials of boreholes (Wang et al., 2010). The structure and physical properties of the ground
are factors affecting temperature, in all zones. The temperature of the ground is a function of the thermal
conductivity, geothermal gradient, water content and water flow rate through the borehole (Florides et al., 2011).

In a vertical U-tube ground heat exchanger a water pump circulates fluid through pipes inserted into a borehole in
the ground. The borehole, after the insertion of the U-tube is usually backfilled with grout or water in order to
ensure good thermal contact with the ground (Florides and Kalogirou, 2008). In Scandinavia groundwater is often
used to fill the space between borehole wall and collector wall, while otherwise it is more common to backfill with
some grouting material. The advantage of using water is cheaper installations and more easy access to the
collector if needed. Grouting is on the other hand required in many counties by national legislation in order to
prevent groundwater contamination or is used to stabilize the borehole wall (Gustafsson et al., 2010). The
circulating fluid is usually water or a water-antifreeze mixture. A borehole heat exchanger is usually drilled to a
depth between 20 and 300 m with a diameter of 10–15 cm. A borehole system can be composed of a large
number of individual boreholes (Florides and Kalogirou, 2008).

1128
In this study, a closed BTES using BHEs at University of Ontario Institute of Technology (UOIT) in Oshawa,
Canada, is analyzed in terms of ground thermal characteristics. The underground BTES system considered in this
article is installed at UOIT. The UOIT campus includes four new buildings that are designed to be heated and
cooled, sometimes with renewable energy in order to minimize greenhouse gas emissions. Test drilling programs
were carried out to determine the feasibility of thermal storage in the overburden and bedrock formations at the
UOIT site. In-situ tests were conducted to determine the groundwater and thermal characteristics. An almost
impermeable limestone formation was encountered from 55 m to 200 m below surface. For meeting the required
energy demand, 370 boreholes, each 200 m in depth were used. In addition, five temperature monitoring
boreholes were installed, increasing the total drilling for the project to 75 km. Some design changes were made to
the borehole heat exchangers (BHE) as a result of the lack of groundwater flow in the rock. The Swedish practice
of water-filled BHEs was utilized instead of the North American practice of grouted BHEs. Waterfilled BHEs
improve the efficiency of the U-tube installation and extend the life of the boreholes indefinitely (Further details,
see Dincer and Rosen, 2007).

GROUND THERMAL CHARACTERISTICS OF OSHAWA

For determination of the thermal interaction of BHE with the ground, it is important to have an accurate
representation of local ground temperature since it is the temperature difference between the ground and the
circulating fluid that drives the heat transfer. The ground temperature near the surface cycles with the time of the
year. These variations disappear at lower depths, where the ground remains at its mean temperature throughout
the year. An equation for calculating the ground temperature at any time of the year was developed by Kasuda
who found that the temperature of the ground is a function of the time of year and the depth below the surface
(Keen Engineering, 2003). The equation is used for predicting soil temperatures, T (z, t), for any depth z and at
any time t in the year, based on the mean ground temperature (Tmean), the surface amplitude (Tamp), and the time
of the minimum surface temperature (t0) (Kasuda and Archenbach, 1965). The vertical ground temperature
distribution can be theoretically modeled based on the following equation:

 π   2π  z 365  
T (z ,t ) = ( Tmean − Tamp ) − exp − z cos t − t 0 −  (1)
 
365α   365  2 πα  
  
where Tmean is the mean soil temperature, Tamp the annual surface temperature amplitude, t the "Julian day", where
t = 0 refers to January 1st and t0 the day of minimal soil surface temperature, and α the thermal diffusivity of the soil
(in m2/day). The annual surface temperature amplitude Tamp can be found using equation below (Jenkins et al.,
2009):
Tsurf ,max − Tsurf ,min
Tamp = (2)
2
The required ground measurement data are the mean or average ground temperatures at approximately 3 m or as
deep as is available. The data at which Tsurf,min occurs (t0) is also available. To calculate the value for t0, it is the
best way counting the number of days since January 1st, including January 1st and the day on which t0 occurs. For
groundwater systems, the entering water temperature can be approximated by the mean ground temperature.
Table 1 contains the ground property data for 15 Canadian locations (Keen Engineering, 2003).

Table 1. Earth Temperature data for various Canadian locations (Keen Engineering, 2003)
Location Tmean (°C) Tamp (°C) t0 (days)
Calgary (AB) 6.4 11.1 22
Edmonton (AB) 7.1 14.9 11
Vancouver (BC) 11.3 12.2 7
Victoria (BC) 11.8 12.2 5
Winnipeg (MB) 6.1 14.0 36
Halifax (NS) 9.0 10.2 50
St. John’s (NL) 6.7 10.0 50
Fredericton (NB) 7.7 11.6 38
Charlottetown (PE) 7.5 15.2 44
Montreal (QC) 6.4 15.1 14
Quebec City (QC) 7.4 13.5 52
Ottowa (ON) 8.9 12.6 36
Toronto (ON) 10.1 12.5 41
Oshawa* (ON) 10.3 12.9 44
Regina (SK) 4.8 15.9 36
Saskatoon (SK) 5.9 17.2 12
* Data for Oshawa was determined using values obtained from Environment Canada (Weather Canada, 2011)

1129
The thermal conductivity of the soil and the thermal diffusivity are the two soil and rock properties that most affect the
design of the heat pump system. The thermal conductivity represents the ability of heat to travel through the ground.
The thermal diffusivity is am measure of the ground’s ability to conduct thermal energy relative to its ability to store
thermal energy. Table 2 provides a simplified subdivision of rock and soil properties accompanied by approximate
thermal and physical properties (Keen Engineering, 2003).

Table 2. Thermal and physical properties of some types of rock and soil (Keen Engineering, 2003)
Layer Material k α ρ Cp
2 3
W/mK m /day kg/m kJ/kgK
Dense rock (Granite) 3.46 0.1114 3204.0 0.8374
Average rock (Limestone) 2.42 0.0891 2803.5 0.8374
Heavy soil – damp (Clay. compacted sand) 1.30 0.0557 2098.6 0.9630
Heavy soil – dry (Clay. compacted sand) 0.87 0.0445 2002.5 0.8374
Light soil – damp (Loose sand. silt) 0.87 0.0445 1602.0 1.0468
Light soil – dry (Loose sand. silt) 0.35 0.0245 1441.8 0.8374

CALCULATION OF THERMAL RESISTANCES

Thermal Resistance inside the Borehole
Thermal resistance of fluid convection is defined as:
1
Rconv = (3)
2πrin hin
where subscript in represents inner flow. The convection coefficient hin, for inner flow is determined by using Nusselt
number:

hin d in
Nu = (4)
k fluid

For a fully developed turbulent flow in smooth tubes, a simple relation for the Nusselt number is given by Gnielinski
(Cengel, 2002):

Nu =
( f / 8)(Re− 1000) Pr (0.5 ≤ Pr ≤ 2000, 3 × 103 < Re < 5 × 10 6 ) (5)
 2 
1 + 12.7 f / 8  Pr 3 − 1
 
 

where ƒ is the friction factor and can be determined for smooth tubes, in turbulent flow from the first Petukhov
equation (Cengel, 2002):

f = (0.790 ln Re− 1.64)−2 (10 4

< Re < 10 6 ) (6)

The fluid properties are evaluated at the bulk mean fluid temperature T=(Tin+Tout)/2. Thermal resistance of conduction
in the pipe then becomes

 rp ,out 
ln 
 rp ,in 
Rcond =   (7)
2πk p

where rp,in and rp,out are the inner and outer radius of the pipe and kp is thermal conductivity of pipe. Since in most of
the GSHP systems, the boreholes are usually grouted, a few of them left with groundwater to fill the space between
heat exchanger pipes and borehole wall. Few models concern the convective heat transfer in the groundwater. Also it
was reported by Gustafsson et al. (2010), that most of them are dealing with regional groundwater flow, which is
shown not to influence the system except in porous ground or high fractured bedrock. None of these models concern
natural convection in groundwater filled borehole situated in solid rock without any larger fractures, which is a
common situation in Sweden. In Scandinavia groundwater is often used to fill the space between borehole wall and
collector wall, while otherwise it is more common to backfill with some grouting material (Gustafsson et al., 2010). Like
in most GSHPs in Sweden, the system at UOIT was also built using water as filling material.

1130
For UOIT, the background temperature of the geologic formations at the site is 10°C. Considering the ground
temperature, the poor quality shale groundwater and the lack of ameasurable groundwater flux in the limestone, it was
concluded that the site was ideally suited for a borehole thermal energy storage (BTES) system. No water-bearing
sand deposits were found in the 44 m of deposits (silt/clay overburden) (Beatty and Thompson, 2004).

Calculation of Thermal Resistances of Filling Material for Grouts

The three effective thermal resistances, Rconv, Rcond and Rgrout are, respectively, the convective resistance inside each
tube, the conduction resistance for each tube and the grout resistance. They are obtained using the analytical
equation proposed by Hellström (1991). Thermal resistance of the grout has the following expression (Philippe et al.,
2010):
  
  
1   rbh   rbh  k grout − k soil  rbh 4 
R grout = ln + ln +
 ln  (8)
4πk grout   r p ,out   Ls  k grout + k soil  4  Ls  
4

  rbh −  2  
    

where subscript bh represents borehole, kgrout is the thermal conductivity of the grout, ksoil is the ground thermal
conductivity and Ls is the center-to-center distance between the two pipes (Fig. 1).

Fig. 1. Cross-section of geothermal vertical borehole (Philippe et al., 2010)

Thermal resistance of grout can be computed by shape factor method and equivalent diameter method. The shape
factor method is used to describe the heat conduction characteristics of a complicated geometry (Jun et al., 2009).
1
R grout = β
(9)
 d bh  1
k grout β 0  

 d out 

where β0 and β1 are the shape factors of Rgrout, whose values depend on the relative location of U-tube pipes in the
borehole. Remund (1999) studied three configurations as shown in Fig. 2, the corresponding values are provided in
Table 3. To be mentioned, configuration B would be an appropriate design assumption in most situations, as it
represents an average spacing along the entire borehole length. While configuration A is a conservative design
assumption, configuration C is a risky design assumption (Remund, 1999).

Fig. 2. Configrations of U-tube pipes in borehole (Modified from Remund, 1999)

Table 3. Values for β0 and β1 for three configurations

Configuration β0 β1
A 20.10 -0.9447
B 17.44 -0.6052
C 21.91 -0.3796

1131
Equivalent diameter method means the two legs in the U-tube are replaced by a single concentric cylindrical heat
source (or sink), the equivalent diameter given by Claesson and Dunand (1983) is given as follows:

d e = 2 d out (10)

The equivalent diameter given by Gu and O’Neal is as follows (Jun et al., 2009):

d e = 2d out Ls (d out < Ls < d bh ) (11)

When equivalent diameter method is used for computing the thermal resistance inside the borehole, thermal
resistance of fluid and pipe should remain constant, the mass flow rate and heat capacity of the fluid should also be
constant. Application of this method is especially advantageous when an analytical solution developed for a single
cylindrical source is available (Jun et al., 2009).

1 d 
R grout = ln bh  (12)
2πk grout  d e 

Another commonly used relation for Rgrout is the multipole model of Bennet et al., which is an extension of the relation
presented by Hellström, and is given by the following expression (Lamarche et al., 2010; Raymond et al., 2011):
 
 
 
 
 2 
 2
 4σ  

R grout =
1
 
ln x1 x 2
1+ 4σ 

−
( x 3 
)

1 −
x 2
4
− 1 



( )
(13)
2πk grout   2 x 4 − 1 σ   
  2   
  
 2 16σ 
 1 + x 3  1 + 2 
  2 
 x − 1  
   
x 2   
2 2
  
where
r
x1 = bh (14)
r p ,out
2rbh
x2 = (15)
Ls
x2
x3 = (16)
2x1
k grout − k soil
σ= (17)
k grout + k soil

Using a numerical model presented Sharqawy et al. (2009) the following alternate definition for Rgrout is obtained:
 - 1.49 
+ 0.656 ln(x1 ) + 0.436 
1
R grout =  (18)
2πk grout  x 2 

Other definitions for Rgrout also include as proposed by Shonder and Beck (2000):
1  x 
R grout = ln 1  (19)
2πk grout  2 

and that of Gu and O’Neal (1998a).

R grout =
1
2πk grout
(
ln x1 x 2 ) (20)

1132
Calculation of Thermal Resistances of Filling Material for Water
For determining the thermal resistnce of filling material for water, we use the following equation (Gustafsson et al.,
2010).

Nu = 0.291 (Ra * )0.244 A −0.238 R 0.442 (1.8 × 10 4

≤ Ra * ≤ 4.21 × 10 7 ) (21)

where Ra* is modified Rayleigh number, A is the aspect ratio, and R is the radius ratio as given below (Gustafsson et
al., 2010):

Ra* = g β bw q" p ,oou

(rbh − rp ,out )4 (22)
k bwν bwα bw
L
A= (23)
rbh − r p ,out
rbh
R= (24)
r p ,out

Here, g is gravitational acceleration, β is thermal expansion coefficient, q” is heat flux, and subscript bw is borehole
water. Thermal expansion coefficient is defined as

(ρ − ρ ref )g ≈ − ρ ref β (T − Tref )g (25)

Note that Littlefield and Desai suggest the following relationship of Nusselt’s number for isothermal boundary
condition (Gustafsson et al., 2010):

Nu = 0.443(Ra * )(0.233−0.009 R ) A −0.245 R 0.44 (1.5 ≤ R ≤ 5 10 ≤ A ≤ 50 N hf ≤ 0.5 Ra* ≤ 10 8 ) (26)

In Eq. 26, Nhf is heat flux number, q"bw/q"p,out.

Thermal Resistance outside the Borehole

Predicting the thermal behavior inside and in the vicinity of boreholes is important to establish the required borehole
length and to determine the resulting fluid temperature. In the literature one can find several methods to calculate heat
flow in the ground around vertical boreholes. These methods can be divided in three distinct categories: analytical,
numerical and semi-analytical models. The main advantage of analytical models is that they are simple to implement
and they lead to relatively fast solutions. However they are restricted to pure conduction cases. For instance, in the
presence of groundwater flow, numerical methods are necessary. A number of studies on the analytical models
presented here can be found in the literature, and are reviewed briefly in the following paragraphs (Philippe et al.,
2010).

Kelvin’s Line Source Model

Ingersoll and Plass (1948) first applied the Kelvin line source theory to buried pipe heat transfer problems. The theory
assumes that heat is obtained from an infinitely long permanent line source of heat or a sink in an infinite medium
(soil) at an initial uniform temperature T0 (Jun et al., 2009).
∞ −β 2
q q bh
I (X )
e
∆T = Tbh − T0 = bh ∫ dβ = (27)
2πk soil X β 2πk soil
Tbh − T0
I (X )
1
R soil = = (28)
q bh 2πk soil
where;
r r
X = = (29)
2 αt 2 Fo

When 0 <X ≤ 1, the I-function is calculated by

(
I(X) = 0.5 - lnX 2 - 0.57721566 + 0.99999193 X 2 - 0.24991055 X 4 + 0.05519968 X 6 - 0.00976004 X 8 + 0.00107857 X 10 (30) )
and when x≥1,

1133
 1 A
I(X) =
2X 2 exp X 2 ( ) B
(31)

where
A = X 8 + 8.5733287X 6 + 18.059017X 4 + 8.637609X 2 + 0.2677737 (32)
8 6 4 2
B = X + 9.5733223X + 25.6329561X + 21.0996531X + 3.9684969 (33)

Finite Line-Source Model

When using finite line theory the heat conduction outside the borehole is treated as a finite line source in a semi-
infinite medium, which has a constant initial and boundary temperature for field and upper surface. In the finite line
source model, end effects are not negligible and the real borehole wall temperature varies along the borehole depth,
especially near the end of borehole region (Pesl, et al., 2007). The analytical solution of the finite line source model
was derived by Eskilson (1987) and Zeng et al. (2002) respectively. The temperature is obtained by integrating
contributions of imaginary point source distributed along the borehole (Eskilson 1987; Xu, 2007).

D+H
qbh 1  r+  1  r 
Tbh (r, z,t) - T0 =
4πk soil ∫  erfc
 r+
 − erfc − dh
 4αt  r−  4αt 
(34)
D

where H stands for depth of borehole, and the other parameters are defined as

r+ = r 2 + (z − h )2 and r− = r 2 + (z + h )2 (35)

Substituting Eq. (35) into Eq. (34) yields

  2   r 2 + (z + h )2 
 erfc r + (z − h )
2
  
  erfc   
H 2αt 2αt
qbh   −   dh
Tbh (r, z,t) - T0 =
4πk soil ∫   (36)
0 r 2 + (z − h )2 r 2 + (z + h )2 
 
 
 

Infinite Cylindrical Heat Source Model (Kavanaugh’s Approach)

The cylindrical source model assumes the single borehole as a single infinite isolated pipe surrounded by an
infinite solid with constant properties. With a constant heat flux at the pipe wall, the temperature of point under
consideration at or near the pipe is calculated as follows (Philippe, et al., 2010);

qbh αt
G (z , p ) z = 2
r
∆T = Tr - T0 = p= (37)
k soil r rbh

where G (z,p) is commonly referred to as the cylinder source integral or function, whose values was tabulated by
Ingersoll et al. (1954), or can be computed by equation given below (Chiasson, 2007; Deng, 2004):

e − βFo − 1
[J 0 ( pβ )Y1 (β ) − J1 (β )Y0 ( pβ )]
J12 (β ) + Y12 (β )
∞
 r  1
G Fo( z ), ( p )  = 2
 rbh  π
∫ β2
dβ (38)
0

By solving this equation, G-factor at the borehole wall as a function of Fo, for 10-1<Fo<106 is given in Table 4.

Table 4. Several expressions of G-Function (Jun et al., 2009)

P G (z,p)
1
z1 2 3
G=10 z1=-0.89129+0.36081Log(z)-0.05508Log (z)+0.00359617Log (z)
2
z2 2 3
G=10 z2=-1.4541+0.89933Log(z)-0.31193 Log (z)+〖0.061119 Log (z)
5
z3 2 3
G=10 z3=-3.0077+2.25606Log(z)-0.7928Log (z)+0.134293Log (z)
10
z4 2 3
G=10 z4=-9.1418+11.7025Log(z)-7.09574Log (z)+0.269837Log (z)

1134
Eskilson’s Model
The borehole wall temperature Tbh varies along the borehole depth, especially near the end; therefore, it is more
reasonable to use the integral mean temperature of the borehole wall to represent the heat transfer between the
borehole and the ground as explained by Eskilson in 1987. The borehole wall temperature for a single borehole
can be calculated by (Xu, 2007):

q l  t r 
Tbh - T∞ = - g  , bh  (39)
2πk soil  t s H 
  H  1  t  5rbh 2
ln  + ln  < t < ts
 t rbh    2rbh  2  t s 
 α H2
g  , ≈
 ts = (40)
 ts H    H  9α
ln 2r  t > ts
  bh 

THERMAL RESPONSE TEST

Soil thermal properties, i.e. thermal conductivity and borehole thermal resistance, represent the main design data
in thermotechnical applications that exploit the ground heating storage. When the heat storage system is vertically
oriented, that is the case considered here, the common method to estimate the soil properties is named Thermal
Response Test. An essential role in the estimation procedure is played by the choice of the reference model
adopted to recover the above-stated unknown quantities; this procedure is based on the comparison of the
experimental data with the solution of the equations describing the model’s behavior (Sanner, 2003).

A thermal conductivity test (Fig. 3) consists of a fully drilled borehole fitted with a u-tube and backfilled. This is
then connected to a defined artificial load and the resultant temperature fluctuations of the circulating fluid are
measured. Data on the energy use of the pump and inlet/outlet water temperature and flow rate can then be used
to determine the thermal properties of the soil. It should be noted though that while on-site tests are accurate they
are also expensive to conduct (Le Feuvre, 2007).

Fig. 3. Schematic of a thermal response test (Sanner, 2003)

The equation for the temperature field as a function of time (t) and radius (r) around a line source with constant
heat injection rate may be used as an approximation of the heat injection from a BHE (Eklöf and Gehlin, 1996):
Q  4α t  Q
Tbh = ln 2  − γ  + Rbh + T0 (41)
4πk soil H   r   4πk soil H

which can be re-written in a linear from as

Tbh = k ln(t ) + m (42)

with
Q
k= (43)
4πk soil H

1135
Therefore, it can be determined from the slope of the line resulting when plotting Tb against ln(t), hence the name
and basis of the evaluation method (Georgiev and Busso, 2006).

RESULTS AND DISCUSSION

For determining the daily soil temperatures at different depths of Oshawa, Eq. (1) is used. According to the data
given in Table 1, the time shift is 44 days, the mean surface temperature is 10.3 °C, the amplitude of the surface
temperature is 12.9 °C are taken for Oshawa, ON. The result is given in Fig. 4. The mean surface temperature
which by definition should correspond to the temperature of the ground at an infinite depth shows a discrepancy.

Fig. 4. Predicted daily soil temperatures at different soil depths for Oshawa.

In BTES systems, shank spacing is an important parameter for placing the BHEs. Assuming the filling material is
grout, thermal resistances are calculated using equations given in previous sections. For a comparative study, the
variation of thermal resistance of grout with shank spacing is given in Fig. 5.

Fig. 5. Thermal resistance of grout calculated using different equations. R1: Hellström (1991), R2: Remund (1999), R3: Claesson and Dunand
(1983), R4: Gu and O’Neal (1998a), R5: Bennet et al. (1987), R6: Sharqawy et al. (2009), R7: Shonder and Beck (2000), R8: Gu and O’Neal
(1998b)

Fig. 6 illustrates the radial temperature profile for different time periods and a given set of parameters for BTES
system in UOIT in Oshawa. These curves were obtained by evaluating the Kelvin’s Line Source Model integral
numerically.

For the Finite Line Source the integral in Eq. (36) is numerically evaluated using a recursive adaptive Simpson
quadrature for the parameters for the BTES system in UOIT. The results are presented in Fig. 7 where the
temperature profile has been plotted after 6 months of operation for four radial distances from the center of the
borehole.

1136
Fig. 6. Temperature distribution in the ground determined by Kelvin’s Line Source Model

Fig. 7. Temperature distribution in the ground determined by Finite Line Source at t=6 months

In order to calculate the thermal resistance outside the borehole, four models are used: Kelvin’s Line Source Model,
Eskilson’s Model, Infinite Cylindrical Heat Source Model (Kavanaugh’s Approach) and Finite Line-Source Model. For
comparison of 4 models, the parameters of GSHP system in for UOIT were used and the results are show in Fig. 8 As
seen from the figure, all fours models are very similar.

Fig. 8. Thermal resistance outside the borehole calculated by for different models. (R1: Kelvin’s Line Source Model, R2: Eskilson’s Model, R3: Infinite
Cylindrical Heat Source Model (Kavanaugh’s Approach), R4: Finite Line-Source Model)

The thermal response test (TRT) for BTES system founded in UOIT were made by Beatty and Thompson (2004) who
were the construction executives during the building process. TRT test results given here are calculated using the
data obtained from reference. According to their data, water was circulated through the closed loop at a rate of 0.75
L/s, with a constant heat injection of 3.2 kW. The heat injection test was conducted for 72 hours, followed by a 72-
hour recovery period. The power input, flow rate, and the loop inlet and outlet temperatures were recorded during the
heat injection and recovery period. The rate of heat injection is 3000 Watts and the length of borehole is 190 m for the

1137
TRT. Fig. 9 shows the resulting plot for the experimental data gathered from Beatty and Thompson (2004) during the
test.

Fig. 9. The results of TRT (Data obtained from Beatty and Thompson (2004))

The thermal conductivity rate was determined from the relationship between the heat rate, the test hole length and the
slope of the mean temperature data vs. the natural log of time. Using Eq. (41) we obtain

Q
Tbh = k ln(t ) + m = n(t ) + m (44)
4πk soil H

and the slope from Fig. 9:

y = 0.6493 ln(x ) + 9.4286 (45)

The slope of curve is:

Q 3000W
y' = 0.6493 = ⇒ 0.6493 = ⇒ k soil = 1.93513 W / mK (46)
4πk soil H 4πk soil 190m

The thermal conductivity was calculated to be in the order of 1.9 W/mK.

CONCLUSIONS
Borehole thermal resistance is a very important parameter that can lead to an erroneous estimate of the required
borehole depth for sizing the BHEs of GSHPs. There are several methods to calculate heat flow in the ground around
vertical boreholes. In this study analyses of thermal resistance of BTES system build at UOIT were investigated. For
this aim, a ground thermal characteristic of Oshawa was calculated primarily using the climate data. For the thermal
resistances inside the borehole, several suggestions from the literature were used in order to define the differences
between the correlations. For determining the thermal resistance outside the borehole, four different models were
used, namely, Kelvin’s Line Source Model, Eskilson’s Model, Infinite Cylindrical Heat Source Model and Finite Line-
Source Model. It was seen from the results that, all models ‘curves were similar to each other. Finally, thermal
response test for BTES system calculated the thermal conductivity as 1.9 W/mK.

For energy savings and reducing the emissions, GSHP systems are more effective than conventional systems. BHEs
are key parameters when deigning these kinds of GSHP systems. Dimensioning such systems requires knowledge
on ground thermal properties, building loads, climate conditions, etc. Since there are several parameters those are
influencing the GSHP systems, it is necessary to make detailed examinations before constructions of such systems.

Acknowledgement
The Authors gratefully acknowledge the financial support from the TUBITAK, Scientific and Technological Research
Council of Turkey, Turkish Academy of Sciences, and the assistance from the University of Ontario Institute of
Technology.

REFERENCES
Beatty, B., Thompson, J. 2004. 75 Km of Drilling for Thermal Energy Storage. Technical Report for BTES system
installed in UOIT/Durham College.

1138
Bennet J., Claesson J., Hellstrom G. 1987. Multipole Method to Compute the Conductive Heat Transfer to and
Between Pipes in a Composite Cylinder. Notes on Heat Transfer 3, Department of Building Physics, Lund Institute of
Technology, Lund, Sweden.

Cengel A.Y. 2002. Heat Transfer: A Practical Approach. McGraw-Hill, NY., 896 pp.

Chiasson A.D. 2007. Simulation and Design of Hybrid Geothermal Heat Pump Systems, PhD. Thesis, Department of
Civil and Architectural Engineering, University of Wyoming.

Claesson J., Dunand A. 1983. Heat extraction from the ground by horizontal pipes: a mathematical analysis. Swedish
Council for Building Research, Stockholm.

De Carli, M., Tonon M., Zarrella, A., Zecchin, R. 2010. A computational capacity resistance model (CaRM) for vertical
ground-coupled heat exchangers. Renewable Energy 35:1537–1550.

Deng Z. 2004. Modeling Of Standing Column Wells In Ground Source Heat Pump Systems. PhD. Thesis, Faculty of
the Graduate College of the Oklahoma State University.

Dincer I, Rosen M.A. 2007. A Unique Borehole Thermal Storage System at University of Ontario Institute of
Technology. in: Paksoy H.O. Thermal Energy Storage for Sustainable Energy Consumption. NATO Science Series
221-228.

Eklöf C., Gehlin S. 1996. TED - A mobile equipment for thermal response test. MSc. Thesis, Lulea University of
Technology, Sweden.

Eskilson, P. 1987. Thermal Analysis of Heat Extraction Boreholes. PhD. Thesis, University of Lund, Lund, Sweden.

Florides, G., Kalogirou, S. 2008. First in situ determination of the thermal performance of a U-pipe borehole heat
exchanger, in Cyprus. Applied Thermal Engineering 28:157–163.

Florides, G.A., Pouloupatis, P.D., Kalogirou, S., Messaritis, V., Panayides, I., Zomeni, Z., Partasides, G., Lizides, A.,
Sophocleous, E., Koutsoumpas, K. 2010. The geothermal characteristics of the ground and the potential of using
ground coupled heat pumps in Cyprus. Energy 36:5027-5036.

Georgiev A., Busso A., Roth P. 2006. Shallow borehole heat exchanger: Response test and charging–discharging
test with solar collectors. Renewable Energy 31:971–985.

Gu Y., O’Neal D.L. 1998a. Development of an equivalent diameter expression for vertical U-tube used in ground-
coupled heat pumps. ASHRAE Transactions 104:347–355.

Gu Y., O’Neal D.L. 1998b. Modeling the effect of backfills on U-tube ground coil performance. ASHRAE Transactions
104:356–365.

Gustafsson, A.M., Westerlund, L., Hellström, G. 2010. CFD-modelling of natural convection in a groundwater-filled
borehole heat exchanger. Applied Thermal Engineering 30:683–691.

Hellström, G. 1991. Ground Heat Storage-Thermal Analyses of Duct Storage Systems. PhD. Thesis, Department of
Mathematical Physics, University of Lund, Lund, Sweden, 310 pp.

Ingersoll, L. R., Plass, H. J. 1948. Theory of the ground pipe heat source for the heat pump. Heating, Piping & Air
Conditioning 20(7):119-122.

Ingersoll, L. R., Zobel, O. J., Ingersoll, A. C. 1954. Heat conduction with engineering, geological and other
applications. McGraw-Hill, New York.

Jenkins D.P., Tucker, R., Rawlings, R. 2009. Modelling the carbon-saving performance of domestic ground-source
heat pumps. Energy and Buildings 41:587–595.

Jun, L., Xu, Z., Jun, G., Jie, Y. 2009. Evaluation of heat exchange rate of GHE in geothermal heat pump systems.
Renewable Energy 34:2898–2904.

Keen Engineering. 2003. Rock and Soil Property Data. UOIT/Durham College GSHP Report, March.

1139
Lamarche L., Kajl S., Beauchamp B. 2010. A review of methods to evaluate borehole thermal resistances in
geothermal heat-pump systems. Geothermics 39:187–200.

Le Feuvre P. 2007. An Investigation into Ground Source Heat Pump Technology, its UK Market and Best Practice in
System Design. MSc. Thesis, Department of Mechanical Engineering, Strathclyde University.

Kasuda T., Archenbach P.R. 1965. Earth Temperature and Thermal Diffusivity at Selected Stations in the United
States. ASHRAE Transactions 71(1):61-76.

Pesl M., Goricanec D., Krope J. 2007. Response Functions and Thermal Influence for Various Multiple Borehole
Configurations in Ground Coupled Heat Pump Systems. WSEAS Transactions On Heat And Mass Transfer 3:61-71.

Philippe, M., Bernier, M., Marchio, D. 2010. Sizing Calculation Spreadsheet Vertical Geothermal Borefields. Ashrae
Journal 52:20-28.

Raymond J., Therrien R., Gosselin L. 2011. Borehole temperature evolution during thermal response tests.
Geothermics 40:69–78.
Remund C.P. 1999. Borehole thermal resistance: laboratory and field studies. ASHRAE Transactions, 105:439–45.

Sanner, B. 2003. Current Status of Ground Source Heat Pumps in Europe. 9th International Conference on Thermal
Energy Storage September 1-4, Warsaw, Poland.

Sharqawy M.H., Mokheimer E.M., Badr H.M. 2009. Effective pipe-to-borehole thermal resistance for vertical ground
heat exchangers. Geothermics 38:271–277.

Shonder J.A., Beck J.V. 2000. Field test of a new method for determining soil formation thermal conductivity and
borehole resistance. ASHRAE Transactions 106:843–850.

Wang, H., Qi, C., Du, H., Gu, J. 2010. Improved method and case study of thermal response test for borehole heat
exchangers of ground source heat pump system. Renewable Energy 35:727–733.

Weather Canada. 2011. http://www.weatheroffice.gc.ca/canada_e.html

Xu X. 2007. Simulation and Optimal Control of Hybrid Ground Source Heat Pump Systems, PhD. Thesis, Graduate
College of the Oklahoma State University.

Zeng, H. Y., Diao, N. R., Fang Z. H. 2002. A Finite Line-Source Model for Boreholes in Geothermal Heat Exchangers.
Heat Transfer-Asian Research 31(7):558-567.

1140
 INVESTIGATION OF DRYING CHARACTERISTICS OF BLUEBERRY USING A NEW
SOLAR AIR COLLECTOR DESIGN
* ** **
Fatih Danısman , Mustafa Aktas , Adem Acır
*
Gazi University Institute of Science and Technology, Ankara, Turkey
**
Gazi University, Faculty of Technology, Department of Energy Systems Engineering,
Teknikokullar/Ankara/TURKEY
* ** **
fsmturk@gmail.com, mustafaaktas@gazi.edu.tr, adema@gazi.edu.tr

ABSTRACT
In this study, a new solar air collector design was made for drying applications. Blueberry (Myrtus
communis) was dried in this drying process based on solar energy. The drying experiments were carried out
at 0.026, 0.0325 and 0.052 kg/s mass flow rate and variable air temperatures during between 8:30 and
18:00 hours. Solar radiation, bluberyy mass, collector inlet-outlet air temperatures, drying chamber inlet-
outlet air temperatures and relative humidity values were measured during the experiments. The drying air
velocity was controlled by using frequency inverter according to set value. Blueberry was dried from initial
moisture content (2.18 g water/g dry matter) to final moisture content (0.685, 0.286 and 0.402 g water/g dry
matter) for 0.026, 0.0325 and 0.052 kg/s mass flow rates, respectively. The average solar collector
efficiencies were found as 66.3 %; 59.9 % and 81.75 % for 0.026, 0.0325 and 0.052 kg/s mass flow rates,
respectively. As a result, the high efficiency solar air collector was obtained as regards thermal performance
and the blueberry drying process was performed successfully.
Keywords: Blueberry, myrtus communis, solar drying, moisture content, moisture ratio

INTRODUCTION
Drying of agricultural products has always been of great importance for the preservation of food by human
beings. Drying of agricultural products provides effective and practical preservation in order to reduce the
losses after harvest. Drying techniques of agricultural products is divided mainly three parts:
 Drying with heat pump
 Drying with solar energy
 Drying with other techniques (infrared energy etc.)
Drying with solar energy is a well-known method as preservation technique. The solar energy is a
renewable energy resource. Therefore, drying with solar energy is contributed in the economic cost and
utilize of the renewable energy. Also, solar drying is a safer and efficient method to preserve of agricultural
products. Solar drying will be reduced the product losses and it will be improved the quality of the final
product. In brief, solar drying methods are advantages due to free and renewable of solar energy according
to other drying methods. In the literature, three have been many studies with drying of agricultural products
using solar dryer. Different authors have performed on the drying of pistachio (Midilli, 2001), potato (Akpinar
et al., 2003), apple (Elicin and Sacilik, 2005; Aktaş et al., 2009), green chilli (Hossain and Bala, 2002),
green pepper, green bean and squash (Yaldiz and Ertekin, 2001) and apricot (Toğrul and Pehlivan, 2003;
Sarsılmaz et al., 2000). El-Sebaii et al. (El-Sebaii et al., 2002) dried green peas by using an indirect type
natural convection solar dryer. Jadhav et al. (Jadhay et al., 2010) used solar cabinet dryer for drying of
pretreated green peas. And also, mathematical modeling of drying characteristics of tropical fruits was
investigated by Ref. (Ceylan et al., 2009).

In this study, the drying characteristics of blueberry using a new solar air collector design have investigated
for mass flow rate and variable air temperatures. Also, efficiencies of a new type solar air collector have
been determined.

MATERIALS AND METHODS

Materials
Blueberry (Myrtus communis) was used in the drying experiments was supplied from the markets. It grows
spontaneously in the Black Sea, the Aegean and Mediterranean coasts in Turkey as shown in Fig. 1. Also,
blueberry contains high A vitamin and also it improves sight.

1141
Fig.1. A picture of blueberries

Experimental setup
The schematic diagram of the experimental setup was shown in Fig.2. An experimental set-up was
constructed and tested in Gazi University, Ankara, Turkey. The main components were solar collector, air
channels, a fan, an air regulating damper, PLC control unit and drying chamber.

1: Solar collector, 2: Data-logger, 3: Solarimeter, 4: Computer, 5: Exhaust air, 6: Blower, 7: Flexible air duct, 8:
PLC controlled drying cabinet, 9: PLC screen.
Fig. 2. A schematic view of the PLC controlled experimental set-up

The collector size used of experiment was 800 mm in length, 500 mm width and 120 mm in height. The
collector was the single pass finned collector. The collector consists of the glass cover, the insulated
container and the black painted copper absorber. The collector was designed as parabolic absorber plates.
Also, technical properties and equipments used drying based on solar energy are given Table 1. Solar
drying experiments were carried out between 5 July and 10 July 2012 (9.00 am–18.30 pm) in Ankara,
Turkey for different mass flow rates.

1142
Table 1. Main properties of equipments used in solar dryer.
Equipment Technical properties
PLC Controller Controlling of air temperature, relative humidity, air velocity and dried product
mass, LCD screen.
3
Blower 160 W, 990 m /h, 2750 d/d, 50 Hz, 230 V IN:0.74A
2 2
Solarimeter SL 100, 1 W/m -1300 W/m measurement range
Data-logger VCC 0,1V, 2.7 V-5.5 V operating range, 110dB minimum
o
Thermocouple Inlet Pt-100, R/T Type Scala: 0-70 C
o
Load-cell Nominal out (mV / V) 2,0 ± 0.2, 5-12 V, Operation temperature: -35 +65 C,
Error (%FS): 0.02

Uncertainty Analysis
Uncertainty analysis in the experiments is very important in order to be able to make reliable measurement.
The instrument selection, condition, calibration, environment, observation, reading, and test planning can be
caused of errors and uncertainties in the experiments. In drying experiments in solar dryer of blueberry,
weather temperature, relative humidity, weight losses, wind speed and solar radiation were measured with
appropriate instruments. During the experiments, total uncertainties of the measured parameters and
calculated experimental parameters were presented in Table 2.

Table 2. The uncertainties during the measurements of the parameters.

Parameter Unit Comment
Uncertainty in the temperature measurement
o
Collector inlet temperature C ±0.173
o
Collector outlet temperature C ±0.173
o
Absorber surface (copper plate) C ±0.173
o o
Ambient air temperature C ±0.173 C

Uncertainty in the time measurement

Temperature values Min ±0.1
Uncertainty in the air velocity measurement m/s ±0.1446
Uncertainty in the measurement of pressure loss Pa ±0.3
Uncertainty in reading values of table % ±0.1-0.2

Considering to relative uncertainties in the individual factors denoted by xn, uncertainty estimation was
made using the following equation (Holman, 1994):


W  ( x1 ) 2  ( x2 ) 2  ...( xn ) 2 
1/ 2
(1)

MATHEMATICAL MODELLING
In this study, initial moisture content, moisture ratio and solar thermal efficiency were investigated. Initial
moisture content (dry basis) of the blueberry was calculated using the following equations (Elicin and
Sacilik, 2005; Ceylan et al., 2009):

Mi  Md
MC  (2)
Md

The moisture ratio of blueberry during drying experiments is calculated using the following equations (Elicin
and Sacilik, 2005; Ceylan et al., 2009):

M  Me
MR  (3)
M0  Me

where MR is the moisture ratio, M the moisture content, Me the equilibrium moisture content, and Mo is the
initial moisture content, respectively.

1143
 Also, the thermal efficiency of the solar collectors (I) is defined as the ratio between the energy gain and
the solar radiation incident on the collector plane (Karsli, 2007; Esen, 2008):

.
m c p (Ta ,out  Ta ,in )
I  (4)
( IAc )

EXPERIMENTAL RESULTS AND DISCUSSION

The drying experiments were carried out at 0.026, 0.0325 and 0.052 kg/s mass flow rate, 40° collector
angle and variable air temperatures during between 8:30 and 18:00 hours. Change in mass was recorded
every 30 minutes in the experiment. In this study, firstly, the dry matter amount, initial and final moisture
contents for each product are determined. The initial moisture content of the Blueberry samples (50 g) was
considered. Initial moisture content (dry basis) of the Blueberry was calculated from Eq. (1). The initial
moisture contents (MC) of Blueberry was calculated as 2.25 g water/g dry matter (Fig. 3).

2,5
Moisture content (g water/g dry matter)

m=0.026 kg/s
m=0.0325 kg/s
2
m=0.052 kg/s

1,5

0,5

0
0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 5,5 6 6,5 7 7,5 8 8,5 9 9,5
Drying time (h)
Fig. 3. The change of moisture content in the experiments.

The final moisture content are obtained as 0.685, 0.286 and 0.402 g water/g dry matter for 0.026, 0.0325
and 0.052 kg/s mass flow rates, respectively. Final drying levels of Blueberry are realized in 9 hours in the
solar dryer. In addition, the moisture ratio (MR) of Blueberry during drying experiments was calculated from
Eq. (2). As shown in Fig 4, the moisture ratio are computed as 0.272, 0.248 and 0.308 for 0.026, 0.0325
and 0.052 kg/s mass flow rates, respectively.

1,2

1
m=0.026 kg/s
m=0.0325 kg/s
Moisture ratio, MR

0,8
m=0.052 kg/s

0,6

0,4

0,2

0
0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 5,5 6 6,5 7 7,5 8 8,5 9 9,5

Drying time (h)

Fig. 4. Variation of moisture ratio in the experiments.

1144
As a result, the optimum values as regards moisture ratio were provided in 0.0325 kg/s mass flow rate of
experiments as seen in Table 3. Also, a picture of blueberries before drying and after drying was given Fig.
5.

a) b)
Fig. 5. A picture of blueberries a) before drying, b) after drying

During the drying experiments, the solar radiation ranged from 680 to 810 W/m2. Generally, the maximum
solar radiation energy was obtained at midday whereas; the minimum solar radiation energy at evening in
the day was seen. The solar collector efficiencies based on obtained solar radiation were changed during
days. In this study, a new collector was designed. This solar air collector was considered as parabolic
absorber plates. In this study, the average solar collector efficiencies were calculated with Eq. 4. The solar
collector efficiencies values during experiment are given in Fig. 6, which are plotted using Eq. (4). As shown
in Table 3, the average thermal efficiency were found as 66.3 %; 59.9 % and 81.75 % for 0.026, 0.0325 and
0.052 kg/s mass flow rates, respectively. When the thermal efficiency obtained in this present study was
compared the results obtained in the previous studies given in literature, the solar collector having parabolic
absorber plates was very good performance exhibition.

Table 3. Experimental results for 40° collector angle

Mass flow Final Final Final T
I η
rate, kg/s YA MC MR
o
2 ( C)
(gr) (g water/g dry matter) (%) (W/m ) (%)
0.026 25.06 0.594 0.272 709.9 5.90 56.97
0.0325 24.21 0.540 0.248 764.0 5.75 63.30
0.052 26.28 0.672 0.308 688.9 3.61 71.59

120
m=0.026 kg/s
100 m=0.0325 kg/s
m=0.052 kg/s
Thermal efficiency, %

80

60

40

20

0
0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 5,5 6 6,5 7 7,5 8 8,5 9 9,5
Drying time (h)
Fig. 6. Variation of thermal efficiency values of the collectors

CONCLUSIONS
In this study, drying characteristics of Blueberry (Myrtus communis) of a solar dryer controlled with PLC was
investigated. The drying process can be continued with solar energy in daytime. Solar energy can provide
free and renewable energy. The initial moisture content of blueberry were calculated as ~ 2.25 g water/g dry
matter whereas; the final moisture content are obtained as 0.685, 0.286 and 0.402 g water/g dry matter for

1145
0.026, 0.0325 and 0.052 kg/s mass flow rates, respectively. The moisture ratio was computed as 0.272,
0.248 and 0.308 for 0.026, 0.0325 and 0.052 kg/s mass flow rates, respectively. As a result, the optimum
values as regards moisture ratio were provided as 0.248 in 0.0325 kg/s mass flow rate. The drying process
can be expected much more quickly than according to nature drying. Since the quality of blueberry dried by
PLC controlled solar dryer were acceptable and drying takes place in a shorter time other drying methods.
In conclusion, it is expected that this system will help reduced the cost of drying and obtain more quality
dried products. Also, a new design solar collector having parabolic absorber plates was very good
performance exhibition.

ACKNOWLEDGEMENT
This work has been supported by Scientific Research Project Unit of Gazi University (G.U.BAP, 07/2012-
42).

NOMENCLATURES

W : uncertainty in the measurement (%)

I :thermal efficiency
MC :moisture content (g water/g dry matter)
MR :moisture ratio
Mi :initial wet weight (g)
Md :final dry weight (g)
Me :equilibrium moisture content (g water/g dry matter)
M0 : initial moisture content (g water/g dry matter)
I :solar radiation (W/m2)
AC :surface area of the collector (m2)
cp : specific heat of air at constant pressure (kJ/kg K)
Ta : air temperature (°C)

1146
REFERENCES
Akpinar, E., Midilli, A. and Bicer, Y., 2003. Single layer drying behaviour of potato slices in a convective
cyclone dryer and mathematical modeling. Energy Conversion and Management. 44: 1689–1705.

Aktaş M., Ceylan İ. and Yılmaz S., 2009. Determination of drying characteristics of apples in a heat pump
and solar dryer. Desalination. 239: 266–275.

Ceylan İ., Aktaş M. and Doğan H., 2009. Mathematical modeling of drying characteristics of tropical fruits.
Applied Thermal Engineering. 27:1931-1936.

Elicin, A.K. and Sacilik, K., 2005. An experimental study for solar tunnel drying of apple. Tarim Bilimleri
Dergisi. 11: 207–211.

El-Sebaii, A.A., Aboul-Enein, S., Ramadan, M.R.I. and El-Gohary, H.G., 2002. Experimental investigation of
an indirect type natural convection solar dryer. Energy Conversion and Management. 43: 2251–2566.

Esen H., 2008. Experimental energy and exergy analysis of a double-flow solar air heater having different
obstacles on absorber plates. Building and Environment. 43: 1046–1054.

Holman JP. Experimental methods for engineers. 6th ed. Singapore: McGraw-Hill; 1994.

Hossain, M.A., Bala, B.K., 2002. Thin-layer drying characteristics for green chilli. Drying Technology. 20:
489–505.

Jadhav, D.B., Visavale, G.L., Sutar, N., Annapure, U.S. and Thorat, B.N., 2010. Studies on solar cabinet
drying of green peas (Pisum sativum). Drying Technology. 28: 600–607.

Karsli S., 2007. Performance analysis of new-design solar air collectors for drying applications. Renewable
Energy. 32: 1645–1660.

Midilli, A., 2001. Determination of pistachio drying behaviour and conditions in a solar drying system.
International Journal of Energy Research. 25: 715–725.

Sarsılmaz, C., Yildiz, C. and Pehlivan, D., 2000. Drying of apricots in a rotary column cylindrical dryer
(RCCD) supported with solar energy. Renewable Energy. 21: 117–127.

Togrul, I.T., Pehlivan, D., 2003, Modelling of drying kinetics of single apricot. Journal of Food Engineering.
58: 23–32.

Yaldiz, O. and Ertekin, C., 2001. Thin layer solar drying of some vegetables. Drying Technology. 19: 583–
596.

1147
 INVESTIGATION OF TURBIDITY EFFECT ON TRANSMISSION OF HEAT IN A SOLAR POND
1 2 1 3
Ayhan Atiz , Ismail Bozkurt , Mehmet Karakilcik and Ibrahim Dincer
1
Department of Physics, Faculty of Sciences and Letters, University of Cukurova, Adana 01330, Turkey.
2
Department of Mechanical Engineering, Faculty of Engineering, University of Adiyaman,
Adiyaman, 02040, Turkey.
3
Faculty of Engineering and Applied Science, University of Ontario Institute of Technology (UOIT), 2000 Simcoe Street,
North Oshawa, Ontario, L1H 7K4, Canada.

ABSTRACT
Renewable energy sources, such as solar, play a key role due to their environmentally benign nature. There
are various types of solar energy applications, including solar ponds. A solar pond consists of three different
density zones, such as Upper Convective Zone (UCZ), Non-convective Zone (NCZ) and Heat Storage Zone
(HSZ). The zones’s density increases from top to bottom of the pond. Each of the zones is important and
affects the performance of the pond. Because of this, many investigators aim to increase the efficiency of the
pond for better performance. There are also some key parameters affecting the efficiency. One of them is
turbidity, which is caused by dirty over time, (e.g., insects, leaf, dust and wind bringing parts fall down). Thus,
the turbidity in the salty water decreases the heat transmission through the zones. In this study, the samples
are taken from the zones of the solar pond and analysed by using spectrometer for one year. As a result, it is
confirmed that the solar pond performance is greatly affected by turbidity.

Keywords: Solar energy, solar pond, efficiency, turbidity, transmission.

INTRODUCTION
Solar energy is the great big energy source. Turkey lies in a sunny belt between 36° and 42° N latitudes. The
2
yearly average solar radiation is 3.6 kWh/m day and the total yearly radiation period is approximately 2640 h
which is sufficient to provide adequate energy for solar thermal applications (Kaygusuz and Sari, 2003).
The energy content of solar radiation can be used as light, heat and electricity. This diversity makes
solar energy an important option to power different energy systems all over the world. Indeed, the interest in
solar energy systems has been increasing in recent years throughout the world. This interest has been due
to several factors such as the environmental damage the effort to minimize countries dependence on fossil-
based, non-renewable fuels and the international agreements demanding reduction in the greenhouse gases
in the earth’s atmosphere (Khan et al., 2005). Heat storage technologies systems and applications are
studied in the field of solar energy (Dincer, 1999; 2002). Solar ponds and collectors are very important solar
energy systems that generate heat energy from solar energy. Solar ponds are relatively simple devices that
operate straightforwardly, require little maintenance (e.g., they need cleaning to maintain water
transparency), and have long life times (Dincer and Rosen, 2011). Solar ponds have been proposed and
demonstrated to be simple in design low in cost and naturally abundant for the large scale collection and
storage of solar energy. Solar enegry is received as both direct and diffuse radiation and distributed primarly
over the wide range of the ultraviolte, visible and infrared spectra. A higher radiation transmission through
the upper zone and gradient in a solar pond results in a higher heat collection effeciency (Wang and
Yagoobi, 1994). The attenuation of light in the clearest natural waters, solar pond and salt water has been
varied. Irradiance just not depend on water and salty. There is another important thing impact, we know this
is turbidity and common problem encountered in salinity-gradient solar ponds is the growth of various types
of algae and bacterial populations which affects the brine clarity and hence reduces thermal performance
(Malik et al., 2011).
A solar pond as a solar collector is dependent on light transmission to the storage zone. Clarity of the
brine is one of the most important factors in achieving good thermal performance and stability the higher the
transparency the better the performance (Xu et al., 1994). Therefore, it is important to maintain the clarity of
solar ponds to allow maximum possible solar radiation to reach the LCZ (Alagao, 1994).
In the present study, we conduct an experimental investigation of a solar pond and the effect of
turbidity on the pond efficiency and study how much effect is possible under various varying conditions.

EXPERIMENTAL APPARATUS AND PROCEDURE

Solar ponds
In general, solar ponds are composed of three zones. The first zone, upper convective zone (UCZ) is the
fresh water layer at the top of the pond. This zone is fed with fresh water in order to maintain its density as

1148
close as to the density of fresh water. The second zone is the non-convective zone (NCZ) between the heat
storage zone (HSZ) and UCZ. NCZ is composed of salty water layers whose brine density gradually
increases towards HSZ. NCZ is the key to the working of a solar pond. The third zone is known as heat
storage zone (HCZ). HSZ is composed of salty water with highest density. Considerable part of the solar
energy is absorbed and stored by this bottom region (Karakilcik et al., 2006).
In the present experimental work, a solar pond with a radius of 0.80 m and a depth of 2 m was built
in Cukurova University in Adana, Turkey (i.e., 35°18 E longitude, 37°05 N latitude). The experimental
temperature distributions were measured using temperature sensors, which were placed into the inner zones
of the solar pond. The bottom and the side-walls of the pond were plated with the iron sheets in 0.005 m
thickness and in between with a glass–wool of 0.10 m thickness as the insulating layer. Fig.1 illustrates the
inner zones of the solar pond. Inner zones consist of the saline water layers with various densities. To
measure the temperature distributions of various regions, the temperature sensors were placed into the inner
zones, starting from the bottom, at 0.10, 0.30, 0.50, 0.70, 0.90, 1.10, 1.30, 1.50, 1.70, 1.90 m heights. The
data acquisition system was connected to a computer for data recording monitoring. Hence the temperature
distribution profiles of these regions at any time were experimentally obtained by a data acquisition system.
The solar pond is filled up with salty water. HSZ is started from the bottom of the pond until 0.80 m height
and its denstity approximately 1180 kg/m³ with salt water. The middle zone, NCZ is composed of salty water
layers whose brine density gradually increases towards HSZ. NCZ has five layer thicknesses, and each layer
is 0.20 m thick. These layers denstity between 1100 kg/m³-1015 kg/m³ decresing as graded. The surface
zone, UCZ is the fresh water layer at the top of the pond.

Fig.1. Inner zones of the solar pond.

The salinity gradient sustainability of the solar pond is very important for heat storage performance.
Therefore, the density of the solar pond becomes critical. To measure the density distribution of the solar
pond put 5 mm diameter 10 unit transparent plastic hoses with 20 cm distance on 2 m length board in the
pond. The salinity water samples of the layers are taken from the hoses to measure density by using
hydrometer.

Salt gradient protection system

The density gradient in the NCZ is very important to keep the stored energy. In order to maintain the density
gradient of the pond an auxiliary operating system on a feedback princible was developed. It generated a
continuous and self regulating operation. This was a passive system based on the natural circulation of
water caused by density difference as it was first proposed by Akbarzadeh and MacDonalds (1982). It was
built a cylindrical polyethylene tank having 1.20 m lenght and 0.30 m diameter. The tank has a connection
from its bottom to the top of the HSZ. Because of the density difference between the storage zone and a
higher zone, the less dense brine entered via a plastic tube to the top of the tank from top of NCZ and
diffused through a salt bed on a mesh within the tank and reached about 1180-1200 kg/m³ density at the
bottom. This more dense brine was injected into the top level of HSZ via a distributing pipe system. The salty
water distribution system was made of circular polyethylene pipe with a diameter of 1.20 m and the thickness

1149
with 0.02 m diameter. It has several holes along them led the higher density brine out at the top of the HSZ
(Bozkurt and Karakilcik, 2012).

Spectrometer
A spectrometer (spectrophotometer, spectrograph or spectroscope) was used to measure the properties of
light over a specific portion of the electromagnetic spectrum as it is used in spectroscopic analysis to identify
materials (Butler and Laqua, 1995). The measurement samples were contained a solution of distillied water
and of normal grade salt. The following salts were studied: MgCl2, Na2SO4, NaNO3, KNO3 and Na2CO3. The
transmission spectra were measured with a spectrometer in the range 300-1200 nm. The quartz–glass
sample containers of 50 mm length were of good quality (Lund and Keinonen, 1984).

Optical properties of salty water

Essentially all of the solar radiation absorption that takes place in most natural waters is attributable to three
components (the water itself, dissolved organic matter, variously referred to as yellow substance gelbstoff
and gilvin) Kirk (1983). In order to increase the energy input, owing into the storage zone of the solar pond,
NCZ and UCZ should be transparent as possible as because both the intensity of incident ray and the
absorptance of salt-water solution depend on the wavelength to calculate the total transmittance precisely (Li
et al. 2000).
Estimation of total solar radiation available at various depths of a salt gradient solar pond is an
integral part of modeling its thermal performance. The radiation flux incident at the surface of the pond gets
partially reflected and the remainder penetrates into the pond. The radiation flux as it propagates through the
body of the pond gets attenuated in its path. (Husain et al. 2004). The relation among absorption, reflection
and transmission coefficients is given as

ρ  α  τ 1 (1)

where α, ρ and τ are the absorption, the reflection and the transmission, respectively.
Most of the solar radiation in the near infrared spectrum is absorbed within the first centimeter of
depth. The radiation associated with wavelengths greater than 1200 nm represented 22,4% of the total
incident radiation and this radiation was totally absorbed in this upper water boundary layer for radiation with
<1200 nm (Rabl and Nielsen, 1975). The radiation emitted by the sun travels through the vacuum of space
unaltered. The percentage of energy associated with certain bandwidths of solar radiation emitted from a
blackbody at 5800 K (the temperature of the sun) (Holman, 1997). The light entering the pond is also
scattered by the various suspended particles. The scattering particles however are assumed not to absorb
energy (Tsilingiris, 1991).
Here, we note that β is the fraction of the incident solar incident that actually enters the pond (Hawlader,
1980).

2 2
 sinθi  sinθ r   tanθi  tanθ r 
β  1  0.6   0.4  (2)
 sinθi  sinθ r   tanθi  tanθ r 

where θi and θr are the incidence angle and the refraction angle. The attenuation of radiation through a pure
medium is described by the Bouger-Beer Law; to measured solar energy with deep giving Lambart-Beer Law
that is depend on wave length  and way, ray of light intensity in the x deep as follows:

I  ( x)  I 0 i exp(   x) (3)

where, I(x) is the wave lenght  bunch of the beam in the x deep. The deep lenght the the light took the way
in water. I 0 i is entered the water all part of sun energy and = −1/ wave lenght of beam absorbing
coefficient  ,  wave lenght of beam coming carecteristic of absorbing coefficient giving the deep. The
formulation can calculate every wave lenght of absorbing energy but this take long time and it is diffucult
application in pratic. Many application of formules divided spectrum of sun energy n part and it is given as

1150
 n n

 I i   I 0i exp(i x)
i 1 i 1
(4)

which can be arranged to find the absorbing rate of function h(x) as

n

I 0i exp(   i x)
h( x )  i 1
n
(5)
I i 1
0i

I 0i
where we accept the n
as i and obtain the following:
I
i 1
0i

n
h( x)   i exp(   i x) (6)
i 1

RESULTS AND DISCUSSION

A solar pond is used to store the heat energy in the heat storage zone. The first part of the experimental
investigation is carried out to determine the temperature and density distribution of the solar pond. In the
second part of the experiment it is aimed to determine the turbidity effect on the performance of the solar
pond by using spectrometer. As seen in Figs. 2-3, the differences of the salinity gradient is kept
approximately stable by using salt gradient protection system. We wanted to show change of the gradient
solar pond due to diffusion of salt molecules throughout the one year.

1200
1180
1160
1140
Density (kg/m3 )

1120
1100
1080
1060
1040
1020
1000
20 40 60 80 100 120 140 160 180 200
Height (cm)

Fig.2. Variation of the density with depth in August, 2009.

Figures 4 and 5 show the transsmission of the experimental turbidity and clear samples from solar
pond at the 10 different point, vertically. The characteristic transmission of ponds appears to be very different
from each other. A UV-Vis spectrometer was used in order to understand characteristic transmission of every
samples (e.g. clear salty water and turbidty salty water). UV-Vis spectrum measurements were performed in
the wavelength range of 400–700 nm by a Perkin Elmer UV-Vis Lambda 2S spectrophotometer at room
temperature. We used two clean quartz–glass sample containers, including the clean water and a turbidty
salty water. The clean water in quartz–glass sample container was used as reference.

1151
 1200
1180
1160
1140
Density (kg/m3)

1120
1100
1080
1060
1040
1020
1000
20 40 60 80 100 120 140 160 180 200
Height (cm)

Fig.3. Variation of the density with depth in August, 2010.

100
1000kg/m3
98
1015kg/m3
96
1030kg/m3
94
1050kg/m3
% Transmission

92
1075kg/m3
90
1130kg/m3
88
1100kg/m3
86
1150kg/m3
84
1180kg/m3
82
1180kg/m3
80
400 450 500 550 600 650 700
Wavelength (nm)

Fig.4. Variation of the tranmission of the wavelength for clean water of solar pond in August, 2009.

Tables 1 and 2 give the absorption coefficients of turbidty and clear water in the solar pond, based
on the samples taken from solar pond at 10 different heights from top to bottom. The absorption coefficients
of the sample were measured by using the spectrometer. The Beer-Lambert Law was used to calculate the
absorption coefficients between 400 and 700 nm. In Eq. 3, the thickness of x was measured as 0.5 cm using
the quartz–glass sample container in the spectrometer. Then, using the transmission data in Figs.4 and 5
with Eq.3 the absorption coefficients listed in Table 1 and 2 were determined. These tables give the
transmission data of light with different depths for turbidty water and clear water in the pond.

1152
 105
1021kg/m3
100
95 1021kg/m3
% Transmission

90 1061kg/m3
85
1100kg/m3
80
1151kg/m3
75
70 1170kg/m3
65 1180kg/m3
60
1180kg/m3
55
50 1180kg/m3

45 1181kg/m3
40
400 450 500 550 600 650 700
Wavelength (nm)

Fig.5. Variation of the tranmission of the wavelength for turbidty pond in August, 2010.

Table 1. Absorption coefficients with respect to transmission data for turbidity solar pond.
Bandwidth (nm)
Height(cm)
600-700 500-600 400-500
0-20 500-600
0.002 0.005 0.015
20-40 400-500
0.002 0.004 0.006
40-60 0.005 0.011 0.004
60-80 0.012 0.024 0.036
80-100 0.025 0.044 0.101
100-120 0.222 0.244 0.159
120-140 0.224 0.247 0.159
140-160 0.265 0.343 0.247
160-180 0.248 0.296 0.170
180-200 0.221 0.244 0.148

Table 2. Absorption coefficients with respect to transmission data for clear solar pond.
Bandwidth (nm)
Height(cm)
600-700 500-600 400-500
0-20 500-600
0.004 0.004 0.004
20-40 400-500
0.002 0.002 0.003
40-60 0.002 0.005 0.008
60-80 0.007 0.001 0.018
80-100 0.013 0.013 0.027
100-120 0.011 0.021 0.018
120-140 0.008 0.012 0.024
140-160 0.012 0.017 0.018
160-180 0.015 0.022 0.030
180-200 0.017 0.022 0.036

1153
 Here, we consider each layer thickness equal to x=20 cm from top to bottom in the pond and the
solar radiation reaching the surface of the solar pond as 100% visible light. We used Eq. 3 to calculate the
transmission values of the first layer. Thus, the calculation was repeated for each layer of the inner zone
down to the bottom. Note that the absorption coefficients listed in the tables are used to calculate the
transmission of the clear and the turbidity water in the pond. The transmission results were then used to
obtain the transmission distrubutions in the salinity water to understand the entering the total visible
spectrum in clear and turbidity water of the solar pond. It is assumed that the sun has a surface temperature
of 5800 K. In this regard, the total emitted energy for given bandwidths was calculated to be 38.5% of all
emmited energy as associated with the visible spectrum in the atmosphere (Holman, 1997). As seen in
Figs.6 and 7, we present the percentage of the entering visible light of the storage zone from 120 cm to 200.
The values of 5.39% and 11.93% were found for turbidty and clear salinity water solar pond for the storage
zone, respectively. Figure 8 shows the temperature distributions of clean water and tubidity water. As seen
there, there is some difference between these two profiles.

45

40

35
% Transmission

30

25

20

15

10

0
0 20 40 60 80 100 120 140 160 180 200
Height (cm)

Fig.6. Transmission of visible spectrum for clear solar pond.

45

40

35
% Transmission

30

25

20

15

10

0
0 20 40 60 80 100 120 140 160 180 200
Height (cm)

Fig.7. Transmission of visible spectrum for turbidity solar pond.

1154
 Fig.8. Variation of the temperature distrubutions for clean and turbidity saline water with depth in one year
period for August.

CONCLUSIONS
Transmission visible spectrum of the light has been investigated in turbidity and clear saline water of a solar
pond. We compare the transmission of the turbidity and clean salinity water to understand both of effect on
performance of the solar pond. The turbidity in saline water of the pond is increased in time due to outdoor
(e.g., insects, leaf, dust and wind bringing parts fall down). This increasing is caused by decreasing
transmission of solar pond. Therefore, especially the temperature distrubutions are decreased in the storage
zone in time. The results show that the performance of the solar pond is affected strongly by the turbidity.
The turbidity is not homogeneous in the layers which is caused variation transmission ratios. The
transmission ratios of saline water is an important parameters performance of the storage zone.

Acknowledgement
The authors are thankful to University of Cukurova for financial support of the present work (Grant No.
FEF2004BAP4, FEF2010YL26, FEF2009D2 and FEF2010BAP5).

NOMENCLATURE
IBW insulated bottom wall
ISW insulated side walls
HSZ heat storage zone
LCZ lower convective zone
n number of the part
NCZ non convective zone
T temperature, °C
UCZ upper convective zone
K kelvin temprature
a absorption coefficient
θi incidence angle
θr refraction angle
h(x) transmittance of the working fluid
I (x) wavelength bunch of beam in the deep
x deep lenght the the light took the way in water.
Ioi entered the water all part of sun energy
 wave lenght of beam absorbing coefficient.
 wave lenght of beam coming carecteristic of absorbing coefficient giving the deep

1155
Greek Letters
ρ reflection coefficient
t transmission coefficient
3
 density, kg/m
 incident beam entering rate into water
 wavelength, nm

REFERENCES
Akbarzadeh, A. and MacDonald, R.W.G. 1982. Introduction of a passive method for salt replenishment in the
operation of the solar ponds. Solar Energy 29:71-76.

Alagao, F.B. 1996. Simulation of the transient behavior of a closed-cycle salt-gradient solar pond. Solar
Energy 56:245-260.

Bozkurt, I and Karakilcik, M. 2012. The daily performance of a solar pond integrated with solar collectors.
Solar Energy 86:1129-1672.

Bryan, H.C. and Colbeck, I. 1977. A Solar pond for London. Solar Energy 19:321

Butler, L.R.P. and Laqua, K. 1996. Nomenclature, symbols, units and their usage in spectrochemical
analysis-IX. Instrumentation for the spectral dispersion and isolation of optical radiation. Spectrochimica
Acta Part B: Atomic Spectroscopy 51:645-664

Dincer, I. 1999. Evaluation and selection of energy storage systems for solar thermal applications.
International Journal Of Energy Research 23:1017-1028.

Dincer, I. 2002. Thermal energy storage systems as a key technology in energy conservation. International
Journal of Energy Research 26:567-588.

Dincer I.. Rosen M.A. 2011. Thermal Energy Storage Systems and Applications. New York:Wiley, Print
ISBN:978-0-470-74706-3.

Holman, J.P. 1997. Heat Transfer (8th edition). McGraw-Hill.

Husain M., Patil S.R. and Samdarshi, S.K. 2004. Simple methods for estimation of radiation flux in solar
ponds. Energy Conversion and Management 45:303-314

Kaygusuz, K. and Sarı, A. 2003. Renewable energy potential and utilization in Turkey. Energy Conversion
and Management 44:459-478.

Khan, F.I., Hawboldt, K. and Iqbal, M.T. 2005. Life cycle analysis of wind–fuel cell integrated system.
Renewable Energy 30:157-77.

Kirk, J. 1980. Spectral absorption properties of natural water's: Contribution of the soluble and particulate
fractions to light absorption in some inland waters of South Australia. Australian J. of Mar. and Freshwater
Res. 31:287-296.

Li, Y.X., Kanayama K. and Baba, H. 2000. Spectral calculation of the thermal performance of a solar pond
and comparison of the results with experiments department of mechanical engineering. Renewable Energy
20:371-387.

Lund, P.D. and Keinonen, R.S. 1984. Radiation transmission measurements for solar ponds. Solar Energy
33:237-240.

Malik N., Date A., Leblanc J., Akbarzadeh A. and Meehan B. 2011. Monitoring and maintaining the water
clarity of salinity gradient solar ponds. Solar Energy 85:2987-2996.

1156
Rabl, A. and Neilsen, C.E. 1975. Solar ponds for space heating. Solar Energy 17:1-2.

Smith. R.C. and Baker, K.S. 1981. Optical properties of theclearest natural waters. Applied Optics. 20:177.

Tsilingiris, P.T. 1991. Radiation transmission through a composite pure water layer of uniform and
depth-dependent extinction characteristics. International Journal of Energy Research 15: 723-730.

Xu H.. Sandoval, J.A., Lu, H., Golding. P. and Swift, A. 1994. Operating experience with the El Paso Solar
Pond. In:Joint Solar EngineeringConference ASME pp:503–511.

Wang, J. and Seyed-Yagoobi, J. 1994. Effect of halobacteria and selected chemicals on radiation
transmission in a salt water. Solar Energy 52:411-418.

1157
 ENHANCEMENT OF GAS TURBINE PERFORMANCE USING COOLED INLET AIR
AND SIMULATION WITH EES SOFTWARE
1 2 3
Saeed Karimi Alavijeh , Keyvan Mafi , Amir Karimi Alavijeh
1 rd
Master of Science in Mechanic Engineering 3 Gas Refinery, South Pars Gas Complex Company (SPGC), Iran
2
Bachelor of Mechanical Engineering Faculty of Mechanical Engineering, Azad University of Takestan, Iran
3
Master of Science in Aerospace Engineering Iran Helicopter Company, Iran
e-mail: Saeed.Karimealvijeh@nigc-spgc.ir, Keyvan.mafi@gamil.com, Amir.Karimi.Alavijeh@gmail.com

ABSTRACT
Having widely used gas turbine in various industries, such as electricity, gas, aviation and military industries turned
it to one of the most important engineering issues. So, the relevant companies try to explore the results of these
power generation systems as possible in order to improve the efficiency and reduce fuel consumption and also
they consider the reduction of environmental pollution. About %35 of country’s annual power generation is
produced by the gas units. However, by simple and extensive instruction in the industry, the cycle of power
generation is along with relatively low efficiency (generally less than %35), thus they need high fuel consumption.
During the warm months, a gas turbine inlet air cooling technique is a useful option for increasing output. Inlet air
cooling increases the power output by taking advantage of the gas turbine’s feature of higher mass flow rate, due
the compressor inlet temperature decays.
In the current study, the effect of compressor intake air cooling with absorption chiller is theoretically studied for a
typically gas turbine. Initially the gas turbine cycle simulated in software then the compressor inlet air cooling
method in order to increase the efficiency studied. By the thermodynamically analysis of the cycle and using EES
software simulation, the effect of this method on increasing the efficiency of cycle showed.
The results obtained with this method are compared with the values of the condition without cooling system.
Results shows after using the cooled inlet air, total power of the gas turbine increases by 518kw also the efficiency
of gas turbine cycle increase by 7% and it makes this cycle more efficient and cost-effective.

INTRODUCTION
Today, by considering the problems of the world about the supply of fuel, the high cost of energy transfer and also
the movement of reduction of underground fuel reserves, any slight increase in the power generation’s efficiency
will cause a significant reduction in fuel assumption and cost saving in generation.
The gas turbine is composed of a compressor that supplies high pressure air for the combustor chamber, which
provides high temperature flue gas at high pressure. Gas turbines are of constant-volume and their power output
is directly proportional and limited by the air mass flow rate. As the compressor has a fixed capacity for a given
rotational speed and volumetric flow rate of air, their volumetric capacity remains constant. Therefore, the mass
flow rate of air enter into the gas turbine varies with their specific mass, which means that it depends on the
temperature and the relative humidity of the ambient air (American Society of Heating, 2008).
Also exit of hot combustion gasses (exhaust gas) is one of the problems in operation of gas turbine which it
produces the gasses in the combustion chamber, after passing the power stage; because it would be discharged
to the environment by high temperature (about 400oC). During this operation regional climate and weather
condition would be affected by irreversible results. In addition, the high temperature of these gasses let to the
energy exhaust of generation cycle as the sensible heat and also energy loss.
Nowadays, one of the main methods used to increase the gas turbine efficiency is the cooling of the compressor
intake air. There are two main commercially options for this method: evaporative cooling and mechanical or
absorption chillers. Santos et al. (2012) studied the effect of different compressor intake air cooling systems both
evaporative media cooling and chiller systems for a single shaft gas turbine.
Gas turbine performance is critically limited by the amounting ambient temperature, mainly in hot and dry regions.
It occurs because the power output is inversely proportional to the ambient temperature.
As gas turbine has been widely used for power generation and the demand for electricity is highest during periods
of elevated ambient temperatures. This factor leads to an increase in power plant peaking capacity. For example,
In the Saudi Electric Company’s (SEC), approximately 42% of the annual energy is generated by gas turbines, and
during the summer these turbines suffer a 24% decrease in their capacity, due to ambient temperature up to 50 oC
(Al-Ibrahim and Varnham, 2010).
Al-Ibrahim and Varnham (2010) reported that refrigerative cooling can uses mechanical or electrical vapor
compression refrigeration equipment.
Jaber et al. (2006) studied the effect of air cooling intake on the gas turbine performance by comparing two
different cooling systems, evaporative and cooling coil. Their results showed that the evaporative cooling and
chiller system present similar improvement in the power output, about 1.0-1.5 MW, but the cooling coil of the

1158
mechanical chiller consumes more energy to run the vapor-compression refrigeration unit and the overall plant
performance decrease.
Alhazmy and Najjar (2004) compared two different techniques of air coolers, water spraying system and cooling
coils, and the results were analyzed for a specific set of operational and design conditions.
Nasser and El-Kalay (1991) suggested the use of a simple Li Br/water absorption system to cool the inlet air of
gas turbine in Bahrain. One of the most effective and useful methods to increasing the gas turbine’s output is air
intake cooling techniques, this method especially for hot region is completely effective and efficient (Al-Ibrahim and
Varnham, 2010).
Ana Paula Santos et al. (2012) showed the effect of cooling intake temperature on gas turbine power output. This
study showed the gas turbine power output increase due to decreasing the air intake temperature.
In this paper a cooling system with absorption chiller for gas turbine was stadied. The effect of cooling the intake
air is studied. Gas turbine in case-base (without compressor intake air cooling system) and new model with intake
air cooling system was simulated and analyzed. The results obtained in this model are compared with the value of
both two conditions.

GAS TURBINE CYCLE

Fig. 1, illustrates the schematic of simple gas turbine cycle. This model employed to simulate the air
thermodynamic procedure from 1 to 4. Pressure and temperature calculations for each point are also determined,
as shown in Fig. 1.

Fig.1. Simple Gas Turbine Cycle

-Compressor:
The compressor inlet temperature in typical gas turbine is equal to ambient temperature. Air has been considered as
ideal gas in all gas turbine cycle, also using the polytropic relation for ideal gas:

k 1
T2 ( )
 (rp ) k
(1)
T1

Where rp is compression ratio in compressor, k is specific heat ratio:

Cp
k (2)
Cv

Where Cp and Cv are specific heat at constant pressure and volume, respectively. Also pressure ratio:

P2
rp  (3)
P1
Using equation for efficiency in compressor, the output temperature (T2) can be calculated as:

k 1
T1
P
T2  [( 2 ) k
 1]  T1 (4)
c P1

Considering the first law of thermodynamic the compressor work estimated as follows:

1159
 . .
W  m a C pc,av (T2  T1 ) (5)

.
Where m a is the air mass flow rate and C pc,av is the specific heat of air at the constant pressure in average
temperature across the compressor (Alhamzy and Najjar, 2004).

-Combustor Chamber:
The heat delivered by combustor is determined from energy balance:

. .
Q  m a C pcc,av (T2  T1 ) (6)

Where C pcc,av is the flue gas specific heat at the combustion as function of the average value (Alhamzy and
Najjar, 2004). Also fuel gas consumption in combustor chamber is defined as:

.
. Qin
mf  (7)
LHV cc

Where  cc is the combustor chamber efficiency and LHV is fuel gas lower heat value.

-Turbine:
The discharge temperature of the exhausted gas that living the turbine is defined as:

k 1
1
T4  T3  tT4 [1  ( ) k
] (8)
P4
P3

Where t is the turbine isentropic efficiency and P4 is the turbine exhaust gas pressure in the last step of
typically gas turbine cycle. Therefore total power produced from the turbine is equal to:

. .
Wt  mT Ct ,av (T3  T4 ) (9)
. . .
Where mT is the total mass flow rate composed of fuel ( m f ) and air mass flow rate ( ma ).
Also Ct ,av is the fuel gas specific heat in average temperature through the turbine (Alhamzy and Najjar, 2004).
The net power obtained from the gas turbine cycle can be calculate by using Eq.(5) and Eq.(5):

. . .
Wn  Wt  Wc (10)

Lastly the thermal efficiency of gas turbine can be calculated as follow:

Wn
th  .
(11)
Qin
All equation was written for ideal Bryton gas turbine cycle. Equations for actual gas were written in the calculations
and for simulation of this study.

ABSORPTION CHILLER SYSTEM

1160
As mentioned before there are two basic methods to cool air inlet of compressor. First and most effective system
is evaporative cooling (Paula Santos et al, 2012). Evaporative coolers make use of evaporation of water to reduce
the gas turbine inlet air temperature. But condensing the humid air at the compressor stage is one of the
disadvantages of this method. Corrosion occurs on the compressor blades because of condensed water.
Therefore using this method is not appropriate in many vital industries.
The second system uses mechanical or absorption chillers to cool air intake compressor. In this method water cool
in chiller then flows through a heat exchanger located in the inlet duct to remove heat from the inlet air.
Fig. 2, shows the simple absorption chiller cycle. In current study temperature and pressure has been calculated
and simulated for all states (from 1 to 10).

Fig.2. Simple Absorption Chiller Cycle

Mass balance in Evaporator:

. . .
m9  m10  m (12)

Absorbed heat from evaporator can be calculated by Energy balance in Evaporator:

. . . . . .
Qe  m10 h10  m11 h11  m9 h9 (13)

mass balance in Absorber:

. . . .
m10  m11  m6  m1 (14)

Rejected heat from absorber can be calculated by Energy balance in absorber:

. . . . .
Q a  m10 h10  m11 h11  m6 h6  m1 h1 (15)

Energy balance in heat exchanger:

. . . .
m2 h2  m4 h4  m3 h3  m5 h5 (16)

Energy consumption in pump:

. .
W  m1 v1 ( p2  p1 ) (17)
Where p2 and p1 are the pressure of liquid after and before of solution pump respectively. Also h is enthalpy of
streams. Also In Eq.(16), assumes that specific volume of solution is constant.
Heat input in generator can be calculated by energy balance:

. . . .
Q gen  m4 h4  m7 h7  m3 h3 (18)

And heat rejected in condenser:

1161
 . .
Q cond  m7 (h7  h8 ) (19)

Finlay coefficient factor of absorption chiller can be estimated using the equation bellow:

.
Qe
COPChiller  .
(20)
Qgen
. .
Where Qe and Qgen are energy exchanged in Evaporator and Generator respectively.

PERFORMANCE ENHANCEMENT
As mentioned before there are different methods for enhancement of gas turbine performance, preheating the air
before combustion chamber by installing heat exchanger (recuperator), injecting steam or water to combustion
chamber due to increasing mass flow, cooling the air between to stage of compressor by installing after cooler are
various ways to rectifying the efficiency of gas turbine cycle.
The purpose of this project is to enhancement of gas turbine cycle, for this purpose an absorption chiller has used
to cool the inlet air. Also this chiller uses its energy from hot outlet exhaust gas. Absorption chillers can use steam
water as inlet energy, also steam produces by using a heat exchanger and using hot exhaust gas.
Fig. 3, illustrates the studied gas turbine cycle. This gas turbine was installed in Lavan Island in Iran. As shown in
Fig. 3, there are two steps compressor include low pressure compressor (Comp1) and high pressure compressor
(Comp2), and an intercooler between them. Intercooler reduces temperature of exit gas from first compressor step.
It helps to reduce second compressor’s work, therefore it increases the efficiency of gas turbine cycle.
Also this typical gas turbine has two steps turbine, high pressure turbine (Turb1) and low pressure turbine (Turb2),
and one combustion chamber before each of them. It helps to increase the temperature of exhaust gas in turbine
also increases the turbine’s power output. Increasing the turbine’s power output causes to enhancement the
efficiency of the gas turbine cycle.
There is a heat exchanger between hot gas outlet exhaust gas (stream 9) and compressed inlet air to first step of
combustion chamber (stream 4). It helps to reduce the fuel consumption by increasing the temperature of
combustion air inlet. Reducing the fuel consumption also increases the gas turbine cycle efficiency. This heat
exchanger herein named Heat Recovery Steam Generation (HRSG). It makes steam without consumption fuel.

Fig.3. Schematic of Gas Turbine Cycle (case study)

Climate properties in Lavan Island are presented in Table 1. Some properties of weather like average of
temperature, both wet and dry bulb temperature, dew point and altitude of gas turbine installed area is mentioned.
Also humidity and moisture of air should be considered in calculation and simulation.

Table 1. Weather properties of Lavan Island

Dry Relative Humidity Dew Specific
Wet bulb Altitud
bulb umidity ratio point olume
o o 3
o
c c m % kg/kg c m /kg
37 32 17 70-86 0.029 13 0.92

1162
Characteristic data of gas studied turbine is presented in table 2. Air flow rate, temperature and pressure of gas
stream are presented in table.

Table 2. Date for studied Gas turbine

Description unit value
1
Air flow (m /s) 31
LHV of Methane (kj/kg) 01131
First Combustor outlet temperature (c) 3011
Second Combustor outlet temperature (c) 3311
Temperature of the intercooler fluid (c) 00
Regenerator effectiveness (%) 01
Intercooler effectiveness (%) 01
First Compressor effectiveness (%) 50
Second Compressor effectiveness (%) 50
First Turbine effectiveness (%) 51
Second Turbine effectiveness (%) 51
First Combustor effectiveness (%) 50
Second Combustor effectiveness (%) 50

SIMULATION IN EES SOFTWARE

By means of calculate the equations and simulation of studied gas turbine as well as cooling system, simulation
and coding was performed in EES software (Engineering Equation Solver). Also the gas turbine cycle with air
intake cooling system has simulated in EES software. EES helps to simulate and analyze the complicated
engineering equation. Fig. 5 shows simulation window in EES software.

Fig.5. Gas Turbine Simulation’s window in EES software

Simulation in this software helps us to change the all parameters and see the results. Also we are able to draw many
graphs and tracing the parameters in comparison with others.

CONCLUSIONS
This paper deals with cooling the air intake of compressor to increases the gas turbine efficiency. By means of
cooling the compressor air intake, an absorption chiller has used before compressor intake duct. High temperature
exhausted gas from gas turbine recovered by a heat recovery exchanger and produce steam without consumption
excess fuel. This steam feeds absorption chiller to make chilled water, chilled water cools the compressor intake
air. This complex finally cools the compressor intake air due to improvement the efficiency of gas turbine.

1163
Calculation and simulation shows that the variation of the air inlet temperature from 38 oC to 15oC generates a
reduction of 11.36 % in the compressor work. Also this leads to reduce fuel consumption in combustion chamber
by 6.4%. Due to reduction the fuel consumption, the efficiency of cycle decrease (with considering constant power
of turbine).
Results for power output, fuel consumption and gas turbine efficiency are also obtained for typically gas turbine
without inlet cooling system and the gas turbine with compressor air intake cooling system
This model has simulated in EES software. Simulation of model helps us to change the variety of inlet temperature
and other parameters. Thus, this method leads to increase the thermal efficiency by 7% for real site condition.

Acknowledgement
Authors are grateful to NIGC (National Iranian Gas Company) by the technical and financial support for the
development of this work.

NOMENCLATURE
LHV lower fuel heat value, J/kg
h specific enthalpy, kJ/kg
ṁ mass flow rate, kg/s
Ԛ heat transfer rate, kW
rp compressor pressure ratio
T temperature, C
Ẇ power output, kW
k specific heat ratio, kJ/kgC
P pressure, kPa
EES Engineering Equation Solver software
COP coefficient of performance of absorption chiller
Greek Letters
η efficiency, %

Subscripts
p pressure
t turbine
v volume
c compressor
th thermal
e evaporator
cond condenser
gen generator
f fuel

REFERENCES
Journal: Omar Othman Badran. 1999. Gas turbine performance improvement. International Journal of Applied
Energy 64:263-273.

Journal: Ana Paula Santos. 2012. Analysis of Gas turbine performance with inlet air cooling techniques applied to
Brazilian sites. International Journal of Applied Thermal Engineering 30: 263-273.

Journal: Al-Ibrahim, A. Varnham. 2010. A review of inlet air-cooling technologies for enhancing the performance of
combustion turbines in Saudi Arabian. Applied Thermal Engineering 30:1879-1888.

Journal: Jaber, Q. M. et al. 2007, Assessment of power augmentation from gas turbine power plants using different
inlet air cooling systems. Jordan Journal of Mechanical and Industrial Engineering 1:7-15.

Journal: Najjar, Y. S. H. 1996. Enhancement of performance of gas turbine engines by inlet air cooling and
cogeneration systems. Applied Thermal Engineering 16:163-173.

Book: H. Cohen. and G.F. Rogers. and H. Saravanamutto. 1987. Gas Turbine Theory. Henry Ling Ltd.

Book: V. Ganapathy. 2003. Industrial boilers and heat recovery steam generators. Florida: CRC Press.

1164
 A STUDY ON SINGLE LAYER DRYING CURVE EQUATIONS
a,* b a
H. Kucuk A.Killic and A. Midilli
a
Energy Division, Department of Mechanical Engineering, Faculty of Engineering, Recep Tayyip Erdoğan University,
53100 Rize, Turkey
b
Department of Food Engineering, Faculty of Engineering, Recep Tayyip Erdoğan University, 53100 Rize, Turkey
haydar.kucuk@erdogan.edu.tr, aydin.kilic@erdogan.edu.tr, adnan.midilli@erdogan.edu.tr

ABSTRACT
In this study, single layer drying curve equations commonly used in the literature between 2003-2013 are
systematically are discussed and evaluated in terms of their applications and selection for single layer drying
processes by considering the model evaluation criteria. As a result of this study, it is noticed that serious
complications, confusions and conflicts in the applications of single layer drying curve equations and their evaluation
criteria. Consequently, it is recommended that the drying curve equations should be applied in accordance with the
forms commonly used in the literature. Also, it is determined that the following drying curve equations give the best
results for single layer drying processes which are the Midilli-Kucuk, Page, Logarithmic, Two-term, Wang and Singh,
Approximation of diffusion, Modified Henderson and Pabis, Modified Page, Henderson and Pabis, Two-term
exponential and Verma et al.

Keywords: Single layer drying curve equation, Midilli-Kucuk model, Mathematical modeling

INTRODUCTION
Thin layer equations have been used to estimate drying times of several products and to generalize drying curves. In
the development of thin layer drying models for agricultural products, generally the moisture content of the material at
any time after it has been subjected to a constant relative humidity and temperature conditions is measured and
correlated to the drying parameters (Midilli et al., 2002). The thin layer drying models can classify as theoretical, semi-
theoretical and empirical. The most widely used theoretical models are derived from Fick’s second law of diffusion.
Similarly, semi-theoretical models are generally derived from Fick’s second law and modifications of its simplified
forms (other semi theoretical models are derived by analogues with Newton’s law of cooling) (Erbay and Icier, 2010).
The empirical models have also similar characteristics with semi-theoretical models. They strongly depend on the
experimental conditions and give limited information about the drying behaviors of the product (Erbay and Icier,
2010). The empirical method is based on experimental data and dimensional analysis. They are easily applied to
drying simulation as they depend on experimental data (Midilli et al., 2002). The theoretical method takes into
account not only the external conditions, but also the mechanism of internal movement of moisture and their
consequent effects (Yücel, 2006). The Semi-theoretical and emprical models consider only the external resistance to
moisture transfer between the product and air (Akpinar et al., 2004).

THIN LAYER DRYING CURVE EQUATIONS

Researchers have developed numerous thin layer models given as below for various agricultural products depending
on product type, pretreatment of the product, drying parameters and drying methods.

Newton (Lewis, Exponential, Single exponential) Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et
al., 2004; Toğrul, 2005; Yurtsever, 2005; Duan et al., 2005; Menges and Ertekin, 2006; Sacilcik and Elicin, 2006;
Yücel, 2006; Ghazanfari et al., 2006; McMinn et al., 2007; Cihan et al., 2007; Özbek and Dadali, 2007; Sobukola et
al., 2007; Vega-Galvez et al., 2008; Mohamed et al., 2008; Ganesapillai et al. 2008; Roberts et al., 2008; Arslan and
Özcan, 2008; Abbasi et al., 2009; Aghbashlo et al., 2009; Ong and Law, 2009; Chin et al., 2009; Pin et al., 2009;
Zomorodian and Moradi, 2010; Jazini and Hatamipour, 2010; Saeed, 2010; Erbay and İcier, 2010; Demirhan and
Özbek, 2010; Sarimeseli, 2011; Corzo et al., 2011; Liu et al., 2011; Rasouli et al., 2011; Evin, 2012; Doymaz, 2012;
Ashraf et al., 2012; Bingol et al., 2012; Khanali et al., 2012; Balbay and Şahin, 2012; Sharifian et al., 2012; Kaleta et
al., 2013; Ruiz et al., 2013; Tee et al., 2013; Abano et al., 2013; Pillai, 2013)

MR  exp  kt )  (1)
Page Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et al., 2004; Toğrul, 2005; Yurtsever, 2005;
Duan et al., 2005; Menges and Ertekin, 2006; Sacilcik and Elicin, 2006; Yücel, 2006; Ghazanfari et al., 2006; McMinn
et al., 2007; Cihan et al., 2007; Özbek and Dadali, 2007; Sobukola et al., 2007; Vega-Galvez et al., 2008; Mohamed
et al., 2008; Ganesapillai et al. 2008; Roberts et al., 2008; Abbasi et al., 2009; Aghbashlo et al., 2009; Ong and Law,
2009; Chin et al., 2009; Pin et al., 2009; Zomorodian and Moradi, 2010; Saeed, 2010; Diamante et al., 2010; Erbay
and İcier, 2010; Demirhan and Özbek, 2010; Sarimeseli, 2011; Corzo et al., 2011; Hemis et al., 2011; Rasouli et al.,
2011; Evin, 2012; Doymaz, 2012; Ashraf et al., 2012; Bingol et al., 2012; Khanali et al., 2012; Balbay and Şahin,
2012; Sharifian et al., 2012; Kaleta et al., 2013; Tee et al., 2013; Abano et al., 2013; Pillai, 2013)

MR  exp  kt n  (2)
Modified Page (Modified Page-I, Overhults et al.) Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Toğrul,
2005; Duan et al., 2005; Vega-Galvez et al., 2008; Saeed, 2010; Jazini and Hatamipour, 2010; Evin, 2012; Khanali et
al., 2012; Sharifian et al., 2012; Kaleta et al., 2013; Tee et al., 2013; Abano et al., 2013)

1165
 
MR  exp  kt
n
 (3)
Modified Page-II Model (Toğrul, 2005)
  t n 
MR exp   c 2   (4)
 L  
 
Henderson and Pabis (Single term, Generalized exponential) Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004;
Lahsasni et al., 2004; Toğrul, 2005; Yurtsever, 2005; Duan et al., 2005; Menges and Ertekin, 2006; Yücel, 2006;
Sacilcik and Elicin, 2006; McMinn et al., 2007; Cihan et al., 2007; Özbek and Dadali, 2007; Hayaloglu et al., 2007;
Sobukola et al., 2007; Mohamed et al., 2008; Ganesapillai et al. 2008; Roberts et al., 2008; Arslan and Özcan, 2008;
Abbasi et al., 2009; Aghbashlo et al., 2009; Ong and Law, 2009; Pin et al., 2009; Amira et al., 2010; Zomorodian and
Moradi, 2010; Saeed, 2010; Diamante et al., 2010; Jazini and Hatamipour, 2010; Erbay and İcier, 2010; Sarimeseli,
2011; Rasouli et al., 2011; Evin, 2012; Doymaz, 2012; Bingol et al., 2012; Khanali et al., 2012; Balbay and Şahin,
2012; Sharifian et al., 2012; Kaleta et al., 2013; Ruiz et al., 2013; Tee et al., 2013; Abano et al., 2013; Pillai, 2013)
MR  a exp  kt   (5)
Logarithmic (Asymptotic, Yagcioglu et al.) Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et al.,
2004; Toğrul, 2005; Yurtsever, 2005; Menges and Ertekin, 2006; Sacilcik and Elicin, 2006; Yücel, 2006; McMinn et al.,
2007; Özbek and Dadali, 2007; Hayaloglu et al., 2007; Sobukola et al., 2007; Mohamed et al., 2008; Ganesapillai et
al. 2008; Arslan and Özcan, 2008; Abbasi et al., 2009; Aghbashlo et al., 2009; Ong and Law, 2009; Pin et al., 2009;
Amira et al., 2010; Zomorodian and Moradi, 2010; Jazini and Hatamipour, 2010; Erbay and İcier, 2010; Corzo et al.,
2011; Liu et al., 2011; Rasouli et al., 2011; Doymaz, 2012; Bingol et al., 2012; Khanali et al., 2012; Sharifian et al.,
2012; Ruiz et al., 2013; Tee et al., 2013; Abano et al., 2013; Pillai, 2013)
MR  a exp  kt  c (6)
Midilli-Kucuk (Midilli, Midilli et al.) Model (Midilli et al., 2002; Akpinar et al., 2004; Lahsasni et al., 2004; Toğrul, 2005;
Yurtsever, 2005; Menges and Ertekin, 2006; Yücel, 2006; Sacilcik and Elicin, 2006; McMinn et al., 2007; Cihan et al.,
2007; Özbek and Dadali, 2007; Hayaloglu et al., 2007; Sobukola et al., 2007; Vega-Galvez et al., 2008; Mohamed et
al., 2008; Ganesapillai et al. 2008; Arslan and Özcan, 2008; Abbasi et al., 2009; Aghbashlo et al., 2009; Ong and
Law, 2009; Pin et al., 2009; Amira et al., 2010; Zomorodian and Moradi, 2010; Demirhan and Özbek, 2010; Jazini
and Hatamipour, 2010; Sarimeseli, 2011; Corzo et al., 2011; Rasouli et al., 2011; Doymaz, 2012; Ashraf et al., 2012;
Bingol et al., 2012; Khanali et al., 2012; Balbay and Şahin, 2012; Ruiz et al., 2013; Tee et al., 2013; Abano et al.,
2013)

MR  a exp  kt n  bt  (7)
Modified Midilli et al. (Ghazanfari et al.) Model (Ghazanfari et al., 2006)

MR  exp  kt n  bt  (8)
Abbasi et al. (Modified Midilli-Kucuk) Model (Abbasi et al., 2009)

MR  a exp  kt n  b  (9)
Demir et al. Model (Aghbashlo et al., 2009; Erbay and İcier, 2010; Kaleta et al., 2013)
MR  a exp  kt  b
n
(10)
Modified Henderson and Perry (Agrawal and Singh) Model (Phoungchandang and Kongpim, 2012)
MR  a exp  kt n  (11)
Three Parameter Model (Phoungchandang and Kongpim, 2012)
MR  a exp  kt n
 (12)
Two-Term Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et al., 2004; Sacilcik and Elicin, 2006;
Hayaloglu et al., 2007; Mohamed et al., 2008; Arslan and Özcan, 2008; Abbasi et al., 2009; Aghbashlo et al., 2009;
Ong and Law, 2009; Zomorodian and Moradi, 2010; Saeed, 2010; Jazini and Hatamipour, 2010; Corzo et al., 2011;
Evin, 2012; Tee et al., 2013; Abano et al., 2013)
MR  a exp  k1t   b exp  k 2 t  (13)
Two-Term Exponential Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et al., 2004; Yurtsever, 2005;
Duan et al., 2005; Menges and Ertekin, 2006; Yücel, 2006; Sacilcik and Elicin, 2006; Hayaloglu et al., 2007;
Mohamed et al., 2008; Ganesapillai et al. 2008; Abbasi et al., 2009; Aghbashlo et al., 2009; Zomorodian and Moradi,
2010; Saeed, 2010; Erbay and İcier, 2010; Corzo et al., 2011; Khanali et al., 2012; Ashraf et al., 2012; Evin, 2012;
Ruiz et al., 2013; Tee et al., 2013; Abano et al., 2013; Pillai, 2013)
MR  a exp  kt  1  a  exp  kat (14)
Verma et al. (Modified Two-Term Exponential) Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et al.,
2004; Yurtsever, 2005; Menges and Ertekin, 2006; Yücel, 2006; Hayaloglu et al., 2007; Mohamed et al., 2008;
Ganesapillai et al. 2008; Abbasi et al., 2009; Saeed, 2010; Erbay and İcier, 2010; Corzo et al., 2011; Rasouli et al.,
2011; Evin, 2012; Liu et al., 2011; Khanali et al., 2012; Ashraf et al., 2012; Sharifian et al., 2012; Kaleta et al., 2013;
Ruiz et al., 2013; Abano et al., 2013; Pillai, 2013)
MR  a exp  kt  1  a  exp  gt  (15)

1166
Approximation of Diffusion (Diffusion Approach) Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et
al., 2004; Toğrul, 2005; Yurtsever, 2005; Menges and Ertekin, 2006; Yücel, 2006; Sacilcik and Elicin, 2006; McMinn et
al., 2007; Cihan et al., 2007; Sobukola et al., 2007; Mohamed et al., 2008; Ganesapillai et al. 2008; Abbasi et al.,
2009; Aghbashlo et al., 2009; Zomorodian and Moradi, 2010; Saeed, 2010; Jazini and Hatamipour, 2010; Erbay and
İcier, 2010; Rasouli et al., 2011; Bingol et al., 2012; Khanali et al., 2012; Balbay and Şahin, 2012; Sharifian et al.,
2012; Ruiz et al., 2013; Abano et al., 2013; Pillai, 2013)
MR  a exp  kt  1  a  exp  kbt (16)
Modified Henderson and Pabis (Three Term Exponential) Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004;
Lahsasni et al., 2004; Menges and Ertekin, 2006; Yücel, 2006; Sacilcik and Elicin, 2006; McMinn et al., 2007;
Mohamed et al., 2008; Abbasi et al., 2009; Aghbashlo et al., 2009; Saeed, 2010; Jazini and Hatamipour, 2010; Erbay
and İcier, 2010; Khanali et al., 2012; Ashraf et al., 2012; Evin, 2012; Ruiz et al., 2013; Abano et al., 2013)
MR  a exp  kt  b exp  gt   c exp  ht  (17)
Thompson Model (Akpinar et al., 2004; Menges and Ertekin, 2006; Zomorodian and Moradi, 2010; Erbay and İcier,
2010; Evin, 2012)
t  aln MR  bln MR
2
(18)
Wang and Singh Model (Toğrul and Pehlivan, 2004; Akpinar et al., 2004; Lahsasni et al., 2004; Toğrul, 2005;
Yurtsever, 2005; Menges and Ertekin, 2006; Yücel, 2006; Ghazanfari et al., 2006; McMinn et al., 2007; Hayaloglu et
al., 2007; Özbek and Dadali, 2007; Mohamed et al., 2008; Ganesapillai et al. 2008; Abbasi et al., 2009; Aghbashlo et
al., 2009; Ong and Law, 2009; Amira et al., 2010; Zomorodian and Moradi, 2010; Erbay and İcier, 2010; Demirhan
and Özbek, 2010; Jazini and Hatamipour, 2010; Sarimeseli, 2011; Corzo et al., 2011; Ashraf et al., 2012; Evin, 2012;
Bingol et al., 2012; Khanali et al., 2012; Ruiz et al., 2013; Tee et al., 2013; Abano et al., 2013; Pillai, 2013)
MR  1  at  bt 2 (19)
Kaleemullah Model (Erbay and İcier, 2010)
MR  exp cT   b t  pT  n 
(20)
Diamente et al. Model (Diamante et al., 2010)
ln  ln MR  a  bln t   cln t 
2
(21)
Hii et al. Model (Kaleta et al., 2013)
 
MR  a exp  kt n  c exp  gt n  
(22)
Simplified Fick’s diffusion (SFFD) Model (Toğrul, 2005; Evin, 2012; Ruiz et al., 2013)
  t 
MR  a exp   c 2   (23)
  L 
Weibull Model (Aghbashlo et al., 2009)
  t b 
MR  exp      (24)
 a 
 
Modified Drying Model (Hemis et al., 2011)
MR  a  exp  kt n   (25)
Aghbashlo et al. Model (Aghbashlo et al., 2009)
 kt 
MR exp   1  (26)
 1  k 2t 
Seiiedlou and Aghbashlo Model (Rasouli et al., 2011)
 kt 
MR exp   1  (27)
 1  k 2t 
Decay Model (Montazer-Rahmati and Amini-Horri, 2005)
 k 
MR exp  1  (28)
 1  k 2 
Chavez-Mendez et al. Model (McMinn et al., 2007)
 1 
MR  1  1  L2 L1t  1 L2 
 
(29)
Geometric Model (Cihan et al., 2007)
MR  a t  n (30)
Parabolic (Polynomial, Quadratic) Model (Abano et al., 2013)

1167
MR  a  bt  ct 2 (31)
Logistic Model (Yurtsever, 2005)
b
MR  (32)
1 a exp kt
Multiple Multiplicative Factor (MMF) Model (Amira et al., 2010)

MR 
a  b  c  t  d

b  t  d
(33)

Noomhorn and Verma (Ranjbaran and Zare, Modified Two Term) Model (Kaleta et al., 2013)
MR  a exp  k1t   b exp  k 2 t   c (34)
Akbulut and Durmuş Model (Akbulut and Durmuş, 2009)
m1 s 1
 T  Tdo   bw  bd  / 2 
MR  a  di n 
 exp  kt 
Vc
  
   (35)
 Tco  Tci    Li  Li 1  / t   Lw  Ld  / 2 
Simplified Fick’s Second Law of Diffusion Model (Wiktor et al., 2013)

8   2 Deff  
MR  2 exp   
  4 L2 
 
(36)

Exponential-hyperbolic Decay (Montazer-Rahmati and Amini-Horri) Model (Montazer-Rahmati and Amini-Horri, 2005)
 k t / t 0  
MR exp   1  (37)
 1  k 2 t / t 0  
Third-degree Polynomial (Cubic) Model (Liu et al., 2011)
MR  a  bt  ct 2  kt 3 (38)
Das et al. (Jena-Das) Model (Ashraf et al., 2012)

MR  a exp  kt  b t  c  (39)
Linear Model (Sharifian et al., 2012)
MR  at  b (40)
Quasi-stationary Model (Khanali et al., 2012)
1
MR 
1  t / x  y
(41)

Balbay and Şahin Model (Balbay and Şahin; 2012)

MR  1  a exp  kt n  b  
(42)
Alibas (Modified Midilli-Kucuk) Model (Alibas, 2012)

M R  a exp  kt n  bt  g  (43)
Polynomial (Meda et al.) Model (Meda et al., 2008)
MR  a  bt 
2
(44)
Exponential (Meda et al.) Model (Meda et al., 2008)

MR exp m  nt 1.5  (45)
Jittanit (Modified Page) Model (Le and Jittanit, 2011)
  A 

MR  exp   Kt N exp     (46)
  Tk  
Jittanit (Modified two-term) Model (Le and Jittanit, 2011)
  B    B 
MR  A1 exp   K1t  exp      A1 exp   K 2 t  exp  
 
 
 (47)
  Tk    Tk 
Regression Model (Costa and Pereira, 2013)
 
M R  exp  at 2  bt  (48)
Yun et al. Model (Tee et al., 2013)

1168
 a  bt  ct 2
MR  (49)
1  dt  ft 2
Kaleta et al. (Modified Verma et al.) Model (Kaleta et al., 2013)
 
MR  a exp  kt n  1  a  exp  gt n   (50)
Ademiluyi Model (Kaleta et al., 2013)
 
MR  a exp  kt
n
(51)
Kaleta et al. (Modified two-term) Model (Kaleta et al., 2013)
MR  a exp  kt   b exp  gt  
n n
(52)
Jaros and Pabis-I Model (Kaleta et al., 2013)
 1   
N
b 
M  Mo
1 b 0.85
1  kt   , b (53)
 1  b  NM o  1 b  1 Mo
 
Jaros and Pabis-II Model (Kaleta et al., 2013)
  k t  t   1 b  
N 1

M  M e  M c  M e  exp  c
1  kt  (54)
 M c  M e  NM o c  
 
Henderson et al. Model (dos Santos, 2007)
 1 
X r  c  exp  Kt     exp  9 Kt  (55)
 9 
Law et al. (New Variable Diffusion Drying) Model (Law et al., 2003)

X  X eq   exp c bc dX dt   

X  X cr 2  ln 1 
cr 2 cr 2
t (56)
c  X cr 2  X eq 
 
Zero Model (Phoungchandang and Kongpim, 2012)
X  X o exp  Kt  (57)
Motta Lima et al.-I Model (dos Santos, 2007)
X r  a  bT 
2
(58)
Motta Lima et al.-II Model (dos Santos, 2007)

  e
X r  a  at  ct 2  dt 3 exp     (59)
 T
Pillai (Pillai, 2013)

MR  a exp  aIMC  MWP SA DT 
b c d
 (60)

EVALUATİON CRİTERİA
There are many evaluation criteria used to choose the best thin-layer drying curve equation for different products,
drying methods and drying conditions.

Correlation coefficient (Cihan et al., 2007; Chin et al., 2009; Erbay and Icier, 2010)
 N  N 
N  MR pre,i MRexp,i     MR pre,i   MRexp,i 
N

r i 1  i 1  i 1  (61)
 2
 2

 N MR 2   MR   N MR 2   MR 
N N N N

   pre,i     exp,i  
pre,i
 i 1
exp,i
 i 1 
 i 1  i 1 
The regression sum of squares (Jazini and Hatamipour, 2010)

SSR   MRcal,i  MRavg 

n
2
(62)
1
The total sum of squares (Jazini and Hatamipour, 2010)
2

SST   MRexp,i  MRavg 

n
(63)
1
The error (residual) sum of squares (Saeed, 2010)

1169
 2

SSE   MRexp,i  MRcal,i 

N
(64)
i 1
Residual sum of squares (McMinn et al., 2007; Özbek and Dadali, 2007; Ganesapillai et al. 2008; Pillai, 2013;
Tahmasebi et al., 2013)

RSS   MRexp,i  MR pre,i 

N
2
(65)
i 1
Coefficient of determination (Saeed, 2010; Jazini and Hatamipour, 2010)
SSR SSE
R2   1  (66)
SST SST
The adjusted R2 (Menges and Ertekin, 2006)

R 2  1 1 R2   N N k 1 1 (67)

The reduced chi-square (Toğrul and Pehlivan, 2004; Lahsasni et al., 2004; Toğrul, 2005; Menges and Ertekin, 2006;
Sacilik and Elicin, 2006; Sobukola et al., 2007; Mohamed et al., 2008; Diamante et al., 2010; Erbay and İcier, 2010;
Evin, 2012; Ashraf et al., 2012; Pillai, 2013; Tahmasebi et al., 2013)

 MR  MR pre,i 
N
2
exp,i
2  i 1
(68)
N n
The root mean square error (Toğrul and Pehlivan, 2004; Toğrul, 2005; yurtsever, 2005; Menges and Ertekin, 2006;
Cihan et al., 2007; Hayaloglu et al., 2007; Sobukola et al., 2007; Akbulut and Durmuş, 2009; Chin et al., 2009; Erbay
and İcier, 2010; Evin, 2012; Rasouli et al., 2011; Doymaz, 2012; Ashraf et al., 2012; Khanali et al., 2012; Tee et al.,
2013; Ruiz et al., 2013)

 MR  MRexp,i 
N
2
pre,i
RMSE  i 1
(69)
N
Residuals (Ruiz et al., 2013)

residuals   MRexp,i  MR pre,i 

N
(70)
i 1
Modeling efficiency (Toğrul, 2005)

 MR  MRexp,ave    MR pre,i  MRexp,i 

N N
2 2
exp,i
EF  i 1 i 1
(71)

 MR  MRexp,ave 
N
2
exp,i
i 1
The standard error of estimate (Saeed, 2010)

 MR  MRcal,i 
N
2
exp,i
SEE  i 1
(72)
N  np
Variance (s2) (Standard error (2)) (Özbek and Dadali, 2007)
 N 2
  MRexp,i  MR pre,i  
 i 1  min
 2  s2  (73)
N n
The reduced sum square error (Saeed, 2010)

 MR  MRcal,i 
N
2
exp,i
RSSE  i 1
(74)
N
The mean bias error (Toğrul and Pehlivan, 2004; Toğrul, 2005; Hayaloglu et al., 2007; Chin et al., 2009)

 MR  MRexp,i 
N

pre,i
MBE  i 1
(75)
N
Mean relative percentage error (Sacilik and Elicin, 2006)

1170
 100 N MRexp,i  MR pre,i
P  MR
N i 1
(76)
exp,i
Mean absulate error (Nazghelichi et al., 2011)
N
1
MAE 
N
x
i 1
pi  xdi 77)

t-value (Hayaloglu et al., 2007)

t  value 
n  1MBE 2 (78)
RMSE 2  MBE 2
Degrees of freedom (Gunhan et al., 2005)

 
 2
exp  
/ nexp   2pre / n pre  2

(79)
 2
exp / nexp   
2 2
pre / n pre 2

nexp  1 n pre  1
F-value (Tahmasebi et al., 2013)
MS treatment
F  value  (80)
MS error

RESULTS AND DISCUSSION

As a result of literature review, it is observed that sixty single layer drying curve equations have been derived by
researchers by taking in to consideration the following parameters: i) product type, ii) pretreatment of the product, iii)
drying parameters such as temperature, air velocity, layer thickness, micro-wave power levels, amount of solar
radiation, vacuum pressure, frequency of the sound wave, excitation amplitude, relative humidity, bed dept, shape of
product, pH, salt content, absolute pressures iv) drying methods. On the other hand, twenty evaluation criteria have
been defined to evaluate the performance of the single layer drying curve equation. It is noticed that serious
complications, confusions and conflicts in the applications of single layer drying curve equations and their evaluation
criteria such as name, equation and nomenclature. In this study, the drying curve equations have been given in
accordance with the forms commonly used in the literature. It is realized that Newton (Lewis, Exponential, Single
exponential) (Eq.1), Page (Eq.2), Henderson and Pabis (Single term, Generalized exponential) (Eq.5), Logarithmic
(Asymptotic, Yagcioglu et al.) (Eq.6), Midilli-Kucuk (Midilli, Midilli et al.) (Eq.7), Two-Term (Eq.13), Two-Term
Exponential (Eq.14), Verma et al. (Modified Two-Term Exponential) (Eq.15), Approximation of Diffusion (Diffusion
Approach) (Eq.16), Modified Henderson and Pabis (Three Term Exponential) (Eq.17), Wang and Singh (Eq.19),
Modified Page (Modified Page-I, Overhults et al.) (Eq.3), Thompson Model (Eq.18) and Simplified Fick’s diffusion
(SFFD) (Eq.23) models are commonly used in the literature. The others single layer drying curve equations have
been rarely used by the researches. Also, It is noticed that, coefficient of determination (R2) (Eq.66), the reduced chi-
2
) (Eq.68), the root mean square error (RMSE) (Eq.69), the mean bias error (MBE) (Eq.75), mean relative
percentage error (P) (Eq.76), residual sum of squares (RSS) (Eq.65), modeling efficiency (EF) (Eq.71) and the
standart error of estimate (SEE)(Eq.72) are commonly used to evaluate single layer drying models in the literature.
However, it is determined that Midilli-Kucuk, Page, Logarithmic, Two-term, Wang and Singh, Approximation of
diffusion, Modified Henderson and Pabis, Modified Page, Henderson and Pabis, Two-term exponential, Verma et al.,
and the others give best results, respectively, for different drying methods, drying parameters and dried products.

CONCLUSIONS
Single layer drying curve equations are commonly used in the literature to estimate drying behavior of the different
products depending on pretreatment of the product, drying parameters and drying methods. It is realized that there is
a complexity both drying methods and evaluation criteria such as name, equation and nomenclature. It is clear the
standardization is necessary for single layer drying curve equations and their evaluation criteria. The comprehensive
review of single layer drying curve models seems that the Midilli-Kucuk, Page, Logarithmic, Two-term, Wang and
Singh, Approximation of diffusion, Modified Henderson and Pabis, Modified Page, Henderson and Pabis, Two-term
exponential and Verma et al., respectively.

NOMENCLATURE
A,A1,a,B,b,c,d,
e,f,g,h,K,K1,K2, drying constants
k,k1,k2,L,L1,L2,m,
n,N,p,s,x,y
b sample length, m (Eq.35)
D moisture diffusion coefficient, m2/s
DT drying time, min

1171
EF modeling efficiency
IMC moisture content
M moisture content (g water/g dry solids), % dry basis
MAE mean absulate error
MBE mean bias error
MR,x,Xr moisture ratio
MS mean squares
MW microwave
MWP microwave output power, W
t drying time, min
L sample thickness (Eq.35), half-thickness (Eq.36), m
N,n number of observations
n,np number of constants
P mean relative percentage error
r correlation coefficient
R2 coefficient of determination
RMSE root mean square error
RSS residual sum of squares
RSSE reduced sum square error
s2 variance
SA surface area, m2
SEE standard error of estimate
SSE error (residual) sum of squares
SSR regression sum of squares
SST total sum of squares
T drying air temperature, oC
Tk drying temperature, K
V drying air velocity, m/s
X moisture content (g water/g dry solids), % dry basis
Greek Letters
2
reduced chi-square
 degrees of freedom
 standard deviation
2 standard error
 drying time, min
Subscripts
ave, avg average
c critical
cal calculated
cr1 first critical
cr2 second critical
ci collector in
ci collector out
d dry
di dryer inlet
do dryer outlet
e equilibrium
eff effective
eq equilibrium state
exp experimental
0 initial
pre predicted
w wet

REFERENCES
Abbasi, S., S. M. Mousavi and M. Mohebbi. 2009. Propose suitable model for modeling of moisture ratio and
estimation of effective moisture diffusivity of onion slices by hot air dryer. The 6th International Chemical Engineering
Congress and Exhibition Kish Island, Iran.
Abano, E .E., M. A. Haile, J. Owusu and F. N. Engmann. 2013. Microwave-vacuum drying effect on drying kinetics,
lycopene and ascorbic acid content of tomato slices. Journal of Stored Products and Postharvest Research 4(1):11-
22.
Aghbashlo, M., M. H. Kianmehr, S. Khani and M. Ghasemi. 2009. Mathematical modelling of thin-layer drying of
carrot. International Agrophysics 23(4):313-317.

1172
Akbulut, A. and A. Durmuş. 2009. Thin layer solar drying and mathematical modeling of mulberry. International
Journal of Energy Research 33(7):687-695.
Akpinar, E.K, C. Sarsılmaz and C. Yildiz. 2004. Mathematical modelling of a thin layer drying of apricots in a solar
energized rotary dryer. International Journal of Energy Research 28(8):739-752.
Alibas, I. 2012. Selection of a the best suitable thin-layer drying mathematical model for vacuum dried red chili
pepper. Journal of Biological and Environmental Sciences 6(17):161-170.
Amira, T., C. Saber and Z. Fethi. 2010. Modelling of the drying kinetics of opuntia ficus indica fruits and cladodes.
International Journal of Food Engineering 6((2)11):1-17. http://www.bepress.com/ijfe/vol6/iss2/art11.
Arslan, D. and M. M. Özcan. 2008. Evaluation of drying methods with respect to drying kinetics, mineral content and
colour characteristics of rosemary leaves. Energy Conversion and Management 49(5):1258-1264.
Ashraf, Z., Z. Hamidi-Esfahani and M. A. Sahari. 2012. Evaluation and characterization of vacuum drying of date
paste. Journal of Agricultural Science and Technology 14:565-575.
Balbay, A. and Ö. Şahin. 2012. Microwave drying kinetics of a thin-layer liquorice root. Drying Technology 30(8):859-
864.
Bingol, G., J. S. Roberts, M. O. Balaban and Y. O. Devres. 2012. Effect of dipping temperature and dipping time on
drying rate and color change of grapes. Drying Technology 30(6):597-606.
Chin, S. K., C. L. Law, C. V. Supramaniam and P. G. Cheng. 2009. Thin-layer drying characteristics and quality
evaluation of air-dried Ganoderma tsugae Murrill. Drying Technology 27(9):975-984.
Cihan, A., K. Kahveci and O. Hacıhafızoğlu. 2007. Modelling of intermittent drying of thin layer rough rice. Journal of
Food Engineering 78(1):293-298.
Corzo, O., N. Bracho and C. Alvarez. 2011. Determination of suitable thin layer model for air drying of mango slices
(Mangifera indica L.) at different air temperatures and velocities. Journal of Food Process Engineering 34(2):332-350.
Costa, A. and H. Pereira. 2013. Drying kinetics of cork planks in a cork pile in the field. Food and Bioproducts
Processing 91(1):14-22.
Demirhan, E. and B. Özbek. 2010. Drying kinetics and effective moisture diffusivity of purslane undergoing
microwave heat treatment. Korean Journal of Chemical Engineering 27(5):1377-1383.
Diamante, L. M., R. Ihns, G. P. Savage and L. Vanhanen. 2010. A new mathematical model for thin layer drying of
fruits. International Journal of Food Science and Technology 45(9):1956–1962.
dos Santos, J. A. B. 2007. Análise Comparativa Entre Técnicas de Processamentos Para Extração de Pigmentos
nas Sementes de Urucum. Ph.D. thesis, Federal University of Campina Grande, Campina, Paraíba, Brazil [in
Portuguese].
Doymaz, İ. 2012. Evaluation of some thin-layer drying models of persimmon slices (Diospyros kaki L.). Energy
Conversion and Management 56:199–205.
Duan, Z., M. Zhang, Q. Hu and J. Sun. 2005. Characteristics of microwave drying of bighead carp. Drying
Technology 23(3):637-643.
Erbay, Z. and F. Icier. 2010. A review of thin layer drying of foods: Theory, modeling, and experimental results, Critical
Reviews in Food Science and Nutrition 50(5):441-464.
Evin, D. 2012. Thin layer drying kinetics of Gundelia tournefortii L. Food and Bioproducts Processing 90(2):323–332.
Ganesapillai, M., I. Regupathi and T. Murugesan. 2008. An empirical model for the estimation of moisture ratio during
microwave drying of plaster of paris. Drying Technology 26(7):963–978.
Ghazanfari, A., S. Emami, L. G. Tabil and S. Panigrahi. Thin-layer drying of flax fiber: II. modeling drying process
using semi-theoretical and empirical models. Drying Technology 24(12):1637-1642.
Gunhan, T., V. Demir, E. Hancioglu and A. Hepbasli. 2005. Mathematical modeling of drying of bay leaves. Energy
Conversion and Management 46(11-12):1667-1679.
Hayaloglu, A. A., I. Karabulut, M. Alpaslan and G. Kelbaliyev. 2007. Mathematical modeling of drying characteristics
of strained yoghurt in a convective type tray-dryer. Journal of Food Engineering 78(1):109-117.
Hemis, M., C. B. Singh and D. S. Jayas. 2011. Microwave-assisted thin layer drying of wheat. Drying Technology
29(10):1240-1247.
Jazini, M. H. and M. S. Hatamipour. 2010. A new physical pretreatment of plum for drying. Food and Bioproducts
Processing 88(2-3):133-137.
Kaleta, A., K. Gornicki, R. Winiczenko and A. Chojnacka. 2013. Evaluation of drying models of apple (var. Ligol) dried
in a fluidized bed dryer. Energy Conversion and Management 67:179-185.
Khanali, M., S. Rafiee, A. Jafari, S. H. Hashemabadi and A. Banisharif. 2012. Mathematical modeling of fluidized bed
drying of rough rice (Oryza sativa L.) grain. Journal of Agricultural Technology 8(3):795-810.
Lahsasni, S., M. Kouhila, M. Mahrouz, A. Idlimam and A. Jamali. 2004. Thin layer convective solar drying and
mathematical modeling of prickly pear peel (Opuntia ficus indica). Energy 29(2):211-224.
Law, C. L., S. M. Tasirin and W. R. W. Daud. 2003. A new variable diffusion drying model for the second falling rate
period of paddy dried in a rapid bin dryer. Drying Technology 21(9):1699-1718.

1173
Le, T. Q. and W. Jittanit. 2011. Drying kinetics of cooked jasmine brown rice during various drying methods. The 22nd
National Graduate Research Conference Bangkok, Thailand.
Liu, W., Y. Zheng, L. X. Huang, C. H. Zhang and P. J. Xie. 2011. Low-temperature vacuum drying of natural gardenia
yellow pigment. Drying Technology 29(10):1132-1139.
McMinn, W. A. M., G. Farrell and T. R. A. Magee. 2007. Prediction of microwave drying behavior of pharmaceutical
powders using thin-layer models. Drying Technology 25(9:1551-1569.
Meda, V., M. Gupta and A. Opoku. 2008. Drying kinetics and quality characteristics of microwave-vacuum dried
saskatoon berries. Journal of Microwave Power and Electromagnetic Energy 42(4):4-12.
Menges, H. O. and C. Ertekin. 2006. Modelling of air drying of Hacıhaliloglu-type apricots. Journal of the Science of
Food and Agriculture 86:279–291.
Midilli, A., H. Kucuk and Z. Yapar. 2002. A new model for single-layer drying. Drying Technology 20(7):1503-1513.
Mohamed, L. A.;, C. S. E .Kane, M. Kouhila, A. Jamali, M. Mahrouz and N. Kechaou. 2008. Thin layer modelling of
Gelidium sesquipedale solar drying process. Energy Conversion and Management 49(5):940-946.
Montazer-Rahmati, M. M. and B. Amini-Horri. 2005. From laboratory experiments to design of a conveyor-belt Drryer
via mathematical modeling. Drying Technology 23(12):2389–2420.
Nazghelichi, T., M. H. Kianmehr and M. Aghbashlo. 2011. Prediction of carrot cubes drying kinetics during fluidized
bed drying by artificial neural network. Journal of Food Science and Technology-Mysore 48(5):542-550.
Ong, S. P. and C. L. Law. 2009. Mathematical modelling of thin layer drying of salak. Journal of Applied Sciences
9(17):3048-3054.
Özbek, B., and G. Dadali. 2007. Thin-layer drying characteristics and modelling of mint leaves undergoing microwave
treatment. Journal of Food Engineering 83(4):541-549.
Pin, K. Y., T. G. Chuah, A. Abdull Rashih, C. L. Law, M. A. Rasadah and T. S. Y. Choong. 2009. Drying of betel leaves
(Piper betle L.): Quality and drying kinetics. Drying Technology 27(1):149-155.
Phoungchandang, S. and P. Kongpim. 2012. Modeling using a new thin-layer drying model and drying characteristics
of sweet basil (Ocimum Baslicum Linn.) using tray and heat pump-assisted dehumidified drying. Journal of Food
Process Engineering 35(6):851-862.
Pillai, M. G. 2013. Thin layer drying kinetics, characteristics and modeling of plaster of Paris. Chemical Engineering
Research and Design first online, http://dx.doi.org/ 10.1016/j.cherd.2013.01.005.
Rasouli, M., S. Seiiedlou, H. R. Ghasemzadeh and H. Nalbandi. 2011. Convective drying of garlic (Allium sativum L.):
Part I: Drying kinetics, mathematical modeling and change in color. Australian Journal of Crop Science 5(13):1707-
1714.
Roberts, J. S., D. R. Kidd and O. Padilla-Zakour. 2008. Drying kinetics of grape seeds. Journal of Food Engineering
89(4):460-465.
Ruiz, C. A., C. Francisco, L.-R. Fernand and R. Aida. 2013. Thin layer drying behavior of industrial tomato bye-
products in a convective dryer at low temperatures. Research Journal of BioTechnology 8(2):50-60.
Sacilik, K. and A. K. Elicin. 2006. The thin layer drying characteristics of organic apple slices. Journal of Food
Engineering 73(3):281-289.
Saeed, I. E. 2010. Solar drying of roselle (Hibiscus sabdariffa L.) Part I: Mathematical modelling, drying experiments,
effects of the drying conditions. Agricultural Engineering International: CIGR Journal 12(3):1-13. http://www.
cigrjournal.org/index.php/Ejounral/article/viewFile/1484/1366.
Sarimeseli, A. 2011. Microwave drying characteristics of coriander (Coriandrum sativum L.) leaves. Energy
Conversion and Management 52(2):1449-1453.
Sobukola, O. P., O. U. Dairo, L. O. Sanni, A. V. Odunewu and B. O. Fafiolu. 2007. Thin layer drying process of some
leafy vegetables under open sun. Food Science and Technology International 13(1):35-40.
Tahmasebi, A., J. Yu, Y. Han, H. Zhao and S. Bhattacharya. 2013. Thermogravimetric study and modeling for the
drying of a Chinese lignite. Asia-Pacific Journal of Chemical Engineering first online, doi: 10.1002/apj.1722.
Tee, Y. M., P. Ifa, T. M. Siti, T. M. Z. Meor, D. W. R. Wan and Y. Zahira. 2013. Drying of oil palm frond particles in a
fluidized bed dryer with inert medium. Chemical Industry and Chemical Engineering Quarterly First online,
doi:10.2298/CICEQ120327094Y
Toğrul, İ.T and D. Pehlivan. 2004. Modelling of thin layer drying kinetics of some fruits under open-air sun drying
process. Journal of Food Engineering 65(3):413-425.
Toğrul, H. 2005. Simple modeling of infrared drying of fresh apple slices. Journal of Food Engineering 71(3):311-323.
Vega-Gálvez, A., A. Ayala-Aponte, E. Notte, L. de la Fuente and R. Lemus-Mondaca. 2008. Mathematical modeling
of mass transfer during convective dehydration of brown algae macrocystis pyrifera. Drying Technology 26(12):1610-
1616.
Wiktor, A., M. Iwaniuk, M. S´ledz´, M. Nowacka, T. Chudoba and D. Witrowa-Rajchert. 2013. Drying kinetics of apple
tissue treated by pulsed electric field. Drying Technology 31(1):112–119.
Yücel, U. 2006. Evaluation of High Pressure Pretreatment for Enhancing the Drying Rate of Selected Fruits and
Vegetables. M.Sc. thesis, Middle East Technical University, Ankara, Turkey.

1174
Zomorodian, A. and M. Moradi. 2010. Mathematical modeling of forced convection thin layer solar drying for cuminum
cyminum. Journal of Agricultural Science and Technology 12(4):401-408.

1175