Applied Thermal Engineering 115 (2017) 501–516

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Using waste heat of ship as energy source for an absorption refrigeration

system
Waltteri Salmi a, Juha Vanttola a, Mia Elg b, Maunu Kuosa a,⇑, Risto Lahdelma a
a
Aalto University, School of Engineering, Department of Energy Technology, P.O. Box 14400, FI-00076 Aalto, Finland
b
Deltamarin Ltd, Postikatu 2, 20250 Turku, Finland

h i g h l i g h t s

A steady-state thermodynamic model is developed for absorption refrigeration in a ship.

Operation profile of B.Delta37 bulk carrier is used as an initial data.

Suitability of water-LiBr and ammonia-water working pairs were validated.

Coefficient of performance (COP) was studied in ISO and tropical conditions.

Estimated energy savings were 47 and 95 tons of fuel every year.

a r t i c l e i n f o a b s t r a c t

Article history: This work presents a steady-state thermodynamic model for absorption refrigeration cycles with water-
Received 6 June 2016 LiBr and ammonia-water working pairs for purpose of application on a ship. The coefficient of perfor-
Revised 10 November 2016 mance was studied with different generator and evaporator temperatures in ISO and tropical conditions.
Accepted 28 December 2016
Absorption refrigeration systems were examined using exhaust gases, jacket water, and scavenge air as
Available online 4 January 2017
energy sources. Optimal generator temperatures for different refrigerant temperatures were found using
different waste heat sources and for the absorption cycle itself. Critical temperature values (where the
Keywords:
refrigeration power drops to zero) were defined. All of these values were used in order to evaluate the
Ship
Absorption refrigeration
cooling power and energy production possibilities in a bulk carrier. The process data of exhaust gases
Ammonia and cooling water flows in two different climate conditions (ISO and tropical) and operation profiles of
Lithium bromide a B. Delta37 bulk carrier were used as initial data in the study. With the case ship data, a theoretical
Waste heat recovery potential of saving of 70% of the electricity used in accommodation (AC use) compressor in ISO conditions
and 61% in tropical conditions was recognized. Those estimates enable between 47 and 95 tons of annual
fuel savings, respectively. Moreover, jacket water heat recovery with a water-LiBr system has the poten-
tial to provide 2.2–4.0 times more cooling power than required during sea-time operations in ISO condi-
tions, depending on the main engine load.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction system and to reduce their emissions. Diesel engine combustion

Maritime transport is a huge energy-consuming sector globally.
Typically onboard energy is produced using diesel engine combus-
tion. This power is used for propulsion and to generate electricity
that is needed onboard. However, environmental and economic
concerns have caused the maritime sector to explore alternative
solutions in order to improve the efficiency of the energy usage
⇑ Corresponding author.
E-mail addresses: waltteri.salmi@aalto.fi (W. Salmi), juha.vanttola@aalto.fi
(J. Vanttola), Mia.Elg@deltamarin.com (M. Elg), mabkuosa@hotmail.com (M. Kuosa),
risto.lahdelma@aalto.fi (R. Lahdelma).
produces pollutants like SOx and NOx [1]. Approximately 3.3% of
global CO2 emissions are caused by shipping. The International
Maritime Organization is the major global regulator for emissions
in maritime transport. Moreover, the European Union has imple-
mented a directive for sulfur emissions in specific control areas.
These factors and the fact that the prices of fossil fuels are rising
continuously have motivated the marine sector to reduce its fuel
consumption [2].
Baldi and Gabrielii [3] described a methodology for performing
a feasibility analysis of the installation of a waste heat recovery
(WHR) system on a vessel. The method based on available
data from ship operational profile. The various types of WHR

http://dx.doi.org/10.1016/j.applthermaleng.2016.12.131
1359-4311/Ó 2017 Elsevier Ltd. All rights reserved.
502 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516
Nomenclature
cp specific heat capacity (kJ/[kg K])
Etot total energy production or demand (MW h)
h enthalpy (kJ/kg)
m _ mass flow rate (kg/s)
n_ molar flow (mol/s)
P compressor power (W)
p pressure (Pa)
Q_ cooling or heating power (W)
T temperature (K)
t operation hour (h)
v specific volume (m3/kg)
X mass percentage of absorbent (kg absorbent/kg solution)
x moles of ammonia/moles of liquid
xH2O moles of water/moles of liquid
y moles of ammonia/moles of gas
yH2O moles of water/moles of gas
Greek symbols
e effectiveness of heat exchanger
gp isentropic efficiency of pumping
q density (kg/m3)
technologies available on board ships were discussed from the per-
spective of technological principle and application feasibility by
Shu et al. [4]. The focus was to provide a better understanding of
the options available for WHR in order to improve fuel economy
and environmental compliance.
At present, some effort has been devoted to the utilization of
the vast amount of waste energy from diesel engines used aboard
ships for refrigeration. There are several types of refrigeration tech-
nology being used for marine applications, including compression
refrigeration, sorption refrigeration, and injection refrigeration. It
needs extra energy to drive the compression and injection refriger-
ation systems, which leads to the increase of fuel consumption of
ships. A sorption refrigeration system is driven by thermal energy
and needs little electricity, and it can utilize the waste heat of the
engine and improve the energy conversion efficiency. Therefore,
fuel can be saved considerably. Both absorption and adsorption
refrigeration are sorption refrigeration technologies [4].
Cao et al. [2] conducted a transient simulation of the engine
waste heat powered absorption cooling system for cargo ship
application on a TRNSYS platform. The absorption cooling model
was developed and validated against the literature. The engine
waste heat model was based on experimental data. Challenges to
the absorption cooling system on shipboard applications were also
addressed by suggesting potential solutions. Fernandez-Seara et al.
[5] designed, modeled and analyzed a gas-to-thermal fluid WHR
system for a trawler chiller fishing vessel. An ammonia-water
absorption refrigeration plant was used for onboard cooling. Syn-
thetic oil was used as a heat transfer fluid. The influence of geomet-
ric design parameters and thermal operating conditions were
studied on heat exchangers and system thermal performance.
The high and constant engine load maintained during the fishing
period allowed the recovery of heat from engine exhaust at a ther-
mal level that exceeds that power required by the absorption
system.
Ezgi [6] conducted a thermodynamic analysis of a HVAC system
consisting of a water-lithium bromide (H2O-LiBr) absorption heat
pump for application on a naval surface ship. A comparison was
made between an absorption heat pump and a vapor compression
heat pump system. The case ship used a vapor compression heat
Subscripts
a absorber
c condenser
e evaporator
g generator
s strong solution
w weak solution
1–13 the state points of the absorption refrigeration cycle
Abbreviations
AC air-conditioning
COP coefficient of performance
HT high temperature
HVAC heating, ventilating, and air-conditioning
LT low temperature
ME main engine
NCR normal continuous rating
WHR waste heat recovery
pump system for both heating and cooling. The heat pump had
coefficient of performance (COPs) between two to four. The electric
motor of the compressor in a vapor compression heat pump was
fed by a diesel generator set on board the ship. The electrical load
of compression directly influenced the overall fuel economy and
emissions. Fuel consumption and CO2 emissions decreased while
maintaining the same cooling power when the absorption heat
pump system was used. Ebrahimi et al. [7] studied a two-phase
cooling system and an absorption refrigeration system as associ-
ated with WHR in data centers. They simulated both water-LiBr
and ammonia-water absorption cooling systems. They replaced
the condenser in the cooling circuit with the generator of an
absorption refrigeration cycle. The verification indicated the supe-
riority of the water-LiBr absorption system for data center/server
operating conditions. Menzela et al. [8] presented an experimental
study of an ammonia-water absorption refrigeration system using
the exhaust gas of an internal combustion engine as an energy
source. The exhaust gas energy availability and the impact of the
absorption refrigeration cycle on engine performance, exhaust
emissions, and power economy were evaluated.
Kaita [9] developed equations for calculating the vapor pres-
sure, enthalpy, and entropy of LiBr solutions at high temperatures
using measured data of vapor pressure and heat capacity which are
very helpful for the modeling and design of triple-effect water-LiBr
absorption chillers; these temperatures and pressures are higher
than those used in traditional double-effect chillers. The developed
equations were valid for concentrations between 40 and 65 wt.%
and also for temperatures of 40–210 °C. Lansing [10] described
and analyzed the dynamic simulation and the computer modeling
procedure of a water-LiBr absorption refrigeration system. The
concentration of refrigerant and solution in the range of interest
was from 0.50 to 0.65 kg LiBr/kg solution. The results from this
simulation were heat rates, line concentrations, pressures, and
the overall COP. Kim et al. [11] carried out an experimental study
on a vertical countercurrent slug flow absorber working with
ammonia-water mixture at low solution flow rates. To predict pre-
cise absorption performance in the real absorption heat pump sys-
tem they developed a data reduction model to obtain local heat
and mass transfer characteristics [12].
 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516
A simulated study for a waste heat powered water-LiBr absorp-
tion cycle cooling system for a cargo ship was made by Cao et al.
[2]. Several core components of a cooling system were modeled
and validated. They used the data of a ship that was equipped with
a WARTSILA 8RT-flex68-D engine for propulsion. The rated capac-
ity of the engine was 25,040 kW. The single effect refrigeration sys-
tem was simulated for space cooling. Heat was recovered by an
economizer in a flue gas flow, using a circulating hot water cycle.
Hot water was used as a heat source for the generator after which
it is used for hot water production to meet shipboard hot water
demands (residential, kitchen, and clinic demands). Both the con-
denser and the absorber were cooled with sea water. The chilled
water cycle was cooled in the evaporator and the space cooling
power was transferred to a heat exchanger to the mixture of out-
door air and recirculated indoor air. The simulations were done
under Miami conditions. The resulting COP is 0.64 for the absorp-
tion refrigeration cycle and for the whole system it was 0.6. The
ship’s fuel consumption and the CO2 emissions were 62% less com-
pared to vapor compression. The simulation was also done for
weather conditions in Abu Dhabi and Baltimore. The fuel consump-
tion and CO2 emissions reductions increased by up to 68% under
warmer conditions (Abu Dhabi) compared to vapor compression.
However under colder conditions (Baltimore) they decreased to
38%. The simulation result was also applied to a cruise ship, where
the cooling demand is higher, and an 8.23% total energy consump-
tion reduction (the fuel saving share) was found compared to vapor
compression.
A different solution was presented by Garimella et al. [13] for a
single-effect water-LiBr absorption and supercritical CO2 vapor
compression coupled cycle. The system was modeled for applica-
tion on a naval ship. The exhaust heat (175–275 °C) of the onboard
gas turbine power plant was used as the power source for the gen-
erator of the absorption heat pump. Megawatt scale cooling loads
are required for advanced naval electronics. In order to reach 1 kW/
cm2 cooling fluxes with plausible technical advances while operat-
ing on large surface areas, it is necessary to have, for example, a
coolant available at 40 °C. This is why vapor compression with
a liquid CO2 tank is needed. Additionally, most naval ships require
medium-temperature cooling (5 °C) for electronics and the AC.
Cascade absorption/vapor compression was conceptualized and
analyzed as an energy saving solution. The heat of the condenser
and absorber was rejected to the sea water with a combined cool-
ing cycle. The evaporation of the absorption refrigeration system
was divided into two different stages. The first stage was used to
chill the liquid CO2 vapor compression cycle in order to meet the
cooling storage needs on demand. This arrangement improves
the efficiency of the vapor compression cycle. The second stage is
used for the steady-state cooling of electronics and used in the
AC. Recuperative and solution heat exchangers were added to the
system to improve its efficiency. The COP of the individual absorp-
tion cycle was estimated to be 0.7803 and for the combined total-
energy-input based COP it was estimated to be 0.594, which means
that all the cooling power produced is divided by all the energy
needed in the cascade system. The electrical COP was also defined
as the total cooling power divided by the system’s electricity
power (compression and pumping) requirements; it was estimated
to be 5.685. The cascaded absorption/vapor compression cycle was
estimated to save up to 31% of the electricity consumption in the
case ship studied when compared to an equivalent two-stage
vapor-compression system.
The aim of this research is to find out the refrigeration potential
of an absorption refrigeration system using the waste heat of ships
as the driving force and to estimate how much fuel consumption
can be reduced by this act. The contribution of the study is the pre-
cise ship data (operation profile versus engine load) on the actual
use. Typically refrigeration is produced using vapor compression,
503
which consumes considerably more electricity compared to an
absorption refrigeration system. This electricity is produced at a
rather low efficiency using diesel engine combustion. The chilling
process is necessary for different requirements in maritime trans-
port. For example the AC, ice-making, and medical or food preser-
vation all need refrigeration [4].
First state-of-art and fundamentals of absorption refrigeration
are introduced. Two of the most conventional absorption refriger-
ation systems are explored with a steady-state thermodynamic
model. These systems use water (as refrigerant) and lithium bro-
mide (as absorbent), and ammonia (as refrigerant) and water (as
absorbent) as working pairs [14]. Equations suitable for higher
temperatures of LiBr solutions are selected [9]. In the ammonia-
water case the model takes operation of the rectifier into account.
The process data of the exhaust gases and cooling water flows in
different climate conditions and operation profiles for a B.Delta37
bulk carrier is studied to estimate the refrigeration potential for
ships in general. Energy production by WHR, cooling power possi-
bilities, and annual fuel savings are evaluated in ISO and tropical
conditions.
2. Fundamentals of absorption refrigeration
Absorption refrigeration was discovered by Nairn in 1777 and
the first commercial application was built 46 years later in 1823
by Ferdinand Carré, who also filed several patents introducing a
machine operating on an ammonia-water working pair. A refriger-
ation system using water-lithium bromide was commercialized in
the 1940s and used 1950s as chillers for the air-conditioning (AC)
in large buildings [15]. Many recent efforts have focused on the
development of absorption cycles for applications as low-grade
heat source [4].
The absorption refrigeration system is based on the physical
phenomenon of the different boiling points of pure fluids and for
a mixture of compounds. For example a 60% water-LiBr solution
has a boiling point under 400 mmHg pressure level of approxi-
mately at 135 °C and for pure water it is approximately 82 °C. This
difference in the equilibrium temperatures is used for the pumping
effect in absorption refrigeration [16]. When heat is conducted to a
vacuum tank containing pure water, the water starts to evaporate.
At the same time the solution tank (of water-LiBr for example) is
cooled below its boiling point. Now the partial pressure of the solu-
tion vapor is lower than the pressure of the vapor in the pure water
content tank. When these two tanks are combined with a pipe
above the liquids, the water vapor flows into the solution using
the partial pressure difference as kinetic energy. This pressure dif-
ference is the basis of absorption refrigerators and heat pumps. The
water vapor is absorbed into the solution and the solution receives
its latent heat. The temperature of the solution increases when
water is absorbed until the boiling point is achieved and a state
of equilibrium is reached. Absorption refrigeration systems use
regenerative cycles to maintain the pressure difference and the
cooling effect of the evaporator [16].
The principle of a single-effect water-LiBr cycle is depicted in
Fig. 1. An absorber is a device, wherein pure water vapor absorbs
into a solution. When heat ðQ_ a Þ is conducted out of an absorber,
the thermodynamic equilibrium is disturbed and pure water vapor
flows from the evaporator. Latent heat ðQ_ e Þ added to the evapora-
tor is absorbed into the solution in the absorber [16]. To maintain
the process, the weak water-LiBr solution in an absorber is cooled
ðQ_ a Þ and part of the solution is pumped to the generator where
high-temperature (HT) heat is added ðQ_ g Þ to the solution in order
to separate pure water from the solution. Water evaporates in
the generator and is led to the condenser. The strong solution of
water-LiBr generated in the generator is led back to the absorber
504 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516
Condenser
8
9
Evaporator
Fig. 1. A schematic of a single effect water-LiBr absorption refrigeration cycle.
(through an expansion valve) to maintain a constant LiBr concen-
tration. Water evaporated in the generator is condensed back to
a liquid state in the condenser and condensing heat ðQ_ c Þ is released
[7]. After the condenser, liquid water flows through the expansion
valve back to the evaporator, where low-temperature (LT) heat is
added ðQ_ e Þ to evaporate water (the refrigeration effect). In this
case, water is called the refrigerant and LiBr is the absorbent. A recu-
perative heat exchanger is added between the generator and the
absorber in order to improve the performance of the system [7].
An absorption refrigeration system is quite similar to traditional
vapor compression cycle. Both systems contain a condenser, where
heat ðQ_ c Þ is conducted out of the system, and an evaporator, where
heat ðQ_ e Þ is added to the system. In absorption refrigeration, a
compressor is replaced with an absorber, generator, pump, and
valve (and a recuperative heat exchanger). The advantage of
absorption refrigeration compared to vapor compression is that
the system consumes less electricity because liquid compression
needs less energy than vapor compression. The driving force of
the system is heat ðQ_ g Þ, which is added to the generator [7].
The pressure levels of the system at different state points and
temperatures are demonstrated in a log p T-chart in Fig. 2. Evapo-
ration of the refrigerant and absorption happens in low-level pres-
sure while the generator and condenser operate at high-level
pressure. The solution concentration at each state point is also
demonstrated in Fig. 2. The solution is weak between state points
1 and 3, and strong between state points 4 and 6.
The principle of an NH3 -water absorption refrigeration cycle is
described in Fig. 3. In this case water is the absorbent and ammonia
the refrigerant. This arrangement enables the cooling process to
reach below-zero temperatures in the evaporator. The process
cycle is quite similar to the water-LiBr cycle described previously.
The difference is that the ammonia-water refrigeration cycle needs
Log p
100%,
7, 8 3 4
9, 10 6 X
1 hðT; XÞ ¼
T
Fig. 2. Log p T -chart of a single-effect water-LiBr refrigeration cycle.
7
Generator
3 4
Heat exchanger
2 5
P Pump
1 6
10
Absorber
a rectification column after the generator in order to condense
water due to the existence of water in the ammonia vapor leaving
the generator. This has a negative effect on the performance of the
system. The weak solution mixture pumped from the absorber is
used to condense water in the vapor after the generator. Then
the liquefied water is lead back to the generator. The weak solution
mixture is led to the heat exchanger before entering the generator,
as was the case previously in the water-LiBr absorption cycle [7].
3. Process modeling of the absorption refrigeration
3.1. The water refrigerant–lithium bromide absorbent cycle
The required initial values for water-LiBr absorption refrigera-
tion calculations are the cooling power of the evaporator ðQ_ e Þ,
the temperatures of the generator, condenser, evaporator, and
absorber and also the efficiency of the recuperative heat exchan-
ger. The indexes presented in Fig. 1 are used to nominate the state
points in the cycle.
Some assumptions are made in order to ease the calculation
process. These assumptions are the following: the cycle is a steady
state system; the temperatures of the evaporator, condenser, gen-
erator, and absorber are uniform; all expansion valves are adia-
batic; there is no heat loss to the surroundings; there are no
pressure drops in the pipes; and the refrigerant is saturated vapor
after the evaporator and saturated liquid after the condenser.
The concentrations of water-LiBr solutions are calculated using
Eqs. (1) and (2) by knowing the temperatures of the condenser,
evaporator, absorber, and generator [10]:
49:04 þ 1:125T g  T c
Xs ¼ ; ð1Þ
134:65 þ 0:47T g
49:04 þ 1:125T a  T e
Xw ¼ ; ð2Þ
40%, 134:65 þ 0:47T a
where X s and X w are mass percentages of LiBr in the strong and
weak solutions respectively (kg LiBr/kg solution) and T g , T c , T a ,
and T e are the temperatures (°C) of the generator, condenser, absor-
ber, and evaporator respectively.
The enthalpies (kJ/kg) of the solutions are calculated using Eq.
(3) [9]:
4 X
3 X
2
ðan X n Þ þ T ðbn X n Þ þ T 2 ðcn X n Þ þ T 3 d0 ð3Þ
n¼0 n¼0 n¼0
where T is the temperature (K) of the solution between
273 6 T 6 463 °C, and X (wt.%) is the concentration of the absor-
 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516 505

8
Condenser 10 Rectifier 9 Generator

4 5
11 2 3
Heat exchanger
P Pump
6

12
1
7
13
Evaporator Absorber

Fig. 3. A schematic view of a refined ammonia-water absorption refrigeration cycle.

Table 1
The constants for enthalpy calculations for the water -LiBr solution [9].

n an bn cn dn
0 954.8 3.293E1 7.4285E3 2.269E6
1 47.7739 4.076E2 1.5144E4
2 1.59235 1.36E5 1.3555E6
3 2.09422E2 7.1366E6
4 7.689E5

bent (LiBr) between 40 6 X 6 65 wt.%. The constants of the Eq. (3) m _ sXs
_ wXw ¼ m ð8Þ
are presented in Table 1.
From these equations we can obtain the rest of the mass flows with
Knowing the temperatures and LiBr concentrations of the absor-
Eqs. (9) and (10):
ber and generator, we can calculate the enthalpies at points 1 and 4.
Knowing the temperatures of the condenser and evaporator and _ H2O
m
assuming that the outlet of the condenser is saturated liquid and _s¼
m ; ð9Þ
Xs
Xw
1
the outlet of the evaporator is saturated vapor, the enthalpies at
state points 8 and 10 are calculated using REFPROP [17]. The energy
m_ H2O
equation for the expansion valve is simplified, so that _w¼
m ; ð10Þ
1  XXws
h8 ¼ h9 : ð4Þ
To define the work done by the pump, the density of the strong
The pressure levels of the condenser and evaporator are calculated solution at point 1 must be calculated. It is gained by using Eq. (11)
using Eq. (5) [7]: [14]:
 
11:78ðT  372:79Þ q1 ¼ 1145:36 þ 470:84X w þ 1374:39X 2w
p ¼ 105 exp ; ð5Þ
ðT  43:15Þ
 ð0:333393 þ 0:571749X w ÞT a ; ð11Þ
where T (K) is the temperature of the condenser or the evaporator.
where q1 (kg/m3) is the density of the solution at state point 1. The
These pressures also define the pressure levels of the whole system.
work done by the pump is
The pressure in the generator equals the pressure in the condenser,
and the pressure in the absorber equals the pressure in the _ w ðp2  p1 Þ
m
P¼ ; ð12Þ
evaporator. q1
The enthalpy of the water vapor at state point 7 (Fig. 3) is cal-
culated using the pressure in the condenser and the temperature where p2  p1 is the pressure (Pa) difference between the condenser
of the generator with REFPROP. and the evaporator. For approximations of the specific heat values
Knowing the cooling capacity needed, the mass flow rate of the of the solution at state points 2 and 4 Eq. (13) is used [9]:
water cycle in the system is defined by the equation for the conser-
vation of energy in the evaporator [7]:
X
3 X
2
cp ¼ ðbn X n Þ þ 2T ðcn X n Þ þ 3T 2 d0 ; ð13Þ
Q_ e n¼0 n¼0
_ H2O
m ¼ ; ð6Þ
h10  h9 where cp is the specific heat capacity (kJ/kg K) of the solution, T is
the temperature (K) of the solution between 273 6 T 6 463 °C,
where Q_ e is the heat transfer rate at the evaporator (W) and m
_ H2O and X (wt.%) is concentration of the absorbent (LiBr) between
(kg/s) is the mass flow rate of the water cycle. The cycle contains 40 6 X 6 65 wt.%. The constants of the equation are specified in
two equations for the conservation of masses. For the whole system, Table 1. The heat capacity rates at state points 2 and 4 are estimated
the conservation of mass can be written as in Eq. (7): in order to find out enthalpies after the heat exchanger at state
m _ H2O þ m
_w¼m _ s; ð7Þ points 3 and 5 [7].
_ w cp2 > m
If m _ s cp4 , then
where m _ w (kg/s) is the mass flow rate of the weak LiBr solution and
m_ s (kg/s) is the mass flow rate of the strong LiBr solution. The con- T 5 ¼ T g  eðT g  T a Þ; ð14Þ
servation of mass for LiBr can be written as
506 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516

where e is the effectiveness of the heat exchanger. The enthalpy at (Pa). The ammonia mole fraction is modified in Eq. (23), until Eq.
state point 5 is calculated with Eq. (3) and from the equation for the (24) applies:
conservation of energy we get Eq. (15):
T  ðT real þ 273:15Þ 6 0:001; ð24Þ
_s
m
h3 ¼ ðh  h5 Þ þ h2 : ð15Þ where T real is the real temperature of the absorber or the generator
_w 4
m (the initial values of calculation in Celsius). The same method is uti-
_ w cp2 6 m
If m _ s cp4 , then lized to calculate the mole fraction during the gaseous phase of the
mixture [18]:
T 3 ¼ T a þ eðT g  T a Þ ð16Þ X
T ¼ T0 f 2 ðy; pÞ; ð25Þ
and the enthalpy at state point 3 can be calculated with Eq. (3). The
enthalpy at state point 5 is calculated using the equation for the where y is the mole fraction of ammonia in the gaseous phase
conservation of energy. The result is Eq. (17): (moles of ammonia/moles of gas). The mole fraction is modified
until Eq. (24) applies. The mole fractions of water are calculated
_w
m using Eqs. (26) and (27):
h5 ¼ ðh  h3 Þ þ h4 : ð17Þ
_s 2
m
xH2O ¼ 1  x; ð26Þ
As previously, the enthalpy change after the expansion valve is
considered to be zero (Eq. (18)). yH2O ¼ 1  y: ð27Þ
h6 ¼ h5 ð18Þ Now that the temperatures, pressures and mole fractions at state
points 1, 5, and, 8 are completely known, it is possible to calculate
Once all the enthalpies and mass flows are known, it is possible
enthalpies by using REFPROP. Because REFPROP has a small differ-
to calculate the heat transfers at each point of the system with Eqs.
ence in their equilibrium states in liquid phases compared to the
(19)–(21) [7]:
source [18] used in the present research, it is necessary to calculate
Q_ a ¼ m
_ h20 h10 þ m
_ s h6  m
_ w h1 ; ð19Þ the enthalpies at a one-degree lower temperature, which may cause
a small error in the model.
The pump is assumed to be adiabatic, so Eq. (28) applies:
Q_ g ¼ m
_ h20 h7 þ m
_ s h4  m
_ w h3 ; ð20Þ
v 1 ðp2  p1 Þ
h2 ¼ h1 þ ; ð28Þ
Q_ c ¼ m
_ h20 ðh7  h8 Þ; ð21Þ gp
where v1 is specific volume and gp the isentropic efficiency of
where Q_ a , Q_ g and Q_ c are the heat transfer rates (W) of the absorber,
pumping. The molar flow of pure ammonia (mol/s) is calculated
generator, and condenser respectively. The COP of the system can
with Eq. (29):
be explored now. In this case, the COP is defined as the ratio of heat
power in the evaporator to heat power in the generator in Eq. (22) Q_ e
[7]: n_ NH3 ¼ ; ð29Þ
h13  h12
Q_ e where h is enthalpy (J/mol) and the sub index is the state point. As
COP ¼ : ð22Þ
Q_ g in the water-LiBr case, Eqs. (9) and (10) can be exploited in calculat-
ing the molar flows (mol/s) of weak and strong solutions, using the
same logic. For the ammonia-water case the equations are (30) and
3.2. Ammonia refrigerant–water absorbent cycle (31):
n_ NH3
The indexes presented in Fig. 3 are used to nominate the state n_ s ¼ ; ð30Þ
Xs
1
points in this cycle. The required initial values and assumptions Xw

are the same as they were for the water-LiBr working pair. More-
over all water is assumed to condense at the rectifier, in such a
way that at state point 10 the refrigerant is pure ammonia. The
pump is also assumed to be adiabatic.
Knowing the temperatures in the condenser and evaporator,
the pressure levels of the system are calculated using Eq. (5).
Enthalpies at state points 10, 11 and 13 are calculated for pure
ammonia, using REFPROP. The outlet of the condenser is assumed
to be saturated liquid and the outlet of the evaporator to be sat-
urated vapor. The enthalpy at state point 12 equals that at state
point 11, because enthalpy is assumed to be constant over the
expansion valve.
The ammonia-water system is calculated on a per mole basis
since such equations were available in the source [18]. To calculate
the ammonia mole fraction in liquid phases in the absorber and
generator, which is equal to the weak and strong solution respec-
tively, an iteration process is used. The temperature of the mixture
is calculated using the following equilibrium equation [18]:
X
T ¼ T0 f 1 ðx; pÞ;
where T 0 (K) is the reference temperature, x is the mole fraction of
ammonia (moles of ammonia/moles of liquid) and p is the pressure
n_ NH3
n_ w ¼ ; ð31Þ
1  XXws
where n_ s is the molar flow (mol/s) of the strong solution, n_ w the
molar flow (mol/s) of the weak solution, X w is the molar fraction
of water in the weak solution and X s is the molar fraction of water
in the strong solution (obtained from Eq. (26)).
From the mass balance of ammonia and the rectifier it is possi-
ble to calculate the molar flows (mol/s) at states 8 and 9 with Eqs.
(32) and (33) respectively:
n_ NH3
n_ 8 ¼ ; ð32Þ
Ya
n_ 9 ¼ n_ NH3  n_ 8 ; ð33Þ
where Y a is the mole fraction of ammonia in the gaseous phase in
the generator, obtained with the previous iteration process (Eqs.
(25) and (24)).
ð23Þ The enthalpy at state point 9 is calculated with REFPROP and
with an assumption that it is saturated liquid water at the temper-
ature of the generator. The enthalpy (J/mol) at state point 3 is cal-
culated with Eq. (34):
 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516 507

n_ 8 h8 þ n_ w h2  n_ 9 h9  n_ NH3 h10 can be seen, the model is very accurate. The maximum deviation
h3 ¼ : ð34Þ
n_ w against data series 1 (for output results, Table 2) is 0.7% (e.g. absor-
ber cooling load) whereas it is 0.1% against data series 2 (Table 3).
Specific heat capacities (J/mol K) for weak and strong mixtures
The model of Ebrahimi et al. [7] has a 5% deviation against data ser-
and temperatures before the heat exchanger are calculated using
ies 1. The present model is more accurate because of the wider
REFPROP. Enthalpies after the heat exchanger are calculated using
range of enthalpy and the heat capacity Eqs. (3) and (13). Data ser-
the same method as previously.
ies 2 uses the generator heat load as input data, so an iteration was
If n_ w cp3 6 n_ s cp5 , then
required for comparison (it is not an initial value in the present
T 4 ¼ T 3 þ eðT g  T 3 Þ ð35Þ model). In other words, the evaporator heat load was changed until
the generator loads were equivalent.
and the enthalpy at state point 4 can be calculated with REFPROP.
The ammonia-water model is compared to the COP graphs of
The enthalpy at state point 6 is calculated using the equation for
Ouadha and El-Gotni [21]. The graphs based on a thermodynamic
the conservation of energy. The result is Eq. (36):
analysis were generated by changing the temperature of the gener-
n_ w h3 þ n_ s h5  n_ w h4 ator while keeping the other temperatures constant. The graphs
h6 ¼ : ð36Þ
n_ s are performed for three different condenser temperatures (20 °C,
30 °C and 40 °C). The comparison is shown in Fig. 4. The curves
If n_ w cp3 > n_ s cp5 , then
of the present study are on the left side of Fig. 6. It can be noted
T 6 ¼ T g  eðT g  T 3 Þ ð37Þ that the curves are in great agreement with the critical values
(where the COP of the system reaches zero point) and the temper-
and the enthalpy at state point 6 can be calculated with REFPROP. atures are approximately 20° above that critical point. The differ-
The enthalpy at state point 4 is calculated using the equation for ence between the models’ behaviors is that the curves of the
the conservation of energy. The result is Eq. (38): present study are descending at high temperatures while the
n_ w h3 þ n_ s h5  n_ s h6 curves of Ouadha and El-Gotni [21] are ascending. This is because
h4 ¼ : ð38Þ the vapor after the generator is assumed to be pure saturated
n_ w
ammonia in their model. In the present study a rectifier is needed
As previously, the enthalpy change after the expansion valve is for removing the evaporated water. Moreover the vapor is not
considered to be zero (Eq. (39)). assumed to be saturated after the generator. At higher tempera-
h7 ¼ h6 ð39Þ tures more water is vaporized and thus, more energy is needed
to condense it. This is why the coefficient of the performance des-
Now that all the enthalpies and mass flows are known, it is possible cends. The models are similar when the amount of vaporized water
to calculate the heat transfers at each point of the system with Eqs. remains in minor quantities (near the critical temperatures). The
(40)–(42): present model is more accurate because it takes vaporizing water
Q_ a ¼ n_ NH3 h13 þ n_ s h7  n_ w h1 ; ð40Þ and the rectifier into account.

Q_ g ¼ n_ 8 h8 þ n_ s h5  m
_ w h4  n_ 9 h9 ; ð41Þ 3.4. The system performance and critical temperature values

Q_ c ¼ n_ NH3 ðh10  h11 Þ;
where Q_ a , Q_ g , and Q_ c are the heat transfer rates (W) of the absorber,
generator, and condenser respectively. The COP of the system is cal-
culated with Eq. (22).
3.3. Model validation
In order to ensure that the thermodynamic model created is
accurate, a comparison against experimental data should be per-
formed. The input and output data series 1 [19] and 2 [20] were
used for comparison to the results of present theoretical study.
The comparison and deviations are presented in Tables 2 and 3.
The input data and the output results of data series 1 and 2 are
the same that Ebrahimi et al. [7] used as a comparison source. As
ð42Þ The performance of the system was studied by changing the ini-
tial values of the process model. The COP values of the systems
were plotted while changing the generator temperatures and keep-
ing other temperatures constant. This plotting is studied with three
different evaporator temperatures. The results are shown in Fig. 5
for water-LiBr and for the ammonia-water system in Fig. 6. In both
cases the temperatures of the condenser and absorber were set to
30 °C.
It can be clearly seen in Fig. 5 that the COP of the water-LiBr sys-
tem is quite stable but drops fast while approaching the critical
generator temperature where the refrigeration power is zero. The
value of the COP of the ammonia-water working pair does not drop
so dramatically. The COP starts to decrease by approximately 15
degrees before the critical generator temperature value while
the decreasing point is less than 10 degrees before the critical

Table 2
The data 1 used by Ebrahimi et al. [7] for water-LiBr absorption refrigeration calculations.

Model comparison against Data series 1 Symbol (unit) Data series 1 Present study Data type (input/output) Deviation (%)
Generator temperature T g ð CÞ 84.8 84.8 Input –
Condenser temperature T c ð CÞ 39.8 39.8 Input –
Evaporator temperature T e ð CÞ 8.6 8.6 Input –
Absorber temperature T a ð CÞ 35.5 35.5 Input –
The effectiveness of the heat exchanger e ð%Þ 70.7 70.7 Input –
Evaporation cooling load Q_ e ðkWÞ 201.29 201.29 Input –
Generator heating load Q_ g ðkWÞ 259.55 257.91 Output 0.6
Condenser cooling load Q_ c ðkWÞ 213.37 213.41 Output 0.0
Absorber cooling load Q_ a ðkWÞ 247.47 245.80 Output 0.7
Coefficient of performance COP 0.7755 0.7805 Output 0.6
508 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516

Table 3
A comparison of water-LiBr absorption refrigeration calculation results with the data 2 of Ebrahimi et al. [7].

Model comparison against data series 2 Symbol (unit) Data series 2 Present study Data type (input/output) Deviation (%)

Generator temperature T g ð CÞ 90 90 Input –
Condenser temperature T c ð CÞ 35 35 Input –
Evaporator temperature T e ð CÞ 5 5 Input –
Absorber temperature T a ð CÞ 35 35 Input –
The effectiveness of the heat exchanger e ð%Þ 60 60 Input –
Generator heating load Q_ g ðkWÞ 100 100 Input –
Evaporation cooling load Q_ e ðkWÞ 75.865 75.777 Output 0.1
Condenser cooling load Q_ c ðkWÞ 80.981 80.874 Output 0.1
Absorber cooling load Q_ a ðkWÞ 94.885 94.904 Output 0.0
Pump work P ðkWÞ 0.001 0.001 Output 0.0
Coefficient of performance COP 0.7587 0.7578 Output 0.1

Fig. 4. A comparison of ammonia-water absorption refrigeration model COP values (left) with the graphs of Quandha and El-Gotni (right).

0.90
0.80
0.70
0.60
0.50
COP

0.40
0.30
0.20
0.10
0.00
50 55 60 65 70 75 80 85 90
Temperature of the generator [°C]
2 celcius evaporator temperature
5 celsius evaporator temperature
10 celcius evaporator temperature

Fig. 5. A comparison of COP values with the different generator and evaporator temperatures of the water-LiBr absorption refrigeration cycle.

temperature value with the water-LiBr system. The critical gener-
ator temperatures are higher for colder evaporator temperatures.
At a 10 °C evaporator temperature the critical value is 52 °C in
the water-LiBr system and at a 20 °C evaporator temperature in
the ammonia-water system the critical temperature is 90 °C.
It was also noted that the COP of the system decreases and the
critical temperature increases if the condensation and absorption
temperature increases. The critical temperatures of the generator
mentioned previously increased to 74 °C and 114 °C, respectively,
when the absorption and condensation temperatures were
increased by 10 °C. This is why the sea-water temperature is an
important factor for the system performance because it is used
for cooling the absorber and the condenser. These temperatures
should be kept as close to the sea water temperature as possible
for a better COP. A cooler sea-water temperature enables better
COP compared to tropical conditions.
The disadvantage of the water-LiBr working pair is that it can-
not produce a below-zero degree coolant. It also has a risk of crys-
tallization at high temperatures when the LiBr concentration is
high. The crystallization limit is shown in Fig. 7, which also pre-
sents an approximate polynomial equation of crystallization limit
between 45 and 103 °C [2].
 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516 509

0.6

0.5

0.4

COP
0.3

0.2

0.1

0
70 75 80 85 90 95 100 105 110 115 120
Temperature of the generator [°C]

-5 celcius evaporator temperature

-10 celcius evaporator temperature
-20 celcius evaporator temperature

Fig. 6. A comparison of COP values with different generator and evaporator temperatures of the ammonia-water absorption refrigeration cycle.

3.5. The optimal temperature values in ISO conditions

The opmal temperature of the

78.0
76.0
The optimal temperatures for the absorber, condenser, and gen-
74.0

generator [°C]
erator with different evaporator temperature levels were defined
by using Excel Solver to reach the maximal COP of the water-LiBr 72.0
absorption refrigeration cycle. The temperatures were variables 70.0
while the cooling power and efficiency of the heat exchanger were 68.0
fixed. Excel Solver uses an evolutionary optimization method to 66.0
find out the optimal values of the temperatures. The optimal tem- 64.0
peratures of the generator and the optimal COP values are shown 62.0
in Figs. 8 and 9 as a function of the evaporator temperature for 2 3 4 5 6 7 8 9 10 11 12
the water-LiBr working pair. During the optimization of the Temperature of the evaporator [°C]
water-LiBr absorption cycle it was noted that the temperatures Fig. 8. The optimal temperatures of the generator of the water-LiBr absorption
of the absorber and the condenser should be kept as low as possi- refrigeration cycle.
ble in order to maximize the COP. In this case the lowest temper-
atures are approximately 30 °C because the sea- water
temperature was in ISO conditions. 0.87
The optimal temperature of the generator decreases linearly as 0.86
The opmal COP

the temperature of the evaporator rises (Fig. 8). This effect was also 0.85
noted when operating under any conditions (not only with optimal 0.84
COP temperatures). It can be noted that the COP improves if the 0.83
0.82
evaporator (cooling) temperature is lifted (Fig. 9). This effect can
0.81
be also noted in Figs. 5 and 6. Excel Solver could not be used for
0.80
the ammonia-water system optimization because of the iteration
0.79
round needed for each temperature change. The temperatures in 0.78
the absorber and in the condenser were set to be fixed as low as 2 3 4 5 6 7 8 9 10 11 12
possible (30 °C) because it was the best solution for water-LiBr Temperature of the evaporator [°C]
system. The optimal temperature of the generator was solved
Fig. 9. The optimal COP of the water-LiBr absorption refrigeration cycle.

numerically, changing the temperature value of the generator for

each evaporator temperature. This graph was shown previously
in Fig. 6. The optimal generator temperatures were 125 °C, 130 °C
and 145 °C for evaporator temperatures of 5 °C, 10 °C, and
crystallized 20 °C respectively.

4. Case calculation: A B.Delta37 bulk carrier ship

liquid
4.1. The water cooling system in a cargo ship

A simplified water cooling system and its main components for

a cargo ship are presented in Fig. 10. The system is divided into two
separate water cycles: an LT and an HT cycle. The LT water cycle
Fig. 7. An approximation of the crystallization limit in a water-LiBr solution for (the blue1 lines in the figure) is mainly used to cool the auxiliary
temperature values 45–103 °C [1]. engines, the air of the main engine (ME), the AC chillers, the
510 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516

Fig. 10. The main components in a cargo ship water cooling system [23].

Table 4
The main operational values of the case ship’s cooling system, including the yearly operating hours and the amount of cooling power required by the cooling system components
(kW) [22].

Yearly Jacket Lubrication Scavenge AC Surplus steam Auxiliary Other LT Central

operation water oil cooler air cooler chiller dumping condenser diesel engine consumers cooler
hours cooler
Dimensioning case, tropical – 950 600 1980 432 450 583 166 5161
conditions
ISO, 100% ME load – 930 590 1940 183 10 465 66 4184
ISO, Scantling draught 74.4% 1226.4 760 540 1490 183 10 465 66 3514
ME load
ISO, Scantling draught 39.0% 1226.4 520 420 440 183 5 465 66 2099
ME load
ISO, Design draught NCR 919.8 720 520 1220 183 5 465 66 3179
67.8% ME load
ISO, Design draught slow 919.8 485 400 330 183 0 465 66 1929
steaming 35.1% ME load
ISO, Ballast draught 63.7% ME 919.8 690 500 1100 183 0 465 66 3004
load
ISO, Ballast draught slow 919.8 415 360 100 183 0 465 66 1589
steaming 25.0% ME load
ISO, Maneuvering 96 0 0 0 183 0 498 66 747
ISO, Harbour loading 936 0 0 0 183 0 1026 66 1275
ISO, Harbour unloading 1176 0 0 0 183 0 1026 66 1275
ISO, Harbour standstill 456 0 0 0 183 0 423 66 672

lubrication oil, the HT water cycle and to condensate the surplus
steam from the boilers. The central cooler chills the cycle with
sea-water. The HT water cycle (the red lines) is used for jacket cool-
ing and to produce fresh, drinkable water [22]. The AC chiller
includes the cooling of the condenser of the vapor compression
cycle. The evaporator of the vapor compression cycle is used to
produce sufficient cold water for its AC purposes. This is the only
cold refrigeration (below sea-water temperature) need where an
absorption refrigeration system could be useful in this case. In other
ship types absorption cooling could also be used for ice-making and
medical or food preservation needs, for example, as these require
below-zero chilling temperatures [22].
4.2. The main operation values and energy sources of the case ship
The main operational ISO values for the cooling system of a B.
Delta37 bulk carrier are shown in Table 4. Yearly operation hours
for the ME are presented with different engine loads and operation
conditions. According to the operation profile, the ME load varies
between 25% and 74% on the sea. The column on the right of Table 4
 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516 511

Table 5 According to mass flows, jacket water has the most potential but
The best potential heat sources at various temperature levels and mass flows and its disadvantage is the LT level. Exhaust gases seem to be the best
with different ME loads in ISO conditions [22].
energy source among the options but the disadvantage of them as a
Temperature levels 208–177 °C 203–68 °C 85 °C source is that they cannot be cooled down below 167 °C because of
ME load (%) Exhaust gas Scavenge air Jacket water the risk of sulfur corrosion [22]. All of the sources are explored to
(kg/s) (kg/s) (kg/s)
compare the gained refrigerating powers and total cooling energies
100 11.8 11.6 14.8 that can be used for cooling purposes in the case ship.
75 10.3 10.1 12.1
65 9.1 9.0 11.0
40 6.1 6.0 8.3 5. Cooling power potential with the LiBr-water working pair
35 5.4 5.3 7.6
25 4.0 3.9 6.5
5.1. The initial values of the calculations

The cooling power of different heat recovery systems are calcu-

lated using the same heat exchanger efficiency (70.7%) [14] that is
Table 6
The initial values and variables of the absorption refrigeration systems.
also used in data series 1 in Table 2, while the temperatures of the
condenser and absorber are set to be constant. In ISO conditions
Water-LiBr Ammonia-water these temperatures are 30 °C and in tropical conditions 40 °C
absorption absorption refrigeration
refrigeration
(because of the higher sea-water temperature). In the previous
chapter it was mentioned that these temperatures should be kept
The temperatures of ISO condition: 30 °C ISO condition: 30 °C
as close as possible to the temperature of the sea- water for a bet-
condenser and Tropical condition: Tropical condition:
absorber 40 °C 40 °C ter COP. Other initial values are the same for both conditions.
The temperature of the evaporator is studied at 2 °C, 5 °C, and
Heat exchanger efficiency 70.7% 70.7%
10 °C temperature levels in the water-LiBr absorption system. In
The temperatures of 2 °C; 5 °C; 10 °C 5 °C; 10 °C; 20 °C
the ammonia-water system the studied temperature levels of the
evaporator
evaporator were 5 °C, 10 °C, and 20 °C. The temperature of
The temperatures of the Variable Variable
the generator and the cooling power, which depends on the heat-
generator
ing load of the energy source, were the variables of the model. All
Cooling power Variable Variable
of the initial values and variables are presented in Table 6.

5.2. Heat recovery in exhaust gases

Generator
208 – 177 °C
Fig. 11. A schematic view of the generator heating process using exhaust gases.
summarizes the heat flows. The LT circuit is cooled by sea-water.
These engine loads and operating hours are used for the calcula-
tions of the total cooling energy production of absorption refriger-
ation systems. The AC chiller’s cooling power in an evaporator is
350 kW for the dimensioning case in tropical conditions, which is
used to estimate the cooling power needed in tropical conditions,
and 150 kW is used for ISO condition calculations. The AC chiller’s
cooling power presented in Table 4 shows the cooling needs of the
condenser in the vapor compression cycle, which requires 350 kW
and 150 kW of evaporator cooling power with the 4.5 COP of the
existing cooling system [22].
According to cooling power needs and temperature levels, the
energy sources with the best potential for absorption refrigeration
are jacket water and scavenge air. The HT exhaust gases of the ME
and auxiliary engines are also potential energy sources, but they
are not cooled by the water cooling system and are thus not shown
in Table 4. The amount of surplus steam is a minor factor in ISO
conditions but in tropical conditions there is more potential for
use as an energy source. The best potential heat sources with the
temperature levels and mass flows with different ME loads are pre-
sented in Table 5 [22]. It can be seen that exhaust gas and scavenge
air both have high operating temperatures and similar mass flows
but exhaust gases have more stable temperature variation.
The temperature of the exhaust gases and mass flows varies
during different engine operation conditions. This HT waste heat
can be used as the input power for absorption refrigeration. The
limitation in use is that the exhaust gases cannot be cooled under
the temperature of 167 °C.
Using the bulk carrier data in ISO conditions, the available heat
is calculated assuming that the flue gas is cooled down to 167 °C. It
was noted in the previous chapter that all optimal generator tem-
peratures are below this temperature (Fig. 10). The maximum cool-
ing power in different loads can be calculated by utilizing the
optimal COP generator temperatures and the flue gas data of the
case ship.
A schematic view of the process is demonstrated in Fig. 11. The
generator is placed in the flue gas flow and it is assumed that the
temperature inside the generator does not heat up over the opti-
mal temperature. The heating power of the flue gas flow is calcu-
lated using its molar composition and the temperature of the
flow before (which varies during loads) and after the generator,
using REFPROP:
X
Q_ g ¼ n_ gas ðhin  hout Þ; ð43Þ
where n_ gas is the molar flow (mole/kg) for each substance in the
exhaust gas. The enthalpies hin and hout (J/mole) are for each flue
gas species at the inlet temperature (taken from the data) and at
the outlet temperature (167 °C). The cooling power is calculated
using Eq. (44) and the optimal COPs from Fig. 9.
Q_ e ¼ Q_ g  COP ð44Þ
Because Q_ e is the input value of the system, a newly calculated Q_ e
was set to the thermodynamic model of the refrigeration system
to calculate a new value of the COP. This iteration was made until
the values converged. In this case the values converged immedi-
ately, so it was noted that the COP is only a function of the temper-
atures and the effectiveness of the heat exchanger.
512 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516
The cooling power ðQ_ e Þ for a 5 °C evaporating temperature is
calculated for different ME loads. It can be noted that the cooling
power demand for the AC of a bulk carrier (approx. 150 kW) can
be reached easily by all engine loads. During scantling draught,
when 74.4% ME load is used, it would be possible to produce
approximately 530 kW cooling power for AC needs, which is three
and a half times the required cooling power. The condenser and
absorber require approximately 1190 kW cooling power produced
by LT water or sea-water. During the lowest ME load (25%) approx-
imately 250 kW of cooling power can be produced, which is still
enough to chill the cabins. The cooling power reduces by 2% at
scantling draught, when the temperature of the evaporator is
2 °C and rises to 5% when it is 10 °C. The COPs are 0.79 and 0.85
respectively.
5.3. Heat recovery in jacket water
ME jacket cooling water is relatively HT (85 °C) water and,
therefore, it could be used as an energy source for the absorption
refrigeration. The generator could be placed before the HT water
cooler, as presented in Fig. 12 where a schematic view of the gen-
erator heating and ME cooling process by jacket water are shown.
An HT water cooler is used as an auxiliary cooler when the gener-
ator cannot use all the available energy to cool the water cycle for
the ME operating conditions. To simplify the case, the energy con-
sumption of the fresh water generator is not taken into account.
The operating temperatures of the HT water cycle are 85 °C and
70 °C (the return and supply water of the ME) [22]. In this case the
generator temperature was set to 65 °C to ensure a 5 °C tempera-
ture difference in the heat exchanger. Lower generator tempera-
tures do not improve the system performance and much higher
temperatures cannot be reached.
It can be noted from Fig. 5 that a 65 °C generator temperature is
quite close to the critical value of 2 °C evaporation temperature.
The COP is 0.68, which is used in calculations, but even a small
temperature change can drop the system performance dramati-
cally. For 5 °C and 10 °C evaporator temperatures the COP values
are 0.79 and 0.85 respectively. The cooling power ðQ_ e Þ of the sys-
tem is calculated using Eq. (44).
The cooling capacity with a 5 °C evaporator temperature for AC
at 74.4% ME load is approximately 600 kW, which is over 50 kW
more than in the case of exhaust gas heat recovery. During scant-
ling draught at 39% ME load approximately 410 kW of cooling
Main
engine Pump
Fresh
water
generator
85 °C
HT –water
Generator cooler
Fig. 12. A schematic view of the generator heating and ME cooling process using
jacket water.
power can be produced, which is still enough to chill cabins over
two and a half B.Delta37 bulk carriers and it is 75 kW more that
can be produced by an absorption refrigeration system using
exhaust gases. On average the overall cooling power potential at
the whole engine scale is 16% more than in the previous exhaust
gas heat recovery case.
The problem of the system is that the generator temperature is
quite close to the boundary limit (at low evaporator temperatures)
of where the system malfunction (Fig. 5). In tropical conditions,
where the absorber and condenser temperatures are higher
(approx. 40 °C) because of warmer sea- water, the refrigeration
system does not work at any evaporator temperature between
2 °C and 10 °C.
5.4. Heat recovery in a scavenge air cooler
Scavenge air is used in diesel the engine combustion process. It
has to be compressed before combustion and during the compres-
sion its temperature raises to 68–203 °C, depending on the ME load
(Table 5), and it needs to be cooled down before entering the ME.
The schematic of the process is quite similar to that of exhaust gas
heat recovery and is shown in Fig. 13.
The input heat power is calculated using Eq. (45) with REFPROP:
Q_ g ¼ m
_ in ðh1  h2 Þ; ð45Þ
where h1 ¼ hðT 1 ; p1 Þ, the inlet temperatures ðT 1 Þ and pressures ðp1 Þ
are obtained from the B.Delta37 bulk carrier data and
h2 ¼ hðT g þ 10  C;p1 Þ. The pressure difference over the generator is
assumed to be zero and the outlet temperature is 10 °C higher than
the temperature of the generator to ensure a good heat exchange
ratio. The cooling power is calculated using Eq. (44).
The temperature of the generator is optimized so that the over-
all refrigerating energy is maximized for yearly operation (using
Excel Solver). Different ME loads with operating hours were pre-
sented previously in Table 4. The overall energy production is cal-
culated using Eq. (46).
The optimal generator temperature for a 5 °C evaporator tem-
perature is 73 °C (as presented in Fig. 10) when optimizing the
COP of the absorption refrigeration cycle. However, more energy
can be produced for cooling purposes in this case when the tem-
perature of the generator is set to be 64 °C. This is the optimal tem-
perature with a 5 °C evaporator temperature when the overall
refrigerating energy is optimized with heat recovery system using
scavenge air with the ME load profiles of the case ship. This means
that it is only reasonable to produce the cooling power with 30%
and higher engine loads in order to maximize the total cooling
energy production. Using the same maximized refrigeration energy
production optimization method, the optimal temperatures of the
generator for 2 °C and 10 °C evaporator temperatures were 68 °C
and 57 °C respectively, which leads to the COPs of 0.75 and 0.82
respectively.
In the case of a 5 °C evaporator temperature, the COP is 0.78 and
the cooling power for 65% and higher engine loads are greater com-
pared to the cooling power of the absorption refrigeration system
using jacket water. The cooling power at scantling draught with
74.4% ME load is 60% higher than with the exhaust gas heat recov-
ery and 42% higher than heat recovery in jacket water. Moreover it
would be even higher if the cooling power was optimized for high
engine loads and without taking overall energy production into
account. However the cooling capacity drops fast and with 39.0%
scantling draught only approximately 116 kW of chilling power
can be produced. This power does not meet the whole cooling
demand of 150 kW.
 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516
1 Generator
203 – 68 °C
Fig. 13. A schematic view of the generator heating process using scavenge air.
6. Cooling power potential with the ammonia-water working
pair
The calculation process of the ammonia-water absorption
refrigeration system with heat recovery in exhaust gases is similar
to the one previously presented for the water-LiBr system. The heat
flows of the exhaust gases with different engine loads remain the
same but the evaporation temperatures and system performance
differ. The optimal temperature values (calculated previously) are
used to obtain maximal cooling power generation. The COP of
the system is 0.5 with a 10 °C evaporator temperature, which
means that half of the exhaust gas energy can be turned into refrig-
eration power. Approximately 330 kW of cooling power is gener-
ated with 75% engine load during the scantling draught.
This means that approximately 2.7 tons of ice might be pro-
duced within an hour for cooling purposes. When operating at
lower engine loads the ice production capacity would still stay
high. Approximately 1.7 tons of ice can be produced within an hour
with a 40% engine load. The ammonia-water absorption refrigera-
tion system using exhaust gases as energy source would be suit-
able for fishing ships. It is estimated by the Food and Agriculture
Organization of the United Nations that 50.2 kg of ice is needed
for chilling fresh fish in a 90 L insulated container [24]. The cooling
power production is high enough for AC demands during all engine
loads and the surplus cooling power capacity during high engine
loads could be used for ice production. For example, during scant-
ling draught with 74.4% ME load approximately 180 kW of cooling
power could be used for ice production purposes, which is enough
to make over 1.4 tons of ice within an hour, while cooling all of the
cabins with a power of 150 kW.
7. Usable cooling energy in a bulk carrier and fuel savings
In ISO conditions, the refrigeration powers during high engine
loads (74.4% of engine capacity) provide cabin cooling power for
1400
1200
Cooling power [kW]
1000
800
600
400
200
0
25% 35% 45%
Fig. 14. The cooling power demand and production of different water-LiBr absorption heat recovery solutions with ME load variation.
513
2 Me charge air cooler
26 – 37 °C
over 3.5 times more than required by the case ship studied by
using exhaust gases or jacket water as energy sources and water-
LiBr as the working fluid (Fig. 14). During low ME usage (25.0%
of engine capacity) the cooling capacity still exceeds the require-
ments of the case ship by 1.5 times. The jacket water heat recovery
system has the potential to provide 2.2–4.0 times more cooling
power than is required for AC cooling, depending on the engine
load. For the exhaust gas heat recovery system the same examina-
tion gives 1.7–3.6 times higher cooling powers than is used in cab-
ins. Scavenge air offers the highest cooling powers peaks but its
overall cooling power production is not that promising under
lower engine loads. Cooling power demand and all of the cooling
power potential (production) created by the water-LiBr absorption
refrigeration system (depending on the energy source) is shown in
Fig. 14 with 25–100% ME load variation.
Here jacket water seems to be the best energy source for the
water-LiBr absorption refrigeration system in ISO conditions
(Fig. 15). It offers quite stable cooling power during all engine
loads. The temperature variation is slight so the cooling power
can be easily adjusted by changing the mass flow of jacket water
through the generator (Fig. 12). In this study, it was assumed that
the jacket water is not used for fresh water production purposes
and is entirely available for the absorption system. This could be
possible if fresh water was produced by reverse osmosis.
The problem of an absorption refrigeration system using jacket
water is that it has a relatively low temperature (85 °C). In ISO con-
ditions the system works efficiently but in tropical conditions there
is no possibility of cooling power production. The reason is that the
absorber and the condenser of the absorption refrigeration system
are cooled with sea-water and its temperature is over ten degrees
higher in tropical conditions than in ISO conditions. This has a huge
effect on the performance of the system. Jacket water should be
approximately 15 degrees warmer to work in tropical conditions.
When using exhaust gases as an energy source for absorption
refrigeration with water-LiBr, this problem is avoided. The cooling
Exhaust gas heat
recovery
Jacket water heat
recovery
Scavenge air heat
recovery
Cooling demand
55% 65% 75% 85% 95%
ME load [%]
514 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516

800

700

600
Cooling power

Power [kW]
500
Heang power of
400
exhaust gases
300 Cooling demand

200

100

0
25% 35% 45% 55% 65% 75% 85% 95%
ME load [%]

Fig. 15. The cooling demand and power production potential of the ammonia-water absorption refrigeration solution in exhaust gases during ME load variation.

Table 7
The cooling power potential of the water-LiBr absorption refrigeration system with different energy sources and ME loads (at a 5 °C evaporator temperature).

ME load Yearly operation AC cooling power The cooling potential of exhaust The cooling potential of jacket The cooling potential of
(%) hours (h) demand (kW) gases (kW) water (kW) scavenge air (kW)
74.4 1226.4 150 533 599 853
39.0 1226.4 150 335 410 116
67.8 919.8 150 478 544 580
35.1 919.8 150 306 379 55
63.7 919.8 150 478 544 580
25.0 919.8 150 252 323 0
0 2664.0 150 0 0 0

potential is a bit less in ISO conditions compared to the jacket
water solution but it still exceeds the needs of the case ship (when
the ME is on). In tropical conditions the absorption refrigeration
system using exhaust gases is the best solution.
Absorption refrigeration system with ammonia-water working
pair provides possibility to reach below-zero temperature cooling
levels. It could be used for AC cooling purposes as well as water-
LiBr refrigeration system but it is not as efficient. The cooling
power of the ammonia-water absorption refrigeration system
using exhaust gases is high enough to reach the cooling demand
of cabins (in a B.Delta37 bulk carrier). The rest of the power could
be used for ice-production purposes in fishing boats, for example
or for medical and food preservation needs. It is estimated that
approximately 1.4 tons of ice could be produced within an hour
whilst also chilling the cabins during high loads of the ME (74.4%
of the engine capacity). The cooling demand and power production
potential of the ammonia-water refrigeration solution compared to
the heating power of exhaust gases during 25–100% engine load
variations are shown in Fig. 15.
The cooling power potential of absorption cycles can be esti-
mated with steady-state calculations. Dynamic simulations are
needed to make more accurate estimations of cooling energy pro-
duction. In this study none of the heat exchanger areas were fixed
and it was assumed that all energy could always be transferred in
every heat exchanger of the absorption cycle (absorber, generator,
condenser, and evaporator). The changes in the mass flows’ and
temperatures’ effects on the heat exchangers should be taken into
account. Moreover, the temperature changes inside the absorption
system affect efficiency. In this study all temperatures inside the
system were assumed to be uniform and no changes were taken
into account when the temperature or mass flow of the energy
source changes in the generator (steady-state calculation). These
adjustments were made to simplify the case for calculating general
cooling potential examination in a short time.
One challenge of absorption refrigeration systems in ship usage
is the motion of the ship. For example the generator should be
horizontal and the movement caused by the waves may deterio-
rate the performance of the system. [25] Any condition that is
off-design increases the irreversibility of the absorption cycle,
which harmfully affects on the performance of the system [2]. It
was observed in a German naval test that an approximately 14%
reduction in the system performance happened in a continuous
±20° dynamic heel compared to normal operation [26]. However
this study shows that even with 40% lower performance of the sys-
tem, it is still possible to produce enough cooling power for the
needs of a bulk carrier in ISO conditions. Further challenges with
an absorption chiller application for cargo ships would also be to
integrate the chiller in the cooling plant, since the current set-up
usually consists of direct expansion cooling compressors. Once uti-
lizing absorption chillers for cooling power production, the exist-
ing system could not be used, but the entire plant should
probably be modified to the chiller-type, which might add the cost
and complexity substantially. Also the COP of the cooling power
produced in a direct expansion compressor plant is higher, when
compared to a chiller compressor plant, due to the secondary cir-
cuit between the compressor and the air-handling unit.
ME loads, their yearly operation hours and AC cooling demand
were presented previously in Table 4. They are compiled with cool-
ing power demand (150 kW) and the potential generated by water-
LiBr absorption refrigeration in different heat recovery systems (at
a 5 °C evaporator temperature) and presented in Table 7.
The cooling powers (Table 7) are multiplied with the yearly
operation hours of different ME loads and added up to calculate
total yearly cooling energy demand and production. The result is
Eq. (46):
X Qe  t
Etot ¼ ; ð46Þ
1000
where Etot ðMW hÞ is the total energy production or demand and
tðhÞ is yearly operation hours. The cooling energy demand and pro-
duction with different ME loads and heat recovery solutions using
the water-LiBr absorption refrigeration system are shown in
 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516 515

1200.0

1000.0

Cooling energy [MWh/a]

800.0 Cooling energy demand

600.0 Cooling energy by exhaust gas

heat recovery system
400.0 Cooling energy by jacket water
heat recovery system
200.0 Cooling energy by scavenge air
heat recovery system
0.0
0% 25.0 % 35.1 % 39.0 % 63.7 % 67.8 % 74.4 %
ME loads [%]

Fig. 16. The cooling energy demand and production during a year with different ME loads and heat recovery solutions using a water-LiBr absorption refrigeration system in
ISO conditions.

1000.0
900.0
Cooling energy [MWh/a]

800.0
700.0
600.0 Cooling energy demand
500.0
400.0 Cooling energy by exhaust gas heat
300.0 recovery system
200.0 Cooling energy by scavenge air heat
100.0 recovery system
0.0
0% 25.0 % 35.1 % 39.0 % 63.7 % 67.8 % 74.4 %
ME loads [%]

Fig. 17. The cooling energy demand and production during a year with different ME loads and heat recovery solutions using water-LiBr absorption refrigeration system in
tropical conditions.

Fig. 16 (under ISO conditions). They are calculated by multiplying increased to over 70%. The efficiency of electricity production by
the cooling powers with the yearly operation hours (of different diesel combustion is estimated to be 37%. This means that approx-
ME loads) without the summary presented in Eq. (46). It can be imately 792 MW h of primary energy is needed for cooling pur-
noted that the cooling potential is much higher than the needs of poses yearly and an absorption refrigeration solution could save
the bulk carrier in the estimated operation conditions. Exhaust approximately 46.8 tons (42.7 MJ/kg) of fuel every year.
gas and jacket water cooling production covers all cooling needs A similar study was performed for the absorption refrigeration
during the ME operation hours. This means that approximately system in tropical conditions (with water-LiBr working fluid).
70% of the cooling need can be covered with these solutions yearly The cooling demand is estimated to be 350 kW and the electricity
when the hours that the ME is off are taken into account. Scavenge consumption 77.8 kW when the cooling power is produced by
air heat recovery system covers approximately 50% of the total vapor compression [22]. The cooling energy demand and produc-
energy need, and the cooling potential is large at high engine loads, tion during a year with different ME loads and heat recovery solu-
but in this case it cannot be exploited because the cooling demand tions using a water-LiBr absorption refrigeration system in tropical
remains constant. Absorption refrigeration system in exhaust gases conditions are shown in Fig. 17. In tropical conditions the cooling
with ammonia-water working pair could cover also 70% of the cool- demand can only be satisfied entirely during high engine loads.
ing demand yearly. However, the COP is lower compared to the The absorption refrigeration system using exhaust gases in tropical
water-LiBr system. If below zero temperatures are needed, then conditions has the potential to cover 61% of the cooling energy
the ammonia-water working pair is the only option between these demand of the case ship. This equals approximately 95.0 tons of
two working pairs. fuel savings, which is twice the savings achieved in ISO conditions.
The cooling power of a B.Delta37 bulk carrier is produced by The total cooling energy coverage is less in tropical conditions
vapor compression. Compressors consume 33.3 kW electricity to compared to ISO conditions but the cooling energy demand is 2.3
chill the cabins with the power of 150 kW [22]. Electricity con- times higher, which enables higher savings.
sumption is approximately 293 MW h yearly. Using water-LiBr
absorption refrigeration in exhaust gases or in a jacket water heat
recovery system it is possible to save approximately 204 MW h 8. Conclusions
electricity (70% of the present compression electricity consump-
tion). This does not take into account the increased pumping State-of-art literature study, fundamentals of absorption refrig-
power needed for chilling the absorber and condenser with sea- eration, and detailed presentation of thermodynamic modeling of
water. However, the saving possibilities are significant. If the water-LiBr and ammonia-water working pairs is made in order to
exhaust gases of auxiliary engines are utilized for cooling purposes study the energy production possibilities from waste heat recovery
while the ME is off, the rate of cooling energy coverage can be for cooling power production in a bulk carrier ship (B.Delta37).
516 W. Salmi et al. / Applied Thermal Engineering 115 (2017) 501–516
Exhaust gases, jacket water, and scavenge air of the main engine
(ME) are used as energy sources, when the ship is operated in an
annual average operation profile on ISO and tropical conditions.
Of the two alternatives studied, the water-LiBr working pair
using exhaust gases or jacket water as energy sources is the best
absorption chiller solution in ISO conditions for energy saving pur-
poses on ships. Ammonia-water working fluid is more suitable for
below-freezing-point cooling requirements. Both solutions offer
more cooling power than is required (while the ME is on) in the
case ship, but the LiBr-water working pair is more efficient. This
means that 70% of the cooling energy required can be produced
by absorption refrigeration. Moreover, this percentage value could
be higher if the exhaust gases of auxiliary engines could be utilized
when the ME is not being operated. This option was not studied.
This study shows that there is a possibility to save 70% of com-
pression electricity usage by using an absorption refrigeration
solution in exhaust gases or in jacket water in ISO conditions if
all of the technical difficulties can be overcome. It means that
approximately 46.8 tons of fuel could be saved every year. Cooling
possibilities were also studied in tropical conditions. The water-
LiBr absorption refrigeration system in exhaust gases seems to be
the best solution for air-conditioning cooling purposes in tropical
conditions. It has the potential to cover 61% of the cooling energy
needs of the case ship. In tropical conditions that means approxi-
mately 95.0 tons of fuel savings every year. The fuel savings
increase by over 100% in tropical conditions compared to ISO con-
ditions, although the total cooling energy coverage of the absorp-
tion refrigeration system is less.
Acknowledgments
This work is supported by the Finnish Funding Agency for Inno-
vation (Tekes) Arctic Seas Programme. Thanks are also given to the
companies participating in the Ship Energy Efficiency Technologies
(SET) project.
References
[1] V. Eyringer, H.W. Kohler, A. Lauer, B. Lembper, Emissions from international
shipping: 2. Impact of future technologies on scenarios until 2050, J. Geophys.
Res. 110 (2005) D17306.
[2] T. Cao, H. Lee, Y. Hwang, R. Radermacher, H.-H. Chun, Performance
investigation of engine waste heat powered absorption cycle cooling system
for shipboard applications, Appl. Therm. Eng. 90 (2015) 820–830.
[3] F. Baldi, C. Gabrielii, A feasibility analysis of waste heat recovery systems for
marine applications, Energy 80 (2015) 654–665.
[4] G. Shu, Y. Liang, H. Wei, H. Tian, J. Zhao, L. Liu, A review of waste heat recovery
on two-stroke IC engine aboard ships, Renew. Sustain. Energy Rev. 19 (2013)
385–401.
[5] J. Fernandez-Seara, A. Vales, M. Vazquez, Heat recovery system to power an
onboard NH3-H2O absorption refrigeration plant in trawler chiller fishing
vessels, Appl. Therm. Eng. 18 (1998) 1189–1205.
[6] C. Ezgi, Design and thermodynamic analysis of an H2O-LiBr AHP system for
naval surface ship application, Int. J. Refrig. 48 (2014) 153–165.
[7] K. Ebrahimi, G.F. Jones, A.S. Fleischer, Thermo-economic analysis of steady
state waste heat recovery in data centers using absorption refrigeration, Appl.
Energy 139 (2015) 384–397.
[8] A.A. Menzela, S.M. Hanriot, L. Cabezas-Gomez, J.R. Sodre, Using engine exhaust
gas as an energy source for an absorption refrigeration system, Appl. Energy 87
(2010) 1141–1148.
[9] Y. Kaita, Thermodynamic properties of lithium bromide-water solutions at
high temperatures, Int. J. Refrig. 24 (2001) 374–390.
[10] F.L. Lansing, Computer Modeling of a Single-Stage Lithium Bromide/Water
Absorption Refrigeration Unit, DSN Engineering Section, JPL Deep Space
Network Progress Report 42-32, pp. 247–257.
[11] H.Y. Kim, B.B. Saha, S. Koyama, Development of a slug flow absorber working
with ammonia-water mixture: part I – flow characterization and experimental
investigation, Int. J. Refrig. 26 (2003) 508–515.
[12] H.Y. Kim, B.B. Saha, S. Koyama, Development of a slug flow absorber working
with ammonia-water mixture: part II – data reduction model for local heat
and mass transfer characterization, Int. J. Refrig. 26 (2003) 698–706.
[13] S. Garimella, A.M. Brown, A.K. Nagavarapu, Waste heat driven
absorption/vapor-compression cascade refrigeration system for megawatt
scale, high-flux, low-temperature cooling, Int. J. Refrig. 34 (2011) 1776–1785.
[14] B. Babu, G. Maruthi Prasad Yadav, Performance analysis of lithium-bromide
water absorption refrigeration system using waste heat of boiler flue gases,
Int. J. Eng. Res. Manage. (IJERM) 2 (2) (2015).
[15] A.A. Manzela, S.M. Hanriot, L. Cabezas-Domez, J.R. Sodre, Using engine exhaust
gas as energy source for an absorption refrigeration system, Appl. Energy 87
(2010) 1141–1148.
[16] H. Watanabe, Field Experience with Large Absorption Heat Pump (Type I) and
Heat Transformers (Type II), Sanyo Electric Tokki Co., Ltd, Tokyo, Japan.
[17] E.W. Lemmon, M.L. Huber, M.O. McLinden, NIST Standard Reference Database
23: Reference Fluid Thermodynamic and Transport Properties-REFPROP,
Version 9.1, National Institute of Standards and Technology, Standard
Reference Data Program, Gaithersburg, 2013.
[18] J. Patek, J. Klomfar, Simple functions for fast calculations of selected
thermodynamic properties of the ammonia-water system, Int. J. Refrig. 18
(1995) 228–234.
[19] C. Rubio-Maya, J.J. Pachero-Ibarra, J.M. Belman-Flores, S.R. Galvan-Gonzales, C.
Mendoza-Covarrubias, NLP model of a LiBr-H2O absorption refrigeration
system for the minimization of the annual operating cost, Appl. Therm. Eng.
37 (2012) 10–18.
[20] S. Aphornratana, I.W. Eames, Thermodynamic analysis of absorption
refrigeration cycle using the second law of thermodynamics method, Int. J.
Refrig. 18 (4) (1995) 244–252.
[21] A. Ouadha, Y. El-Gotni, Integration of an ammonia-water absorption
refrigeration system with a marine diesel engine: a thermodynamic study,
Proc. Comp. Sci. 19 (2013) 754–761.
[22] M. Elg, M. Kuosa, K. Tammi, P. Mäkipeska, A. Kinnunen, P. Rahkola, Supporting
the energy efficient ship design with energy flow simulations: case efficient
cooling water system, Energy Efficient Ships 2015, Rotterdam, the
Netherlands, 4th November 2015.
[23] M. Elg, M. Kuosa, M. Lampinen, R. Lahdelma, P. Mäkipeska, J. Raita, Z.
Guangrong, K. Tammi, Advanced auxiliary cooling system for energy efficient
ships, in: Eemods’15 Conference, Energy Efficiency in Motor Driven Systems,
15–17 September 2015, Helsinki, 2015.
[24] Calculations and Examples for Insulated Containers and Fish Holds <www.fao.
org/docrep/006/y5013e/y5013e0a.htm#TopOfPage>, 2016 (accessed 20.01.16).
[25] K.-H. Kim, K.W. Yang, S. Jeong, K. Nam, D. Chang, Absorption refrigeration
system utilising engine exhaust gas for bulk gas carriers, Ships Offshore Struct.
9 (2014) 380–386.
[26] F. Taboas, M. Bourouis, M. Valles, Analysis of ammonia/water and ammonia/
salt mixture absorption cycles for refrigeration purposes in fishing ships, Appl.
Therm. Eng. 66 (2014) 603–611.