12th International Symposium On District Heating and Cooling

12thInternationalSymposiumonDistrictHeating
andCooling
September5
th
September7
th
,2010
Tallinn,ESTONIA
ISBN: 978-9949-23-015-0

PREFACE
The 12
th
International Symposium on District Heating and Cooling is now held for the first time in Tallinn, Estonia.
District heating systems are dominant to heat the buildings in Estonia. Estonian district heating systems are with
small, average and big loadings annual thermal loading from less than 5000 MWh to over 100 000 MWh. The
largest district heating networks are situated in Tallinn, Tartu, Narva and Prnu. The district heating is organized by
municipalities. The development and implementation of the energy policy is organized by the Ministry of Economic
Affairs and Communications and the energy market is supervised by Energy Market inspectorate.
Preservation of district heating system in working order is the basic precondition for combined heat and power
generation, accordingly for fuel consumption and environment pollution reduction.
Additionally to Tallinn and Narva the heat produced with combined production is used for district heating in
Kohtla-Jrve and Ahtme and also in some plants with remarkable lower capacity.
New green-field bio fuel and peat fired combined heat and power plants nearby Tallinn and Tartu in Luunja. The
capacities of theirs plants are as follows: 25 MW
el
and 50 MW
th
(in Tallinn CHP with condenser to 70 MW
th
). The
construction new Fortum bio fuel and peat fired CHP in Prnu. Plant will start the operation in the end of 2010.
CHP power plants can also be an efficient source to supply district cooling. This is very promising concept for
conditions where both low winter temperatures and high summer temperatures prevail. Even in Estonia, with just a
few weeks of hot weather in normal summer (not, as this year`s summer), large shopping centres and office buildings
seem to be a good application area.
The research in the district heating and cooling field is very important and at this symposium we will hear forty five
technical presentations, divided into ten sessions: conceptions and studies in district heating and cooling, efficiency
issues of district heating and cooling, district heating in areas for low density, district heating systems-pipes
properties, renewable district heating and cooling, the aspects of district heating-CHP, district heating-case studies
and optimisation and stimulation in district heating. Two plenary speakers cover the activities on district heating and
cooling in the European Union: the prospects for district heating and cooling seen from the EU commission point of
view and the prospect of district heating and district heating research in Germany.
This symposium is organized at the Tallinn University of Technology in cooperation with the Scientific Committee
consisting of members of a Nordic Research group called Primary Energy Efficiency (PEE), one member from Tallinn
University of Technology and Riga Technical University.
On behalf of the organizers we want to express our sincere thanks to the members of the Scientific Committee, the
Local Organizing Committee and the members of the Advisory Committee who gave us valuable support. We also
thank Nordic Energy Research and Tallinn University of Technology for financial support. We would like to thank the
individual authors for their submitted papers and the reviewers for their time and help.
We hope that you will experience an enjoyable stay in our sweet old Tallinn and also that this conference will improve
further cooperation in the field of district heating and cooling research and development.

Aadu Paist, professor
Andres Siirde, professor

TABLE OF CONTENTS
PREFACE ...................................................................................................................................................................... 1
INTEGRATION OF AN IP BASED LOW-POWER SENSOR NETWORK IN DISTRICT HEATING SUBSTATIONS ... 4
J. Gustafsson, H. Mkitaavola, J. Delsing and J. van Deventer
ON THE RADIAL CONTACT PRESSURE OF PARALLEL BURIED PIPES FOR DISTRICT HEATING ................... 12
I. Weidlich, M. Achmus
ANALYSIS ON FLAT STATION CONCEPT. PREPARING DHW DECENTRALISED IN FLATS .............................. 16
Thorsen, Jan Eric
IMPROVED TEMPERATURE PERFORMANCE OF RADIATOR HEATING SYSTEM CONNECTED TO DISTRICT
HEATING BY USING ADD-ON-FAN BLOWERS ........................................................................................................ 22
Per-Olof Johansson, Janusz Wollerstrand
PRIMARY ENERGY EFFICIENCY AND SYSTEMS ENGINEERING ......................................................................... 31
M.Berner, R. Ulseth, J.Stang
ENHANCED DISTRICT HEATING AND COOLING SYSTEMS REALISATION OF THE LOW-EX CONCEPT ..... 39
Stefan Bargel, Clemens Pollerberg, Armin Knels, Li Huang, Dirk Mller and Christian Dtsch
APPLICATION OF EXERGOECONOMICS TO THE OPTIMIZATION OF BUILDING HEATING SYSTEMS
CONNECTED TO DISTRICT HEATING NETWORKS ................................................................................................ 45
C. W. Snoek and S. C. Kluiters
SLIMNET: AN INNOVATIVE INTEGRAL APPROACH FOR IMPROVING EFFICIENSIES OF DISTRICT HEATING
NETWORKS ................................................................................................................................................................ 53
M. W. P. van Lier
A DIRECT HEAT EXCHANGER UNIT USED FOR DOMESTIC HOT WATER SUPPLY IN A SINGLE-FAMILY
HOUSE SUPPLIED BY LOW ENERGY DISTRICT HEATING .................................................................................... 60
Marek Brand, Jan Eric Thorsen, Svend Svendsen and Christian Holm Christiansen
CHALLENGES ON LOW HEAT DENSITY DISTRICT HEATING NETWORK DESIGN ............................................. 69
M. Rm and K. Sipil
DESIGN OF LOW TEMPERATURE DISTRICT HEATING NETWORK WITH SUPPLY WATER RECIRCULATION 73
Hongwei Li, Alessandro Dalla Rosa, Svend Svendsen
STEADY STATE HEAT LOSSES IN PRE-INSULATED PIPES FOR LOW-ENERGY DISTRICT HEATING ............ 81
A. Dalla Rosa, H. Li, S. Svendsen
TRANSIENT THERMAL CONDUCTIVITY OF FLEXIBLE DISTRICT HEATING TWIN PIPES ................................. 90
C. Reidhav and J. Claesson
DISTRICT HEATING PIPES 200 MM BELOW SURFACE IN A STREET WITH HEAVY TRAFFIC .......................... 96
Anders Fransson

and Sven-Erik Sllberg
STUDY ON THE HEAT LOSS REDUCTION METHOD FROM THE SECONDARY PIPELINES IN THE APARTMENT
COMPLEX .................................................................................................................................................................. 105
Byung-Sik Park, Yong-Eun Kim,

Sung-Hwan Park, Yong-Hoon Im, Hyouck-Ju Kim, Dae-Hun Chung,
Mo Chung
HEAT LOSS OF FLEXIBLE PLASTIC PIPE SYSTEMS, ANALYSIS AND OPTIMIZATION ..................................... 112
EJ.H.M. van der Ven, R.J. van Arendonk

COMPARISON OF COMPETITIVE (SEMI) FLEXIBLE PIPING SYSTEMS BY MEANS OF HEAT LOSS
MEASUREMENT ....................................................................................................................................................... 119
I.M. Smits, J. Korsman, J.T. van Wijnkoop and E.J.H.M. van der Ven
EFFECTIVE WIDTH THE RELATIVE DEMAND FOR DISTRICT HEATING PIPE LENGTHS IN CITY AREAS .. 128
Urban Persson, Sven Werner
INTEGRATING RENEWABLE ENERGY INTO LARGE-SCALE DISTRICT HEATING SYSTEMS .......................... 132
Peter Begerow, Dr. Stefan Holler
SOLAR DISTRICT HEATING (SDH): TECHNOLOGIES USED IN LARGE SCALE SDH PLANTS IN GRAZ
OPERATIONAL EXPERIENCES AND FURTHER DEVELOPMENTS ...................................................................... 140
M. Schubert, C. Holter and R. Soell
BIOENERGY COMBINES IN DISTRICT HEATING SYSTEMS: PROSPECTS FOR A FUTURE GROWTH
INDUSTRY? ............................................................................................................................................................... 143
E. Axelsson, A. Sandoff, C. Overland
SEA WATER DISTRTICT COOLING FEASIBILITY ANALYSIS FOR TALLINN ....................................................... 153
A. Hani, I. Britikovski, H. Voll

and T.-A. Kiv
ANALYSIS FOR THE OPERATION BEHAVIOR AND OPTIMIZATION OF CHP SYSTEM IN DISTRICT HEATING
AND COOLING NETWORK ....................................................................................................................................... 157
Yong Hoon Im,

Hwa-Choon Park, Byung-Sik Park and Mo Chung

IMPROVED PRIMARY ENERGY EFFICIENCY OF DISTRICT HEATING NETWORKS BY INTEGRATION OF
COMMUNAL BIOMASS-FIRED COMBINED HEAT AND POWER PLANTS WITH BIOMASS PYROLYSIS ........... 168
T. Kohl, N.A. Pambudi, T. Laukkanen and C.-J. Fogelholm
CHP OR POWER STATION? QUESTION FOR LATVIA ....................................................................................... 177
D. Blumberga, G. Kuplais, F. Romagnoli and E. Vigants
LCA OF COMBINED HEAT AND POWER PRODUCTION AT HELLISHEII GEOTHERMAL POWER PLANT WITH
FOCUS ON PRIMARY ENERGY EFFICIENCY ........................................................................................................ 184
Marta Ros Karlsdottir, Olafur Petur Palsson, Halldor Palsson
FLEXIBILITY FROM DISTRICT HEATING TO DECREASE WIND POWER INTEGRATION COSTS .................... 193
J. Kiviluoma and P. Meibom
DAILY HEAT LOAD VARIATION IN SWEDISH DISTRICT HEATING SYSTEMS .................................................... 199
H. Gadd and S. Werner
DISTRICT HEATING AS PART OF THE ENERGY SYSTEM: AN ENVIRONMENTAL PERSPECTIVE ON
PASSIVE HOUSES AND HEAT REPLACING ELECTRICITY USE ....................................................................... 202
Morgan Frling and Ingrid Nystrm
ADAPTIVE CONTROL OF RADIATOR SYSTEMS FOR A LOWEST POSSIBLE RETURN TEMPERATURE ........ 206
P. Lauenburg and J. Wollerstrand
POLICIES AND BARRIERS FOR DISTRICT HEATING AND COOLING OUTSIDE EU COUNTRIES ................... 215
A. Nuorkivi and B. Kalkum
BARRIERS TO DISTRICT HEATING DEVELOPMENT IN SOME EUROPEAN COUNTRIES ............................... 223
Dag Henning and Olle Mrdsj
IMPACT OF THE PRICE OF CO2 CERTIFICATES ON CHP AND DISTRICT HEAT IN THE EU27 ...................... 229
Markus Blesl
CONSIDERATIONS AND CALCULATIONS ON SYSTEM EFFICIENCIES OF HEATING SYSTEMS IN BUILDINGS
CONNECTED TO DISTRICT HEATING .................................................................................................................... 238
Maria Justo Alonso, Rolf Ulseth and Jacob Stang
HEAT LOAD REDUCTIONS AND THEIR EFFECT ON ENERGY CONSUMPTION ................................................ 244
Christian Johansson and Fredrik Wernstedt
VERIFICATION OF HEAT LOSS MEASUREMENTS ............................................................................................... 250
J.T. van Wijnkoop, E. van der Ven
DISTRICT HEATING AND COOLING WITH LARGE CENTRIFUGAL CHILLER-HEAT PUMPS ............................. 258
Ulrich Pietrucha
NEW ECONOMICAL CONNECTION SOLUTION FOR FLEXIBLE PIPING SYSTEMS ........................................... 261
Christian Engel, Gerrit-Jan Baars
COMPETITIVENESS OF COMBINED HEAT AND POWER PLANT TECHNOLOGIES
IN ESTONIAN CONDITIONS..................................................................................................................................... 267
E. Latov

and A. Siirde
DISTRIBUTION OF HEAT USE IN SWEDEN ........................................................................................................... 273
Margaretha Borgstrm, Sven Werner
DAMAGES OF THE TALLINN DISTRICT HEATING NETWORKS AND INDICATIVE PARAMETERS FOR AN
ESTIMATION OF THE NETWORKS GENERAL CONDITION .................................................................................. 277
Aleksandr Hlebnikov, Anna Volkova, Olga Duba, Arvi Poobus, lo Kask
EFFICIENCY OF DISTRICT HEATING WATER PUMPING IN FINLAND ................................................................ 283
Antti Hakulinen, Jarkko Lampinen and Janne Lavanti
MODELLING DISTRICT HEATING COOPERATIONS IN STOCKHOLM AN INTERDISCIPLINARY STUDY OF A
REGIONAL ENERGY SYSTEM ................................................................................................................................. 288
D. Magnusson, D. Djuric Ilic
CUTTING COSTS OF DISTRICT HEATING SYSTEMS BY USING OPTIMIZED LAYING TECHNIQUES ............. 297
Alexander Goebel, Dr. Stefan Holler
ANALYSIS OF HEAT TRANSFER IN HEAT EXCHANGERS BY USING THE NTU METHOD AND EMPIRICAL
RELATIONS ............................................................................................................................................................... 305
O. Gudmundsson, O. P. Palsson and H. Palsson
HEAT LOSS ANALYSIS AND OPTIMIZATION OF A FLEXIBLE PIPING SYSTEM ................................................. 310
J. Korsman, I.M. Smits and E.J.H.M. van der Ven
FREE OPTIMIZATION TOOLS FOR DISTRICT HEATING SYSTEMS .................................................................... 318
Stefan Gnchtel, Sebastian Gro
The 12th International Symposium on District Heating and Cooling,
September 5
th
to September 7
th
, 2010, Tallinn, Estonia

4

INTEGRATION OF AN IP BASED LOW-POWER SENSOR NETWORK
IN DISTRICT HEATING SUBSTATIONS
J. Gustafsson, H. Mkitaavola, J. Delsing and J. van Deventer

Div. of EISLAB, Dept. of Computer Science and Electrical Engineering
Lule University of Technology, 971 87 Lule, SWEDEN

ABSTRACT
In this study, the implementation of a wireless, low-
power, sensor network with IP capabilities in a district
heating substation was evaluated. The aim of the
study was to show that an open standard solution is
technically feasible. Low-power wireless
communication was established using IPv6/6LoWPAN
on an IEEE 802.15.4 wireless network. An
experimental district heating substation was equipped
with sensor platforms in vital devices located within or
near a district heating substation. As a result, all
connected devices could obtain a direct internet
connection.
A system with open standards facilitates the
introduction of new energy services such as individual
measurements and improved space heating control.
In this study, we found that resource-limited battery-
powered devices possess a life expectancy of over
10 years, using small batteries while participating in
IPv6 compatible communication.
INTRODUCTION
Embedding low-power wireless devices in district
heating substations and surrounding equipment such
as temperature sensors could provide useful services
to consumers and producers. Currently, many
different substation control systems on the market can
connect to the internet and have various wireless
sensor reading systems. However, these systems
tend to be specialized and are only compatible with
equipment from the same manufacturer. Moreover,
internet-compatible control systems are often also
relatively expensive, and provide bad scalability.
In general, commercially available heat meters cannot
communicate through the current infrastructure; thus,
specialized communication methods such as mbus,
pulse, and infrared readings must be employed.
Therefore, poor communication standards limit the
current usage of heat meters and other equipment in
the substation. However, by sharing information with
other devices in the substation, the heat meter could
provide useful feedback and sensing information,
which can be used to improve the substation control
functionality.
Fig. 1 provides an overview of the development of
sensor networks over the last 20 years. Unfortunately,
most equipment currently used in district heating
substations is antiquated.

Fig. 1. Evolution of wireless sensor networks. Although
the scalability of the sensor network has increased, many
industries still use vendor-specific cable solutions. (The
figure was obtained from the literature [1])

If heat meters, control systems, and other non-district
heating equipment could communicate, new services
that have impact on both economy and the
environment could be developed.
The infrastructure required to achieve wireless device
communication may be attained with low-power
wireless technology. Small sensor platforms with
direct internet access through standardized wireless
technology can provide a solid platform for new
services.
A lack of standardized communication protocols is
commonly encountered when connecting electronic
devices from different vendors. In general, devices
manufactured by different companies use different
communication protocols, which limits the functionality
of the substation.
District heating substations can be divided into
sections based on metering, space heat control, and
tap water control. For a visual overview of a common
parallel connected district heating substation, see
Fig. 2, this is also the substation type used in the
study. Typically, information is not shared between
these sections; thus, each system can only be locally
optimized. To achieve complete substation
optimization, information must be shared between
sections. To this end, wireless sensor-platforms were
installed in temperature-sensors, heat-meters,
circulation pumps, and control valves, and new control
methods and services were tested. This empowers us
September 5
th
to September 7
th

5

to develop new control methods, and implementing
new services to heat suppliers, building owners and
end users.

Fig. 2. A systematic overview of a parallel coupled
district heating substation divided into three sections:
metering, heating and hot water system.

SERVICES
To control or reduce their energy bill, district heating
customers require specific information to determine
the appropriate action. Currently, the only information
available to the customer is the information provided
in the bill or on the heat-meter display.
If information on all devices was available online,
customer could easily monitor their usage and interact
with the substation. Examples of services that could
be provided by the substation are explained in the
following sections.
Improved substation control
Combined heat and power plants are becoming more
common; thus, the importance of the distribution
system T is increasing. In a combined heat and
power plant with a flue-gas condensation system, a
high T is even more important to obtain satisfactory
fuel efficiency.
To maintain high energy efficiency, the hot water
produced by the plant must be delivered to customers
with a minimal heat loss. Once the hot water is
transported to the customer, a maximum amount of
energy per volume of water should be extracted and
used for heating purposes, such as hot tap water and
space heating. To achieve a maximum T, energy
transfer between the distribution medium to the point
of consumption should be maximized, while the
temperature of the returning distribution medium
should be minimized.
Unfortunately, there are many challenges in
maintaining the efficiency of a district heating network.
Problems related to the equipment that controls the
temperature of radiator water and hot tap water are
often encountered. These devices tend to be
calibrated to satisfy the desires of the customer only;
thus, the effects on the energy efficiency of the entire
district heating system are often ignored. One key
factor in obtaining a high T across a district heating
substation is the radiator circuit supply temperature.
The radiator circuit supply temperature does not only
affect the indoor comfort, but also the primary return
temperature as the returning radiator circuit media
cools the primary media through the heat exchanging
unit. Specifically, water returned from the radiator
circuit cools the primary supply through the heat
exchange unit. Currently, the radiator circuit supply
temperature is based on the local outdoor
temperature, which produces a stable indoor
temperature. However, the primary supply
temperature also affects the ideal radiator supply
temperature and the radiator circuit flow. The
relationship between outdoor temperature and primary
supply is often assumed to be linear (colder outdoor
air leads to a warmer primary supply). However,
significant deviations from the ideal curve are
common. More information on the effect of primary
supply temperature and radiator control on the indoor
air temperature and T of the system can be found
in [2].
Adaptive radiator control is another intelligent way of
controlling the radiator circuit and obtaining a high T.
More information on this method can be found in
previous studies by Lauenburg [3].
Fault detection
Control valves in the district heating substation often
possess inappropriate dimensions, resulting in
intermittent control, pressure shocks, and high return
temperatures. Due to the high thermal time constant
of a building, the indoor temperature is not directly
affected. Therefore, an error in the control valve may
go unnoticed for a considerable amount of time.
Error identification can be achieved by evaluating high
frequent meter readings, which to some extent are
done today.
A fouling valve that is stuck or does not move in
accordance with the control signal may also be difficult
to detect. A direct comparison of the valve control
signal with the heat meter, which measures the
primary flow through the district heating substation,
can be used to identify a broken fouling valve [4].
September 5
th
to September 7
th

6

Individual measurements
Individual measurements are common in some
countries and are gaining interest in others. To obtain
measurements of each apartment, tap point, or
radiator, new metering devices must be installed. The
most straight forward method is to install flow meters
at each tap point and/or radiator. In general, high
resolution flow meters are quite expensive; thus,
installing one on every tap point/radiator can be cost-
prohibitive.
An alternative method has been evaluated by Yliniemi
[5]. In this method, temperature sensors were
installed at each tap point, and one central flow meter
was used to measure the flow through a section,
which contained up to 40 tap points. The flow
recorded by the meter and the temperature measured
at the tap points were synchronized, and the integrity
of each tapping point was verified by installing
inexpensive temperature sensors at each site and a
limited number of central flow meters throughout the
building.
Load balancing
Dynamic load balancing is a method used to remove
heat load peaks and divide power consumption
between buildings. Dynamic load balancing is based
on the presence of a large thermal time constant of
each building. For instance, in a building with a high
thermal time constant, the heating system can be
turned off when the price of heat is high or during
peak energy hours. An online automatic and
independent auction system is used to decide which
buildings will be shut down or provided a limited
amount of thermal power. In this system, all
connected buildings are involved in the bidding
process. Specific details on dynamic load balancing
are provided in the literature [6].
Visualized energy efficiency
If a large number of district heating substations were
connected to the internet, the performance of different
substations could be compared. For instance, the
supply/return temperature, T, energy usage, etc. of
all substations could be plotted in a graph, table or
map. Fig. 3 displays a map of the return temperature
of a substation, which allows the consumer to
compare the performance of their house to others in
the area. Moreover, the map provides the utility
company with an overview of the network and
improves the detection of leaks and short circuits.
Moreover, the utility company can identify
deteriorating substations or individual installations that
perform poorly.
NETWORK TECHNOLOGY
A common method of visualizing a network
communication protocol is in the form of stack. A stack
consists of layers that are separated by function; thus,
a communication stack contains different layers of
tasks related to data transportation. The layers can be
divided and visualized in many ways. For example,
the five-layer internet model has been used
extensively in previous studies and is displayed in
Fig. 4 [7]. In this paper, only the layers that are
significant to the results of this research will be
discussed. Thus, the network, link, and physical layers
are considered in more detail.

Fig. 3. Performance of a district heating substation
visualized on a map. The red square can represent the
supply/return temperature, energy usage, or heat flow in
the connected building.

Fig. 4 A generic five-layer internet model and its
implementation in an IEEE 802.15.4 wireless network.

IP (Network Layer)
The internet protocol (IP) is the most well-known and
commonly used network protocol in the world. All
traffic on the internet is currently routed through IP.
Today, there are two co-existing versions of IP,
including IPv4, the older version of IP, and IPv6, the
latest version.
September 5
th
to September 7
th

7

IPv4
Currently, internet protocol version 4 is the most
widely used IP, and almost all computers connected
to the internet use this version.
An IPv4 address is 32 bits long and is typically written
in 4 sections divided by dots (e.g., 192.168.100.123).
The theoretical number of IPv4 addresses is 232
(approximately 4.2 billion); however, a fraction of
addresses is reserved and cannot be used for online
purposes. The total number of usable IPv4 addresses
is approximately 3.7 billion. As the number of devices
connected to the internet increases, IPv4 addresses
are beginning to run out.
Technology such as network address translation
(NAT) and port address translation (PAT) have
postponed the depletion of IPv4 addresses; however,
the number of available IPv4 addresses decreases
every day.
IPV6
IPv6 was developed to compensate for the limited
number of IPv4 addresses. IPv6 uses a longer
address than IPv4 and has several convenient
features. IPv6 uses a 128 bit address, which means
that there are 2128 possible addresses. Thus, the
number of address per square millimeter of the earths
surface is 6.71017. Hopefully, the addresses
obtained through the implementation of IPv6 will last
for a long time.
With the additional address space, it is possible to
give every small device its own unique IP number
without implementing NAT. Thus, direct
communication over the internet can be achieved
without any special gateways. However, the new
address space increases the overhead of data
packages, which negatively impacts small, low-power
devices because more battery energy is wasted on
header data in every wireless data transmission.
However, a new adaptation layer (6LoWPAN) was
developed to limit the amount of lost energy. More
information on 6LoWPAN can be found in the next
section and in [1].
In addition to a wider address space, IPv6 also
includes stateless autoconfiguration, which is a
function that can be used to automatically configure
newly connected devices without any special servers.
To obtain stateless autoconfiguration, newly
connected devices broadcast a router solicitation (RS)
message to every listening device. When a router
receives the message, it responds with a router
advertisement (RA) message. The device adds the
IPv6 prefix from the router to the local link layer
address, creating a complete IPv6 address. To ensure
that another device does not possess the same IP
address, the device broadcasts a neighbor solicitation
message to search for a duplicate address. If another
device has the same IP number, the new device shuts
down.
6LOWPAN
6LoWPAN is an adaptation layer that separates the
network and data link layer of the protocol stack. The
purpose of the layer is to compress IPv6 headers and
minimize unnecessary data transmission while
maintaining IPv6 compatibility. According to the
literature, [8] the 6LoWPAN header uses less than
10% of the total energy used during packet
transmission.
IEEE 802.15.4 physical and data link layers are often
used in combination with 6LoWPAN; however, other
standards can also be applied.
802.15.4 (Link and Physical Layer)
The most common data link and physical layer used
with 6LoWPAN networks is IEEE 802.15.4; however,
6LoWPAN is also compatible with other layers.
Moreover, IEEE 802.15.4 is also the basis for ZigBee,
Wireless HART, and MiWi. The IEEE 802.15.4
standard specifies operation at low frequency bands
such as 868 MHz (EU), 915 MHz (US), and 950 MHz
(JP), and high frequency bands including 2.4 GHz
(World Wide) [9]. The main practical differences
between low and high frequency bands are the
bandwidth and communication range. The 2.4 GHz
band supports a higher bandwidth but the range is
limited, especially in armored concrete buildings. The
low frequency bands have a moderate bandwidth and
a considerably larger range. In a district heating
substation, bandwidth usage can be minimized
because rapid changes are uncommon (compared to
many other control/measurement situations) and low
frequency groups are preferred. However, only 2.4
GHz sensor platforms were available at the beginning
of this study; thus, these platforms were used in most
of the tests.

September 5
th
to September 7
th

8

Fig. 5. Schematic overview of the experimental setup. Squares marked with an N in the 6LoWPAN network are sensor
nodes (motes).

Fig. 6. A stack view of the experimental setup.

Fig. 7. A stack view of how internet connected systems is connected with proprietary communication protocol.
Network setup
The networking hardware used in this study included
sensor nodes, a Linksys WRT54GL router, and an
Ethernet to IEEE 802.15.4 edge router. A schematic
depiction of the experimental setup is provided in Fig 5.
To achieve a network setup that was compatible with
IPv6, some reconfiguration was necessary. Currently,
every internet service provider in Sweden does not
supply native IPv6 support. Unfortunately, the ISP that
was available at the test site did not support native
IPv6. However, IPv6 internet can be accessed by
constructing an IPv6 to IPv4 (6to4) tunnel using
customized firmware for the Linksys broadband router.
In this study, such a tunnel was established in the
router to supply global IPv6 functionality to the LAN
side. A router advertisement daemon (radvd), was also
installed on the router; thus, IPv6 enabled devices were
configured through stateless autoconfiguration. As a
result, full functional IPv6 internet access was provided
to all devices.

The physical connection between the Ethernet network
and the IEEE 802.15.4 wireless network is done at the
edge router (in Fig. 5). Here is where the 6LoWPAN
header compression & decompression is performed to
the passing IPv6 packages. This can also be viewed in
Fig 6, where a stack view of the experimental setup is
depicted. As a comparison to Fig. 6, Fig. 7 illustrates
an example of how vendor specific products currently
connect to the internet, using proprietary protocols,
which makes them incompatible with devices from
other manufacturer. The scalability is hence limited to
devices from the same manufacturer, which tends to be
a short tem solution.
Sensor platforms
The sensor platform used in the study, is the Mulle v5.2
[10]. Other similar available sensor platforms are
among others Micaz [11], AVR Raven [12] and
Sensinode [13]. We choose to work with the Mulle
platform since it has a very good performance to
energy ratio.
September 5
th
to September 7
th

9

To obtain IPv6/6LoWPAN functionality in the Mulles,
the lightweight operating systems Contiki [14] and
TinyOS [15] have been successfully ported to the Mulle
platform. Both operating systems were specifically
designed to be compatible with resource limited
embedded systems such as Mulle. Moreover, Contiki
and TinyOS both support IPv6 and 6LoWPAN.
However, TinyOS was selected for this study because
stability issues due to edge-routing problems with
Contiki.
Sensor platform energy usage
Obtaining an acceptable life expectancy is one of the
biggest challenges to battery powered, wireless
devices. In Sweden, heat meters are inspected every 5
to 10 years, depending on the size of the meter. The
life expectancy of wireless devices should be
equivalent to the inspection period to avoid frequent
and expensive battery replacements. All sensor nodes
do however not need to be battery powered. In the
case of available electric power in close proximity, e.g.
for platforms mounted in pumps or valves there is no
explicit need for batteries since there are electricity
available. At other sensor platforms, battery power is
the only feasible solution, for instance outdoor
temperature sensors.
To determine the amount of energy used by a wireless
sensing device, the current at the sensor platform
associated with IPv6/6LoWPAN communication was
measured. To measure the current used by the device,
a 1 ohm high precision resistor was connected in series
to the Mulle power connector. The voltage drop
generated across the resistor was amplified 100 times
with a MAX4372H amplifier circuit. Using an analog
acquisition card, the amplified signal was measured
and stored in an ordinary PC. Due to poor precision at
very low current, complementary measurements were
performed with a high precision ampere-meter to
determine the current usage of the Mulle, when it was
in deep sleep mode.
To evaluate the energy cost of transmitting data
packets with UDP on IPv6/6LoWPAN, packets with
payload sizes between 1 and 100 bytes were
transmitted, and the expected lifetime of the sensor
was calculated. Fig. 8 displays the expected lifetime of
a sensor with a 500 mA battery and a 15 minute
transmission interval. Out of curiosity, both TinyOS and
Contiki were programmed to transmit UDP packets of
different sizes at consecutive time intervals to observe
any differences in energy usage between the two. The
results indicated that the energy usage of 50 to 80-byte
payloads in Contiki and Tiny OS were significantly
different. The observed difference between operating
systems is most likely related to the method of header
compression. Specifically, Contiki uses HC1, while
TinyOS is based on HC01. However, both methods are
a part of the 6LoWPAN standard. Additionally, TinyOS
uses short addressing, while Contiki employs long
addressing. The type of addressing and header
compression used by the OS can be changed, but in
this particular test, default settings were used.
For payload sizes greater than 60/90 bytes, the IP
packet had to be divided into two separate 802.15.4
frames because the maximum frame size of IEEE
805.15.4 is 127 bytes. The separation of IP packets
increased energy usage and decreased the expected
lifetime of the sensor. Thus, software developers
should consider the maximum frame size if absolute
maximization of sensor lifetime targeted. However
increased payload sizes can of course be
compensated with a larger battery.
As shown in Fig. 8, the fixed transmission interval was
set to 15 minutes, and the effect of transmission
interval on the expected lifetime of the sensor was
analyzed. Additionally, sensor lifetime was evaluated at
various transmission frequencies and a fixed payload
of 80 bytes, as shown in Fig. 9. In accordance to theory
the results indicated that a low transmission frequency
has a positive effect on sensor lifetime. In the case of
context aware sensors, which only transmit data when
required e.g. when a measured temperature exceeds a
set threshold, sensor life expectancy will in most cases
be increased. However, the impact of the
sleep/standby energy usage will make up a larger
percentage of the total energy usage, which hence will
mean that the importance of keeping the sleep current
low will be even bigger.

Fig. 8. The effect of payload size on the expected lifetime
of a sensor platform at a transmission rate of
4 transmissions per hour (1 to 100 bytes).

September 5
th
to September 7
th

10

Fig. 9. The effect of transmission frequency on the
expected lifetime of a sensor platform at a payload
of 80 bytes.

The predictive life expectancy calculations did not take
into account the fact that batteries loose energy over
time, even if they are not in use. Depending on battery
type, this can significantly reduce the expected lifetime
of a sensor.
SENSOR INTEGRATION
To provide wireless accessibility to devices in the
district heating substation, some simple interface
electronics were developed to integrate Mulle with
device hardware. As shown in Fig. 10, a heat meter
was integrated with a Mulle in the bottom module
location.
When digital communication interfaces were available
(heat meter and circulation pump), the corresponding
application protocols were kindly provided by the
vendors (Kamstrup and Grundfos). The control valve
(Siemens SQS-65) was not equipped with any digital
communication interface; however, an analog 010 V
input used to control the position of the valve and a
010 V output used to read the position of the valve
were available.

Fig. 10. A Mulle sensor platform integrated with a
Kamstrup Multical 601 heat meter.
Mulle is marked by a blue square, and the interface
card is indicated by a purple square.
All of the substation devices used in this study were
module-based, which allows manufacturers to produce
6LoWPAN module for large scale deployment.
RESULTS
Wireless devices in a district heating substation were
successful integrated to support a IPv6/6LoWPAN
network. Due to the range limitations of 2.4 GHz
modules, deployment of several platforms was
restricted. However, new 868 MHz platforms are now
available and show excellent preliminary results.
2.4 GHz platforms will be replaced with 868 MHz
platforms during the spring/summer of 2010.
A lifetime of 10+ years can be achieved with 500 mAh
battery and an average transmission interval of
15 minutes using IPv6 compatible communication;
thus, the life expectancy of battery powered sensors
did not have a negative effect on integration.
CONCLUSION
Integrating an IPv6/6LoWPAN wireless network in a
district heating substation can significantly increase the
functionality and scalability of the substation and supply
new services to both producers and consumers.
Using an open, well documented, and tested protocol
increases the possibility of interoperability between
products of different manufacturers. This study
revealed that available technology can be used to
achieve IP-based wireless communication. However, a
considerable amount of work on smart application
layers must be conducted before wireless sensor
networks in district heating substations can be
deployed and used to its full potential.
FUTURE WORK
To achieve complete device compatibility, the
application layer(s) of the integrated network must
further developed. One interesting approach is to adapt
the service oriented architecture in web-based services
to low-power sensors. Available service oriented
architectures (SOA) such as DPWS
1
are developed
primarily for large enterprises and are not intended to
be used with a resource limited device that possesses
a low-bandwidth link. However, the functionality of this
architecture would support a convenient solution for
direct sensor integration in enterprise systems.
The integration of sensors and SOA such as DPWS is
a challenging but intriguing task.

1
Device Profile for Web Services
September 5
th
to September 7
th

11

REFERENCES
[1] Z. Shelby and C. Bormann, 6LoWPAN: The
Wireless Embedded Internet, November 2009.
[2] J. Gustafsson, J. Delsing, and J. van Deventer,
Improved district heating substation efficiency with
a new control strategy, Applied Energy, vol. 87,
no. 6, pp. 19962004, 2010. [Online]. Available:
http://www.sciencedirect.com/science/article/-
B6V1T-4Y648K9-1/2/-
14e2e71a60c1335c8def21f6328bb9a0
[3] P. Lauenburg, Improved supply of district heat to
hydronic space heating systems, Ph.D.
dissertation, Dept. och Energy Sciences, Lund
University, P.O Box 118, SE-22100, Lund,
December 2009.
[4] K. Yliniemi, Fault detection in district heating
substations. Licentiate thesis, Div. of EISLAB, Dep.
of Computer Science and Electrical Engineering,
Lule University of Technology, 971 87 Lule,
Sweden: Lule University of Technology, 2005.
[5] K. Yliniemi, Individuell mtning av varmvatten-
frbrukning,
http://www.svenskfjarrvarme.se/download/4774/Ki
mmo Yliniemi.pdf, 2007.
[6] F. Wernstedt, P. Davidsson, and C. Johansson,
Demand side management in district heating
systems, in AAMAS 07: Proceedings of the 6th
international joint conference on Autonomous
agents and multiagent systems. New York, NY,
USA: ACM, 2007, pp. 17.
[7] J. Kurose and K. Ross, Computer Networking a
Top-Down Approach featuring the Internet, 2nd ed.
Pearson Education International, 2003.
[8] G. Mulligan and 6lowPAN Working Group, The
6lowpan architecture, in Proceedings of the 4th
workshop on Embedded networked sensors, June
2007.
[9] IEEE 802.15.4-2006 standard,
http://standards.ieee.org/getieee802/802.15.html,
April 2010.
[10] Embedded internet system technology botnia AB,
http://www.eistec.se/, March 2010.
[11] Crossbow technology, http://www.xbow.com,
March 2010.
[12] AVR raven, http://www.atmel.com, April 2010.
[13] Sensinode, http://www.sensinode.com, April 2010.
[14] Contiki, http://www.sics.se/contiki/, March 2010.
[15] Tinyos, http://www.tinyos.net, March 2010.

September 5
th
to September 7
th

12

ON THE RADIAL CONTACT PRESSURE
OF PARALLEL BURIED PIPES FOR DISTRICT HEATING
I. Weidlich
1
, M. Achmus
2
1
AGFW, German Heat and Power Association, Research & Development,
Stresemannallee 28, 60596 Frankfurt am Main, i.weidlich@agfw.de
2
Institute of Soil Mechanics, Foundation Engineering and Waterpower Engineering,
Leibniz University of Hannover, Appelstr. 9A, 30167 Hannover, achmus@igbe.uni-hannover.de

ABSTRACT
For the design and calculation of buried district heating
pipe systems the magnitude of radial contact pressures
acting on the pipes is of importance, since these
pressures affect the friction forces which may be
mobilized. For parallel buried pipes, the stress
distribution is generally expected to be different from
the case of a single pipe. The present investigation
compares radial stresses according to current design
directives for buried single pipes with numerically
calculated stresses for parallel buried pipes. The
calculations show a deviation of the radial stress
distributions in particular for the springline area. The
results are compared with former theoretical investi-
gations, which predicted a reduction of radial contact
pressures between the two pipes. This is verified for
small-diameter pipes. With larger pipe diameters a
stress increase was identified between the pipes.
However, with regard to the average radial pressure
only slight differences between single pipes and
parallel buried pipes were found.
INTRODUCTION
As a part of the underground infrastructure of modern
settlements, district heating pipe networks are an
important medium of economic heat transportation. Hot
water is pumped in a flow pipe from the supply station
to the consumer at a high temperature and under high
pressure, and the used water is pumped back to the
supply station in a return pipe.
For buried district heating pipes the earth pressure on
the pipe, respectively the radial contact pressure, is an
important value for the design, since it affects the
friction forces which may be mobilized. The friction
forces determine the axial deflections of the pipe and
the distribution of normal stresses, which are induced
by the temperature loading of the pipe. According to
the European Standard EN 13941, the normal stress
on the pipe coating is calculated for single pipe trench
conditions dependent on the overburden weight of the
soil, the diameter, the pipe weight and an earth
pressure coefficient [1]. However, in practice flow and
return district heating supply pipes are buried side by
side in the same trench. Fig. 1 shows a typical situation
for buried district heating pipes according to Floss [2].

Fig. 1. Typical trench condition for DH-pipes after
FLOSS [2]
The distance between the two pipes depends on the
requirements of the laying technique and procedure.
For small distances between the two pipes an
interaction between the two pipes is to be expected.
PREVIOUS WORK
Previous theoretical investigations were based on the
calculation method developed by Leonhardt, taking into
consideration the deformation behaviour of pipe and
soil and their influence on each other [3]. Leonhardt
introduced the shear resistant beam on elastic
bedding theory, in which the backfill above the pipe is
considered to be a shear resistant beam, which is able
to transfer shear loads, but no bending moments.
Using this model it is possible to determine the shear
forces activated by the deformation of the shear
resistant beam caused by different stiffnesses of the
pipe and the surrounding soil, which leads to a
redistribution of stresses in the soil with corresponding
concentration factors .
For practical application in Germany regulation
ATV A 127 was published employing Leonhardts
theory for buried pipes [4]. This regulation can be
applied analogously to all kinds of buried pipes. The
special application of regulation ATV A 127 for buried
preinsulated district heating pipes was first investigated
by Beilke [5].
September 5
th
to September 7
th

13

For the case of parallel buried pipes former analytical
calculations by Rizkallah and Achmus using
Leonhardts theory showed a reduction of the vertical
stresses between the two pipes [6]. The system used
for these calculations with the shear resistant beam on
elastic bedding for two parallel buried pipes is shown
in Fig. 2.

Fig. 2. Shear resistant beam for parallel buried pipes

The concentration factors
B
. B for the area beside and
between the pipes were found with the following
assumptions:
- The influenced area for the determination of
the concentration factor
B
beside the pipe
(
B1
) is defined by a line with an inclination of
60 shown in Fig. 3. This angle coincides with
the theoretical slope inclination of a non-
cohesive soil with an internal angle of friction
of m' = 30.
- Between the pipes (
B2
) the full interspace is
taken as the area of influence.

Fig. 3. Method to determine the concentration factors for
parallel buried pipes
The calculations by Rizkallah and Achmus showed only
small deviations for the concentration factor
B
of single
pipes and the concentration factor
B1
for parallel
buried pipes. However, a significant reduction of the
vertical stresses (i.e.
B2
<
B
) was determined between
the pipes. As an example, the ratio of the stress factors
is shown in Fig. 4, dependent on the relative
overburden height H/D and the relative distance of the
pipes A/D.

Fig. 4. Ratio B2/B dependent on relative overburden
height and pipe distances

A calculation method for parallel buried pipes in
stepped trenches was proposed by Hornung and Kittel
[7]. With this calculation method the total loading on
one pipe is derived from the sum of the partial loadings,
which correspond to the trench shape to the right and
left of the pipe. The typical trench condition for district
heating pipes provides a non stepped trench with the
flow and return pipes installed on the same
underground level. The presented study was therefore
carried out without employing the Hornung and Kittel
calculation method.
NUMERICAL INVESTIGATIONS
Numerical calculations were carried out with the two
dimensional finite element program PLAXIS, version
8.6. Two standard situations with different outer pipe
diameters D (DN65, D=140 mm; DN250, D= 400 mm)
of two parallel buried district heating pipes were
investigated. The distance between the pipes was
chosen to be A=10 cm (see Fig. 1). The overburden
height of the backfill material of the trench was
H/D=3.0. The finite element mesh used for the DN65
pipe is shown in Fig. 5 as an example.

Fig. 5. Finite element mesh for the case DN65, H/D=3
The installation process was simulated by a staged
construction process, considering a retained trench
and the backfilling procedure with several layers. The
compaction process was accounted for by applying a
static distributed load of p=10 kN/m on each of the
layers. Ground water was not considered in this
investigation.
September 5
th
to September 7
th

14

Sand in a medium dense to dense state was assumed
as backfill material. The mechanical behaviour of the
soil was modelled with the Mohr-Coulomb constitutive
law, which is a linear elastic / ideal plastic material
model. The parameters used for the model are shown
in Table I.
Table I. Soil parameters used for sand in Mohr-Coulomb
material model
Definition and Unit Size
Unit weight [kN/m] 18
Oedometric Elasticity Modulus Eoed [MPa] 70.5
Poissons ratio v 0.3
Internal angle of friction m [] 40
Angle of dilatancy v [] 10
Interface friction Rinter [1] 0.536
Between pipe and soil, the Coulomb friction law with a
according to Eq. (1).
' tan * tan
inter
m o R
i
=
(1)
In order to keep the model as simple as possible the
pipes were assumed to be rigid.
In the numerical calculations the initial soil stress state
due to the soil unit weight was established first. The
installation procedure was then simulated and the
results were evaluated.
In the first model of pipes with an outer diameter of
D=140mm (DN65), no significant stress concentration
between the pipes was observed. The radial contact
pressure obtained for both pipes is shown in Fig. 6.

Fig. 6. Contact pressure on the DN65 pipes, H/D=3

However, in the second numerical model of pipes with
an outer diameter of D=400 mm (DN250), a stress
concentration between the pipes was evident. The
distribution of horizontal effective stresses acting after
the installation process is shown in Fig. 7. The stresses
are significantly larger between the pipes than beneath
them.

Fig. 7. Horizontal effective stresses oh (DN250 pipe,
H/D=3)
In Fig. 8 the stress concentration is shown by the
distribution of the radial contact pressure for the left-
hand pipe. In the springline area a maximum value of
o
r
=21.44 kN/m for the radial pressure was obtained.
Compared with the calculated average radial pressure
of
o
r,avg,calc
= 18.81 kN/m the deviation is about 12.2%.

Fig. 8. Contact pressure on the left-hand DN250 pipe,
H/D=3
From the DIN EN 13941 regulation the average radial
pressure on a single buried pipe can be derived for the
investigated trench condition according to Eq. (2).
|
.
|
\
| +
|
.
|
\
|
+ =
2
1
*
2
*
13941 , ,
k D
H
avg r
o (2)
In Table II the results of the numerical investigation are
compared to the expected radial pressure from the DIN
EN 13941 regulation.
Table II. Average contact pressure or,avg for H/D=3.0
DN Single pipe according to
DIN EN 13941
Parallel buried pipe
according to numerical
results
65 6.15 kN/m 7.25 kN/m
250 17.58 kN/m 18.81 kN/m
Regarding the average radial contact pressure the
difference between the expected values from the DIN
September 5
th
to September 7
th

15

EN 13941 regulation for single pipes and the calculated
values from the numerical simulations for parallel
buried pipes is rather small. The values for the parallel
buried pipes are slightly larger.
CONCLUSION
The earth pressure on district heating pipes is an
important design value and should be determined as
exactly as possible. In the presented work the earth
pressure on parallel buried pipes was investigated.
The evaluation of the radial stresses on the pipe in
numerical calculations showed a stress concentration
between two pipes buried in the same trench for short
pipe distances and large diameters. However, former
theoretical investigations led to a reduction of radial
contact pressure between the two pipes, which was
observed in the numerical simulations for small
diameters and small overburden heights.
Because typical trench conditions with two parallel
buried pipes are not considered in current design
directives for district heating pipes the numerical results
were compared with the values derived from the
current design regulations. For the observed systems
only small deviations regarding the average normal
pressure between single pipe and parallel buried pipes
were found. Thus, as long as the exact distribution of
stresses along the pipe perimeter is not of particular
relevance, current calculation directives are also
suitable for parallel buried pipes. Only for conditions
with large pipe diameters and small distances between
the pipes and also relatively large overburden heights
is a significant deviation to be expected.
Furthermore, inhomogeneous backfill compaction,
which is probable for small pipe distances under in situ
conditions, affects the contact pressure. In order to
take into account the real compaction process within
the trench, only direct measurements seem to lead to
correct results. Further research work is necessary at
this point.
In order to avoid large deviations in the contact
pressures, good and consistent backfill compaction and
a certain minimum distance between the flows and
return pipes is recommended.
REFERENCES
[1] DIN EN 13941, Berechnung und Verlegung von
werkmig gedmmten Verbundmantelrohren fr
Fernwrme, Deutsches Institut fr Normung e.V.,
Beuth Verlag Berlin, 2003.
[2] R Floss, Handbuch ZTVE-StB 94/1997,
Kommentar mit Kompendium Erd- und Felsbau, 3.
Auflage, Kirschbaum-Verlag, 2006.
[3] ]G. Leonhardt, Belastung von starren
Rohrleitungen unter Dmmen, PhD Thesis,
Institute of Soil Mechanics, Foundation
Engineering and Waterpower Engineering,
University of Hannover, 1973.
[4] ATV A 127, Richtlinie fr die statische Berechnung
von Entwsserungskanlen und -leitungen,
Arbeitsblatt A 127 der Abwassertechnischen
Vereinigung e.V., 2000.
[5] O. Beilke, Interaktionsverhalten des Bauwerks
FernwrmeleitungBaugrund, Institute of Soil
Mechanics, Foundation Engineering and
Waterpower Engineering, University of Hannover,
1993.
[6] V. Rizkallah, M. Achmus, Zur Gre der
Reibungskrfte an erdverlegten Fernwrmelei-
tungen, Forschungsvorhaben Wechselwirkungen
Fernwrmeleitung Bettungsmaterial, Institute of
Soil Mechanics, Foundation Engineering and
Waterpower Engineering, University of Hannover,
1993.
[7] K. Hornung, D. Kittel, Statik erdberdeckter
Rohre, Bauverlag GmbH Wiesbaden und Berlin,
ISBN 3-7625-2039-9, 1983.

September 5
th
to September 7
th

16

ANALYSIS ON FLAT STATION CONCEPT.
PREPARING DHW DECENTRALISED IN FLATS
Thorsen, Jan Eric

Senior Project Manager, M.Sc., Danfoss District Energy, DK-6440 Nordborg
jet@danfoss.com

ABSTRACT
In some countries the flat station concept is becoming a
common way of realising heating and domestic hot
water (dhw) installation in blocks of flats. Anyhow, in
other countries it is at the very beginning. Experience
from those countries reveal a number of questions when
understanding and evaluating the flat station concept.
A number of parameters can be addressed to and be
evaluated to disclose qualities and performance of the
flat station concept in relation to traditional concepts for
heating and dhw installations.
This paper aims at analysing main parameters
regarding quality (comfort) and performance of the flat
station concept, covering block distribution system, flat
station itself and flat installation. Parameters in focus
are: riser system, instantaneous dhw principles, heat
losses, comfort of dhw, investments and energy
savings, metering and hygienic issues for dhw.
INTRODUCTION
Areas of district heating distribution systems, building
heating installations and domestic hot water (dhw)
installations show a high degree of conservatism and
traditions, which are reasonable due to their lifetime. But
this also implies a number of questions when new
concepts like the flat station concept are to be
introduced. Not only questions addressed to the flat
station concept but also to existing systems, where
detailed knowledge is faded out due to the maturity of
concepts. This paper aims at analysing main
parameters regarding quality and performance for the
flat station concept, covering block distribution system,
flat station itself and flat installation.
THE PARAMETERS ADRESSED
Investments:
Distribution system
Basement sub station versus flat stations
Energy meters
Energy Savings:
eat loss in primary distribution system
dhw circulation pump consumption
Comfort:
dhw temperature stability and variation
dhw recovery time after idle period
Hygienic issues
considerations on Legionella related to the systems
physical layout.
Investments
Reference for comparing the flat system concept with a
conventional concept is based on modern way of
making block pipe distribution systems [1]. In both
cases it is a horizontal pipe layout in flats with a vertical
pipe tunnel for distribution. Pipe distribution systems are
shown in fig. 1. Main differences are to be seen in the
number of pipes installed. Since dhw is prepared
decentralized in flats, dhw pipe and dhw circulation pipe
are eliminated. Centrally located dhw station in the
basement is replaced by decentralized flat stations.
Balancing valves for heating as well as for dhw
distribution is saved for the flat station concept.
Regarding metering then the dhw meter is eliminated,
since the primary supply to the flat station covers flat
heating and dhw as well. According to measurements of
more than 2500 dwellings in Denmark, including
detached houses as well as multi storey buildings,
individual metering, say individual billing, resulted in
savings of 1530%, [2]. Therefore, this analysis
assumes metering of all thermal energy deliveries to
flats.

Fig 1. Pipe distribution systems in blocks of flats. C:
Modern reference principle. F: Flat station principle.
A recent investment example comparing flat station
concept (F) with traditional system (C) is included in
fig. 2. Data are based on a Danish case from rhus
area where a block, built in 4 levels and a basement
level consisting of 24 flats, will generally be modernised.
Investments compared are based on concepts
presented in fig. 1. Main conclusion is that investment
level is approx. break-even for the two systems, for this
September 5
th
to September 7
th

17

typical Danish case. For other countries implying other
components/costs levels, level could change. In
general, the experience is that flat stations are on
break-even cost level or slightly higher. This is valid for
new buildings as well as for renovation projects.
-
2
0
0
0
0
-
1
5
0
0
0
-
1
0
0
0
0
-
5
0
0
0 0
5
0
0
0
1
0
0
0
0
Euro
1
I
t
e
m
Invest ment s comparison Flat st at ion concept
rhus Case, block of 24 flat s
Saved eneregy met er for
dhw
Saved balancing vavles
for dhw circulat ion
Saved balancing valves
for heat ing dist ribut ion
Saved dhw pipes, incl.
circulat ion
Saved dhw preparat ion
cent rally locat ed
Invest ed in Flat St at ion

Fig. 2. Investment balance for traditional system C and flat
system F. Block of 24 flats.
Energy savings
Main contribution to energy saving is originated from
installed hot distribution pipes. To begin with, it is
assumed that half the yearly distribution energy loss is
net loss (summer time), meaning not contributing to
heating up the building. Wintertime temperatures are
assumed to be identical for the two concepts, because
for this period the heating system defines temperature
levels. To quantify losses a room temperature of 20 C
is assumed. Danish Technical Insulation Standard [3]
requires minimum allowable heat loss constants (W/m),
depending on temperatures, annual operation time and
pipe diameter. These constants turn out to be quite
similar to all pipes in question. To simplify preconditions
a heat loss coefficient of 0.20 W/mK has been chosen
for all hot pipes. Table 1 shows a comparison of pipe
temperatures, heat loss and electrical dhw circulation
pump.
Flats in this first case are provided with floor heating in
bathrooms; therefore, heating is active all year. Due to
floor heating, temperatures for the traditional concept
are lower during summer season compared to the flat
station concept, since floor heating typically operates at
lower temperatures. For the flat station concept a dhw
temperature at 45 C is assumed, demanding a primary
temperature of 55 C.
Comparing the two systems regarding heat loss, then
favour is towards the flat station concept. For the rhus
case it means approximately 4200 kWh/year savings
corresponding to 210 Euro/year (ex. pump. costs). This
means a saving of approx. 2 kWh/m2/year. This
represents a saving of approx. 2% of the yearly heat
demand for a 1970 Danish block building (not including
energy for dhw).
Table 1. Energy losses for traditional system C, and flat
system F based on the rhus case.

Secondly, a situation is analysed where heat loss is not
utilised in the building distribution system at all. Winter
energy losses for the flat station is assumed to be
usable and no floor heating is active during summer.
Table 2. Energy losses for traditional system C, and flat
system F based on the rhus case.

Concept Pipe T T E E Net E Energ. Energ.
length pipe amb. loss loss/y loss/y price costs
[m] [W/mK] [C] [C] [W] [kWh] [kWh] [/kWh] [/year]
Trad. con. C Sum. flow 120 0.20 20 20 0 0 0 0.05 0
Trad. con. C Sum. return 120 0.20 20 20 0 0 0 0.05 0
Flat st. con. F Sum. flow 120 0.20 55 20 840 7358 3679 0.05 184
Flat st. con. F Sum. return 120 0.20 30 20 240 2102 1051 0.05 53
Trad. con. C Winter flow 120 0.20 70 20 1200 10512 5256 0.05 263
Trad. con. C Winter return 120 0.20 30 20 240 2102 1051 0.05 53
Flat st. con. F Winter flow 120 0.20 70 20 1200 10512 5256 0.05 263
Flat st. con. F Winter return 120 0.20 30 20 240 2102 1051 0.05 53
Trad. con. C Unit heat loss 1 pcs. 300 W/unit 2628 2628 0.05 131
Flat st. con. F Unit heat loss24 pcs. 25 W/unit 3816 1908 0.05 95
dhw circ. C Sum. + win. 240 0.20 53 20 1584 13876 13876 0.05 694
dhw circ. elec. Sum. + win. - - - - 30 260 - 0.25 65
Trad. c. C Total 22811 1141
Flat st. c. F Total 12946 647
Diff. C-F Total (ex. electrical consumption) 9865 493

Comparing the two systems regarding heat loss, then
favour is again towards the flat station concept. For the
rhus case it means approximately 9900 kWh/year
savings corresponding to 490 Euro/year (ex. pump.
costs). This means a saving of approx. 4 kWh/m2/year.
This represents a saving of approx. 4% of the yearly
heat demand for a 1970 Danish block building.
Additionally, as for the flat station concept there is no
need for dhw circulation pump, thus no need for the
electric energy of 260 kwh/year. A part of this saving is
anyhow spent for the flat station concept due to
additional circulation of primary water. It is assumed that
this is approx. half the electric energy for dhw circulation
pump of 130 kwh/year.
Concept Pipe T T E E Net E Energ. Energ.
length pipe amb. loss loss/y loss/y price costs
[m] [W/mK] [C] [C] [W] [kWh] [kWh] [ /kWh] [ /year]
Trad. con. C Sum. flow 120 0.20 40 20 480 4205 2102 0.05 105
Trad. con. C Sum. return 120 0.20 25 20 120 1051 526 0.05 26
Flat st. con.F Sum. flow 120 0.20 55 20 840 7358 3679 0.05 184
Flat st. con.F Sum. return 120 0.20 30 20 240 2102 1051 0.05 53
Trad. con. C Unit heat loss 1 pcs. 300 W/unit 2628 1314 0.05 66
Flat st. con.F Unit heat loss 24 pcs. 25 W/unit 3816 1908 0.05 95
dhw circ. C Summer 240 0.20 53 20 1584 13876 6938 0.05 347
dhw circ. elec. Sum. + win. - - - - 30 260 - 0.25 65
Trad. c. C Total 10880 544
Flat st. c. F Total 6638 332
Diff. C-F Total (ex. electrical consumption) 4242 212
September 5
th
to September 7
th

18

When looking at annual energy consumption savings in
percent, figures might appear rather low and of minor
impact. In this respect it has to be remembered that
energy saving relates to a typical 1970 building.
Present building codes require energy savings in the
order of 50% reduction for 2010 established buildings
and another 50% for 2015 established buildings. This
means savings in relative numbers for the flat station
concept will triple towards 2015 compared to 1970
building standards. Range of relative savings goes from
2-4% to 8-16% towards 2015.
Comfort
Comparing the two ways of preparing dhw, i.e. by
storage tank and by heat exchanger [4]/[5], it is obvious
that dynamics of control tasks is quite different. At
continuous tapping from full charged storage tank
temperature will be constant and also independent on
tapping flow changes until colder layers (cold water)
have refilled the storage tank. At this point comfort
drops drastically. If tappings are made periodically and
in shorter duration then temperature will be constant
within each tapping, but will vary between tappings due
to mixing of temperature layers. A typical question
regarding instantaneous prepared dhw is how stable are
temperatures when applying dynamics. Regarding
dynamic control performance an example is included in
fig. 3:

Fig. 3. Dynamic control performance (step test) for
thermostatic and pressure controlled heat exchanger for
dhw production [6]
Fig. 3 shows that stability, temperature peaks at load
change and total dhw temperature (T22) variation is
limited to 34 C. Regarding oscillations at low tapping
flow it should be noted that T22 is measured at heat
exchanger outlet. As example a 5 m 22 mm pex pipe
reduces peaks and amplitudes additionally, dependent
on frequencies, but typically 50%. This example is for
very high primary supply conditions. Oscillations appear
at tap flow of 100 l/h or below. This level shall be seen
in relation to the fact that a typical tapping flow for one
tap is 200400 l/h.
Another relevant question is how fast dhw temperature
is on desired level if supply is in idle condition. Here
dynamics are heavily influenced by idle bypass
thermostat setting. Also pump dynamics are influencing,
meaning how fast is the primary circulation pump
reacting on rapid changes of hydraulic conditions, say
opening of primary valve.

Fig. 4. Dynamic control performance (idle recovery) for
thermostatic and pressure controlled heat exchanger for
dhw production. Heat exchanger is cold during idle. [6]

Fig. 4 shows a flat system with cold heat exchanger.
Bypass temperature setting corresponds to primary
supply temperature (Tf.dh) of 40 C and primary return
temperature (Tr.dh) of 30 C. This setting is in the very
low end, but in the high end regarding energy saving.
Available differential pressure is 1 bar, but drops to 0.25
bar at the beginning of the tapping. In this case
temperature in circulation (Tsupply) is approx. 67 C.
Primary branch pipe from supply to the flat station is
4m, 20 mm.
Measurements show that primary supply has a delay of
approx. 7 sec. to reach a level of 55 C. Additional delay
is then caused by heating up the heat exchanger and
dhw water, this delay is additional approx. 3 sec. to
reach a minimum demanded level of 45 C. After
5 meter of pex pipe of 22 mm additional delay is
approx. 7 sec. By this the total delay from tapping the
start to reach 45 at the tap is approx. 17 sec. In this
example a very long idle branch pipe length is used,
more realistic would be 02 m, resulting in a primary
side delay of not more than a few seconds. Also
diameter of secondary dhw pipe is rather big and
represents a typical shared pipe dimension,
representing one pipe for several taps.
Anyhow, this delay is only relevant for the first tapping,
since thermal capacities combined with efficient
insulation is maintaining temperature, typically with time
constants of 12 hr.
September 5
th
to September 7
th

19

Comfort level is increased by applying a higher bypass
thermostat setting and/or a hot heat exchanger during
idle. Fig. 5 shows an example of flat station with hot
heat exchanger and thermostatic controlled heat
exchanger [7]. Idle temperature is approx. 50 C
corresponding to dhw tapping temperature.

Fig. 5. Dynamic control performance (idle recovery) for
thermostatic controlled heat exchanger for dhw production.
Heat exchanger is warm during idle. [7]
Fig. 5 shows a flat system with hot heat exchanger at
idle. Bypass temperature setting corresponds to a
primary supply temperature (T11) of 58 C and primary
return temperature (T12) of 44 C. This setting is the
high end, meaning in high end regarding comfort. For
this system there are no primary delays, and dhw
tapping temperature at the flat station is available after
approx. 2 sec. Additional delay due to dhw piping
towards tap would be similar to previous example.
In many practical matters a compromise between the
two examples regarding idle temperature setting fulfils
demands for good comfort with reasonable energy
consumption.
In the following a general trade off is included between
branch pipe length, dhw pipe length, idle condition for
heat exchanger and temperature delay on dhw, based
on dynamic simulations. Pipes are simplified by simple
delay models with no heat loss. Heat exchanger is
based on a lumped capacity model described in [5].

Fig. 6. Basic application for flat station, including boundary
conditions for dynamic simulations.
For simulations a branch pipe flow (Q1) of 800 l/h is
assumed. This represents a situation where the
thermostat is fully open until the desired set temperature
is reached. Further a step vice flow change from zero to
Q1 or zero to Q2 is assumed. Tapping flow is assumed
to be on a high level flow for one tap, which is typically
applied when opening the dhw. Q2=400 l/h for all
simulations.
Del ay unt i l reachi ng 45C
L2=5m & 10m (i nt ernal 10mm) - Heat Exchanger hot & col d at i dl e
0
2
4
6
8
10
12
14
16
0 1 2 3 4 5 6
L1 [m] (i nt ernal 20mm)
d
T

[
s
e
c
]
hot - dt at T2 - L2=5m hot - dt at T3 - L2=5m
hot - dt at T4 - L2=5m hot - dt at T4 - L2=10m
col d - dt at T3 - L2=5m col d - dt at T4 - L2=5m
col d - dt at T4 - L2=10m

Fig. 7. Dynamic simulation for hot and cold heat exchanger
during idle. Delay (dt) for dhw temp. of 45 C.

Heat exchanger simulated is Danfoss XB06H-40 [6]. It
can be seen from figure 7, that influence on hot or cold
heat exchanger is in the range of 2 sec. delay. Branch
pipe length (L1) has minor impact on time delay. This is
due to the fact that temperature is maintained with a
temperature gradient along pipe during idle, reflecting
T1 to T2. Basically water in branch pipe is heated to a
certain level already before tapping. Anyhow, due to
energy loss and return temperature, idle bypass
temperature is lower than dhw tapping temperature in
this case.
Main influence on time delay is dhw pipe diameter and
length (L2). Connection in flats shall be of star
coupling principle where every tap has its own supply
pipe with a small inner diameter. Temperature in dhw
pipe water is assumed to be room temperature prior to
tapping. In general, additional delays of typically 3 to 6
seconds shall be expected due to thermal interaction
with thermal capacities along the way to tap and
hydraulic dynamics on branch pipe side and hydraulic
dynamics on dhw side.
Simulated waiting time for a dhw temperature of 40 C
is included in figure below:

September 5
th
to September 7
th

20

Del ay unt i l reachi ng 40C
L2=5m & 10m (i nt ernal 10mm) - Heat Exchanger hot & col d at i dl e
0
2
4
6
8
10
12
14
16
0 1 2 3 4 5 6
L1 [m] (i nt ernal 20mm)
d
T

[
s
e
c
]
hot - dt at T4 - l "=5m hot - dt at T4 - L2=10m
col d - dt at T4 - l 2=5m col d - dt at T3 - L2=5m
col d - dt at T4 - L2=10m

Fig. 7. Dynamic simulation results for hot and cold heat
exchanger during idle. Delay (dt) for dhw temp. of
40 C.

First of all it can be seen that time delay for reaching
40 C at tap is only a bit shorter than reaching 45 C.
This is due to the fact that the T4 temperature profile
has an almost step vice nature, i.e. if temperature goes
up after the dhw pipe is flushed through, it goes almost
like a step.
Different dhw controllers have different performance
regarding time delay. In case of pure proportional
control for dhw system, time delay is longer at part load.
This is because primary flow is proportional to
secondary flow, and the lower the flow the longer the
waiting time. Looking at the example for Q1=800l/h,
Q2=400 l/h, L1=4 m, L2=5 m then time delay (dt) at T4
is 6.9 sec to reach 45 C dhw temperature. In case of
proportional controller with parameters Q1=400 l/h,
Q2=400 l/h, L1=4m, L2=5 m then dt=11.0 sec to reach
45 C. This has considerable effect on time delay as L1
gets longer. In case of a thermostatically controlled dhw
system or a combination of a thermostatically and
proportionally controlled dhw system, time delay is
shorter because no matter how small tapping is, as long
as the desired set point temperature is not reached, the
primary valve will be fully or almost fully open resulting
in high primary flow. Regarding delay to reach a dhw
temperature of 40 C this is only related to dhw pipe
dimension since 40 C is the bypass temperature if heat
exchanger is hot during idle. In case the heat exchanger
is cold during idle, then this introduces an additional
time delay as described above. In all cases, time delay
is dependent on dhw flow, resulting in delay in the dhw
pipe.
Hygienic considerations
Legionella is a well-known bacterial risk for dhw
systems. Normally it is not the question whether
Legionella is present in the dhw system or not, but
rather Legionella bacteria concentration in the dhw.
Facts influencing on potential for Legionella
concentration growth are dhw temperature, exchange
rate of hot water in distribution pipes, and volume of
dhw water in the entire hot system. Also other factors
are influencing, e.g. systematic cleaning of shower
outlets, but this will be not addressed to here, since the
effect is similar for concepts compared.
Comparing volumes of dhw in pipes for concepts, the
flats station solution has significantly lower volume
compared to the traditional system. Furthermore dhw
pipes should be star connected, meaning one small
(diameter) pipe from the flat station to each individual
hot tap. This eliminates problematic dead end or low
flow areas.
Typically volume of heat exchanger is 0.25 to 0.50 litre.
Typical dhw pipe volume is 0.10 l/m, equal to 1.0 litre for
10 m pipe. In total this is a volume of 1.5 to 2 litre
pr. flat. The comparable centrally placed dhw system
with dhw distribution will have a volume of 57 litre pr.
flat. By installing a dhw storage tank this will increase
significantly. The German DVGW regulations states that
heating dhw up to 60 C, due to e.g. Legionella, is not
required if volumes of heat exchanger or volume of dhw
pipes is less than 3 litres [8]. Based on those physical
concept differences Legionella bacteria risk is reduced
for the flat station concept.
Future energy supply/demand perspective
One important challenge for DH is to convert to 4th
generation DH systems. Intention is to realise efficient
DH systems for urban areas where heat demands will
decrease due to modernisation and new building energy
saving codes. In this context one way to go is to reduce
temperatures in DH networks [9]/[10]. This allows for
cost effective geothermal sources as well as other
renewable low temperature sources. For dhw, normal
temperature level is 45 to 60, where higher
temperatures typically are based on considerations
towards Legionella. A way to reduce temperature levels
in DH networks is to set dhw temperature at 45 C. By
this a primary temperature at sub station of 50 to 55 C
will be sufficient. A precondition for this is to use heat
exchangers for dhw production, like the flat station
concept.
CONCLUSION
The two pipe flat station concepts, consisting of
decentralised instantaneous dhw production, open the
possibility of reducing general DH net work temperature,
which for the future will be even more relevant due to
decreasing building heat demand and increased
availability of renewable energy. For building owners,
the investigated case shows that the flat station concept
is on brake-even investment level compared to
September 5
th
to September 7
th

21

traditional systems. The flat station concept has a net
energy saving due to less installed hot pipes. Energy
savings are in the range of 2 to 4 kWh/m^2/y for the
investigated cases. Comfort level has been
investigated, revealing well acceptable dynamic control
performance. Dhw temperature recovery after an idle
period for the instantaneous preparation of dhw is,
however, a trade-off between comfort and energy
saving. Related to Legionella, then risk can be reduced
when installing flat stations as presented in this paper.
REFERENCES
[1] Kristjansson, H. Comparing Distributions Systems
in Blocks of Flats, SDDE 2009, Slovenia
[2] Gullev, L., Poulsen, M. The Installation of Meters
Leads to Permanent Changes in Consumer
Behaviour, the magazine News from DBDH,
#3/2006.
[3] DS 452, Code of practise for thermal insulation of
technical service and supply systems in buildings,
2. Revision, Dansk Standard, 1999
[4] Thorsen, J.E. Cost considerations on Storage Tank
versus Heat exchanger for htw preparation, The
10th International Symposium on District Heating
and Cooling 2006.
[5] Thorsen, J. E. Control Concepts for DH Compact
Stations Investigated by Simulations, The 9th
International Symposium on District Heating and
Cooling 2004.
[6] http://www.danfoss.com/Products/Categories/List/H
E/Temperature-Controllers/Temperature-
controllers/IHPT-and-XB-06/b1c8a73c-59f1-4fef-
8b52-f49c97b6019b.html
[7] http://www.danfoss.com/Products/Categories/Group
/HE/District-Heating-Substations/Substations-
Direct-Heating/Flat-Stations/8f81605b-bab9-4644-
961b-51a3f0503f05.html
[8] DVGW regulations, Germany, Arbeitsblatt W551,
April 2004
[9] Olsen, P.K., Lambertsen, H., Hummelshj, R.,
Bhm, B., Christiansen, C.H., Svendsen, S.,
Larsen, C.T., Worm, J. A new Low-Temperature
District Heating System for Low-Energy Buildings,
The 11th International Symposium on District
Heating and Cooling 2008.
[10] Paulsen, O., Jianhua, F., Furbo, S., Thorsen, J. E.
Consumer Unit for Low Energy District Heating Net
Works. The 11th International Symposium on
District Heating and Cooling 2008.

September 5
th
to September 7
th

22

IMPROVED TEMPERATURE PERFORMANCE OF RADIATOR HEATING SYSTEM
CONNECTED TO DISTRICT HEATING BY USING ADD-ON-FAN BLOWERS
Per-Olof Johansson
1
, Janusz Wollerstrand
2

1 & 2
Lund University, Department of Energy Sciences, Division for Efficient Energy Systems
Corresponding author: per-olof.johansson@energy.lth.se, Energy Sciences, Lund University, P.O. Box 118,
221 00 Lund, Sweden, Phone: + 46 46 222 40 43, Fax: + 46 46 222 47 17

ABSTRACT
District heating (DH), which is the most common heat
source in multifamily houses and commercial buildings
in Sweden, can be produced in several different type of
production units.
In order to gain thermal efficiency in a DH system it is
important that DH supply and return temperatures are
kept low. The temperature demand in the DH system is,
during the heating season, dependent on the
temperature level in the heating system of the DH
connected buildings. Many production units benefit from
a lowered DH return temperature, while others are more
affected by a reduced supply temperature. In a CHP-
station the heat to power ratio will increase when the
DH supply temperature is decreasing. In order to reduce
the temperature demand, low temperature heating
systems are of interest, as well as systems resulting in a
low DH return temperature.
To increase the heat output in an existing radiator
heating system, the radiators can be complemented
with small electric fans resulting in an increased share
of forced convection in the heating system. Field studies
have shown that the heat output, with constant supply
temperature and mass flow through the radiator, can
increase with more than 50%.
INTRODUCTION
For many years, return temperatures in DH networks
have been an important issue for DH research. A low
DH return temperature is in many cases favorable for
the DH production units. However, if also the supply
temperature could be kept at a low level the share of
electricity produced in a CHP station could increase.
This would lead the way towards an increased share of
electricity produced by non fossil fuels. In Sweden more
than 30 % of the DH is produced in CHP stations [4].
In many reports the gain from a reduced temperature
level in the DH network has been discussed and
quantified in economic terms, see e.g. [12], [13].
The DH supply temperature level in the DH network is,
during heating season, dependent of the temperature
demand in the DH-connected buildings heating system.
In modern buildings low temperature heating systems
are common, which may allow reduced DH temperature
level. In order to reduce the temperature demand in
existing buildings the idea of using small add-on-fan
blowers placed under the radiator to increase the heat
output due to an increased share of forced convection
came up.
Objective
The field study presented in this paper investigates the
possibility to reduce the space heating temperature
program and estimates the impact on the DH supply
and return temperature. Possible reduction of the DH-
flow rate is also calculated.
This paper is focusing on buildings indirectly connected
to the DH network through a substation with heat
exchangers (HEX).
DESCRIPTION OF ADD-ON-FAN BLOWER
The add-on-fan blower that is tested in this study
consists of several regular DC motor driven fans,
originally used for cooling, mounted under a radiator,
see Fig. 1. In the study, two different kinds of radiators
were tested, a panel and a column radiator.
m
s
T
ss
T
sr
Add-on-fan
Increased air flow
Space heating
radiator
O
u
t
e
r

w
a
l
l
Floor

Fig. 1 The add-on-fan blower mounted on a panel radiator.
The add-on-fan blowers in this study are provided by a
Swedish company: A-energi AB (the product is called
flktelement in Swedish). The company describes the
features of the add-on-fan blower as a possibility to
reduce the temperature program without replacing the
radiators, with the aim to reduce the electricity demand
for buildings supplied with heat from heat pumps [5].
September 5
th
to September 7
th

23

THEORY
In this section a theoretic analysis of the impact of
forced airflow on heat output from radiator with a length
of (L) 1 m and height of (h) 0.59 m is described. The
radiator is in this study approximated by a flat vertical
plate. The indoor temperature is assumed to be
constant at 21C and equal to T
inf
.
Heat output
The heat output from the radiator to the room arises
from convection and radiation. The heat transfer
process from heating water to the room through a
radiator is summarized in equation 1 [7], [8].
u A = = ) ( ) ( A k T T c m Q
sr ss p s
(1)
k is the heat transfer (convection) from the water to the
surrounding metal, conduction through the metal and
convection from the outer surface of the radiator to the
room according to equation 2.
rad conv metal
metal
metal water
k o o
o
o +
+ + =
1 1 1
(2)
The dominating parameters in this equation are the
convection and radiation between the radiator and the
room (
conv
and
rad
), while the other terms, in this case,
can be neglected. This results in a new equation for
energy output, see equation 3.
u o u o A + A = + =
rad rad conv conv conv rad
A A Q Q Q
(3)
The temperature is the logarithmic mean
temperature difference according to equation 4.
i sr
i ss
sr ss
T T
T T
T T
= A
ln
u
(4)

Radiation
According to Trschel [8] the heat output from radiation
can be estimated according to equation 5.
u
c c
o c
u o
A
+

~
~ A =
rad m
rad
radiator
rad
rad
rad
rad rad rad
A T
A
A
A Q
3
) 1 (
4
(5)
Where the temperature, T
m
, is the mean temperature of
the radiator surface and the surfaces in the rooms, see
equation 6. For a panel radiator the A
rad
/A
radiation
=1 [8].
2
2 / ) (
,
i sr sf
surface room radiator m
T T T
T T T
+ +
~ + =
(6)
Since the A
rad
and emissivity,
rad
, are constant for a
specific radiator, the relation can be simplified to
equation 7.
u A ~
3
m rad
T C Q
(7)
Convection
The convection that arises due to the temperature
difference between the radiator surface and the
surrounding air is a function of the Nusselt number (Nu),
see equation 8.
h Nu
conv
/ o =
(8)
The heat output due to convection is divided into three
sections, natural, mixed and forced convection.
For natural convection, the Nu number is dependent on
the Rayleigh number (Ra), which is a product of the
Prandtl number (Pr) and the Grashof number (Gr). For
air, Pr can be considered constant, Pr=0.71, wile
i
T T
h
g Gr
/ 1 / 1
inf
2
3
~ =
A =
|
v
u |
(9)
where g is the gravity force, is kinematic viscosity and
is the coefficient of expansion.
Several empirical relations describing Nu are available.
In this study a relation described by Churchill has been
used [9], see equation 10 and 11.
9 / 4 16 / 9
25 . 0
] Pr) / 492 . 0 ( 1 [
67 . 0
68 . 0
+

+ =
Ra
Nu

9
10 < Ra (10)
27 / 8 16 / 9
6 / 1
5 . 0
] Pr) / 492 . 0 ( 1 [
387 . 0
825 . 0
+

+ =
Ra
Nu

9
10 > Ra (11)
For forced convection the Nu number is calculated by
equations described by Holman [10], see equations 12
and 13.
1/3 0.5
Pr Re 0.664 = Nu
5
10 5 Re < (12)
) 871 Re 037 . 0 ( r
8 . 0 1/3
= P Nu

7 5
10 Re 10 5 < < (13)
and the Reynolds number, Re, is described as:
v
L u
= Re
(14)
The product of Gr/Re
2
describes the dominating type of
convection. If Gr/Re
2
>10, natural convection is
dominating, if Gr/Re
2
1, both natural and forced
convection is of importance and if Gr/Re
2
<<1, forced
convection is dominating. When a mix of forced and
natural convection occurs, the Nusselt number is
calculated according to equation 15 [11].

3 / 1
3 3
) ( natural forced Nu Nu Nu + =
(15)

September 5
th
to September 7
th

24

Impact of air speed on space heating temperature
Results from the theoretical analysis, using the
equations above, are shown in Fig. 2 to Fig. 4. The heat
output for a radiator designed for the temperature
program 60/45 C is illustrated as a function of the air
speed in Fig. 2. The supply temperature and the mass
flow through the radiator are kept constant. Two cases
have been derived, one with heat output only from
convection, and one with heat output from both radiation
and convection. With =0.9, the share of heat output
from radiation will be 65% at DOT.
0 2 4 6 8 10 12 14 16
0
50
100
150
air speed (m/s)
C

a
n
d

%
T
sf
(C)
T
sr,conv
(C)
% additional Q, conv
T
sr,rad&conv
(C)
% additional Q, rad&conv
0 2 4 6 8 10 12 14 16
0
10
20
30
air speed (m/s)
o

(
W
/
m
2
K
)

Fig. 2 Calculated heat output improvements at Tss=60 C,
Tsr0=45 C with increasing airspeed. ms is kept constant.
As seen, the additional heat output from the radiator is
increasing rapidly when the air flow is increased. With
radiation taken into account, the increase is somewhat
lower since the mean temperature, T
m
, is decreased,
see equation 7.
In Fig. 2 the heat output is increasing. In Fig. 3 and
Fig. 4 the supply temperature to the radiator is reduced
instead to keep the heat output constant. New T
ss
and
T
sr
can now be calculated under the assumption that the
total heat output and the mass flow (m
s
) through the
radiator are constant. The impact of the air flow is
described for three different heat loads (Q
rel
=100%,
50%, 25%) with standard 60/45 C temperature
program as a reference. See Fig. 3.
0 5 10 15 20 25 30
20
25
30
35
40
45
50
55
60
air velocity (m/s)
T
e
m
p
e
r
a
t
u
r
e

(
C
)
T
ss0
= 60C , Q
rel
= 100 %
T
sr0
= 45C , Q
rel
= 100 %
T
ss0
= 43.1C , Q
rel
= 50 %
T
sr0
= 35.6C , Q
rel
= 50 %
T
ss0
= 33.6C , Q
rel
= 25 %
T
sr0
= 29.9C , Q
rel
= 25 %

Fig. 3 Possible Tss and Tsr to for three heat load situations
at different air speeds. Qrad=65% at DOT.
New temperature programs have been derived for some
moderate air speeds, see Fig. 4. As seen the impact of
an increased air flow, expressed in C, is larger at high
relative heat load.
0 20 40 60 80 100
15
20
25
30
35
40
45
50
55
60
65
relative heatoutput (%)
T
e
m
p
e
r
a
t
u
r
e

(
C
)
NEW TEMPERATURE PROGRAM
T
sf0
U= 0.0m/s
T
sr0
U= 0.0m/s
U= 0.5m/s
U= 1.0m/s
U= 2.0m/s
U= 3.0m/s

Fig. 4 New space heating temperature programs at
different air speeds. Red lines: Tss, Blue lines: Tsr.
Qrad=65% at DOT.
In the calculations performed, the radiator is assumed to
have the same heat output from both sides of the
radiator. The air flow is assumed to be uniformly
distributed through the length and height of the panel
radiator. This is not the case in the real add-on-fan
blower applications, however, one can expect results
following the same pattern.
EXPERIMENTAL STUDY
To investigate the performance of the add-on-fan
blower, two radiators of different type were supplied with
such device during the heating season 2009/2010. The
power supply to the fans was scheduled to switch on
and off while the mass-flow (m
s
) through the radiator
was kept at a constant level.
Field study object
The radiators are situated in two offices at Lund
University. The original temperature program for the
September 5
th
to September 7
th

25

radiators in the building is running at 60/45C at DOT
(represented by narrow black lines in Fig. 8 trough
Fig. 13).
The radiator types tested were:
- Panel radiator, see Fig. 5
- Column radiator, see Fig. 6.

Fig. 5 Add-on-fan blower mounted on a panel radiator.

Fig. 6 Add-on-fan blower mounted on a column radiator.
For each radiator the fans have been run at two
different rotation speeds. The net electric power
consumption (P
fan
) has been measured. See Table 1 for
P
fan
and the design heat energy output at DOT. Note
that the electric power to the add-on-fan blower is
constant and not dependent on the relative heat output.
Table 1 Radiator and add-on-fan blower design.
Radiator
type
P
fan
(el)
Q @ DOT, 60/45C
(Heat)
Panel 2.7 W
1.9 W
360 W*
Column 3.0 W
2.2 W
430 W*
* Calculated for new radiators of the same dimensions
manufactured by Lenhovda radiator factory [3]
The radiators are located in traditional office
environment in a building built in 1960.
Data acquisition
Measured parameters in the test were: secondary
supply and return temperature (T
ss
and T
sr
), indoor
temperature (T
i
) and outdoor temperature (T
out
).
The impact on return temperature for a given supply
temperature was then calculated.
In Fig. 7, a screenshot from the logger software is
shown. The return temperature is decreased by 5C
when the fan is switched on. This results in an
increased heat output by more than 60 %.
fan
T
ss
(C)
T
sr
(C)
T
i
(C)
T
out
(C)
U
fan
(V)
T
sr,0
=39
T
sr,Fan
=34

Fig. 7 Log file from field study.
New reduced temperature program will be derived in
next section.
MODIFYING SPACE HEATING TEMPERATURE
PROGRAM
Method
When the add-on-fan blower is switched on, the T
pr
is
decreasing, causing an additional heat output since m
s

is kept constant. See Fig. 8.
0 0.2 0.4 0.6 0.8 1
20
25
30
35
40
45
50
55
60
Relative heatload
T
e
m
p
e
r
a
t
u
r
e

[
C
]
T
ss
T
sr,0
T
sr,Fan
Original cooling in radiator
Additional cooling with
add-on-fan-operation

Fig. 8 Increased cooling of secondary system.
September 5
th
to September 7
th

26

The relative heat output from the radiator with and
without fan operation is calculated from equation 1.
)
`
=
=
DOT Fan sr ss p s Fan rel
DOT sr ss p s rel
Q T T C m Q
Q T T C m Q
) (
) (
, ,
0 0 ,
(1)
Since the temperature drop in the radiator is increasing
with the fan in operation, the radiator now could be
considered oversized. Then, with the same type of
reasoning as in e.g. [2], the T
ss
program or m
s
needs to
be adjusted in order to avoid overheating of the building.
In this paper, the m
s
has been considered constant,
allowing us to compute the new relative space heating
load for a given T
ss
according to equation 2. Q
rel,0
is
computed using the original space heating temperature
program.
0 ,
0
,
,
) (
) (
rel
sr ss
Fan sr ss
Fan rel
Q
T T
T T
Q
=
(2)
Knowing Q
rel,Fan
, a new temperature curve, which will
result in correct heat output from the radiator with the
fan in operation, can be calculated. The curve appears
to the right in the diagram, see Fig. 9.
0 0.2 0.4 0.6 0.8 1 1.2 1.4
20
25
30
35
40
45
50
55
60
Q
rel
T
s

T
ss,0
T
sr,0
T
ss,Fan
T
sr,Fan

Fig. 9 Modified secondary temperature program.
New space heating temperature program - results
The procedure described above has been applied to all
collected data. Results are shown in Fig. 10 and Fig. 11
for the panel radiator, and Fig. 12 and Fig. 13 for the
column radiator.
As seen in the figures, the temperature program is now
significantly reduced for both the panel and the column
radiator. The new temperature program shows a similar
pattern for both types of radiators. For the panel
radiator, the measured return temperature is more
concentrated, especially at the higher fan speed. This
could be explained by a more favorable air flow pattern
due to the physics of the radiator.

0 0.2 0.4 0.6 0.8 1 1.2 1.4
20
25
30
35
40
45
50
55
60
Q
rel
T
s

[
C
]
Panel radiator P
fan
= 2.7 W
T
ss,0
T
sr,0
T
ss,Fan
T
sr,Fan

Fig. 10 Modified temperature program for panel radiator,
Pfan=2.7 W.

0 0.2 0.4 0.6 0.8 1 1.2 1.4
20
25
30
35
40
45
50
55
60
Q
rel
T
s

[
C
]
Panel radiator P
fan
= 1.9 W
T
ss,0
T
sr,0
T
ss,Fan
T
sr,Fan

Fig. 11 Modified temperature program for panel radiator,
Pfan=1.9 W.
0 0.2 0.4 0.6 0.8 1 1.2 1.4
20
25
30
35
40
45
50
55
60
Q
rel
T
s

[
C
]
Column radiator P
fan
= 3.0 W
T
ss,0
T
sr,0
T
ss,Fan
T
sr,Fan

Fig. 12 Modified temperature program for column radiator,
Pfan=3.0 W.

September 5
th
to September 7
th

27

0 0.2 0.4 0.6 0.8 1
20
25
30
35
40
45
50
55
60
Q
rel
T
s

[
C
]
Column radiator P
fan
= 2.2 W
T
ss,0
T
sr,0
T
ss,Fan
T
sr,Fan

Fig. 13 Modified temperature program for column radiator,
Pfan=2.2 W.
INFLUENCE ON DH NETWORK
Knowing the reduced temperature level on the
secondary side of the HEX, the impact on the DH
network can be estimated. The impact is calculated
based on two different strategies:
1. Primary supply temperature (T
ps
) is kept at the
same level as before
2. The primary flow (m
p
) through the HEX is kept
constant
By applying the first strategy, both T
pr
and the mass flow
in the DH network is reduced. The second strategy
results in a reduced T
ps
and T
pr
without changing the
flow rate in the DH network.
Results so far will now be applied to a DH substation
dimensioned as recommended by the Swedish district
heating association [1]. The calculations are made with
a parallel connected DH substation serving a building
with 20 apartments. The substation is providing the
building with heat and domestic hot water (DHW) and
DHW circulation. The assumed DHW usage is
125 l/apartment&day, space heating load at DOT is
3 kW/apartment. The heat loss from DHW circulation is
assumed to be 0.1 kW/apartment. For each space
heating load a flow-weighted mean value for T
ps
and T
pr

is calculated for a time period of 24 h, including heat
load from both DHW and DHW circulation. The
reference DH supply temperature, dependent on the
space heating load, is assumed as illustrated in Fig. 14.

1 0.8 0.6 0.4 0.2 0 0
60
70
80
90
100
110
T
p
s

[
C
]
rel heatload
DH primary supply temperature

Fig. 14 DH primary supply temperature.

Results
The first control strategy is in Fig. 15 Fig. 20 noted as
T
ps
unchanged, and the second strategy is noted as
m
p
unchanged.
In Fig. 15 and Fig. 16 the possible reduction of DH
supply temperature is shown.
1 0.8 0.6 0.4 0.2 0
0
5
10
15
rel heatload
T
p
s

s
a
v
i
n
g

[
C
]
Panel radiator T
ps
saving

P
fan
= 2.7 W m
p
unchanged
P
fan
= 1.9 W m
p
unchanged

Fig. 15 Resulting Tps reduction with panel radiator.
1 0.8 0.6 0.4 0.2 0
0
5
10
15
rel heatload
T
p
s

s
a
v
i
n
g

[
C
]
Column radiator T
ps
saving

P
fan
= 3.0 W m
p
unchanged
P
fan
= 2.2 W m
p
unchanged

Fig. 16 Resulting Tps reduction with column radiator.
In Fig. 17 and Fig. 18 the reduction of T
pr
is shown. As
seen the reduction of T
pr
is of the same magnitude
independently of which control strategy is used.
September 5
th
to September 7
th

28

However, by keeping the DH supply temperature
constant (strategy 1) the flow rate in the DH network is
affected, see Fig. 19 and Fig. 20.

1 0.8 0.6 0.4 0.2 0
0
5
10
15
rel heatload
T
p
r

s
a
v
i
n
g

[
C
]
Panel radiator T
pr
saving

P
fan
= 2.7 W T
ps
unchanged
P
fan
= 1.9 W T
ps
unchanged
P
fan
= 2.7 W m
p
unchanged
P
fan
= 1.9 W m
p
unchanged

Fig. 17 Resulting Tpr reduction with panel radiator.
1 0.8 0.6 0.4 0.2 0
0
5
10
15
rel heatload
T
p
r

s
a
v
i
n
g

[
C
]
Column radiator T
pr
saving

P
fan
= 3.0 W T
ps
unchanged
P
fan
= 2.2 W T
ps
unchanged
P
fan
= 3.0 W m
p
unchanged
P
fan
= 2.2 W m
p
unchanged

Fig. 18 Resulting Tpr reduction with column radiator.
1 0.8 0.6 0.4 0.2 0
0
5
10
15
20
rel heatload
m
p

s
a
v
i
n
g

[
%
]
Panel radiator m
p
saving (%)

P
fan
= 2.7 W T
ps
unchanged
P
fan
= 1.9 W T
ps
unchanged

Fig. 19 Resulting mp reduction with panel radiator.

1 0.8 0.6 0.4 0.2 0
0
5
10
15
20
rel heatload
m
p

s
a
v
i
n
g

[
%
]
Column radiator m
p
saving (%)

P
fan
= 3.0 W T
ps
unchanged
P
fan
= 2.2 W T
ps
unchanged

Fig. 20 Resulting mp reduction with column radiator.
Annual gain in T
ps
and T
pr
during heating season
In order to evaluate the annual impact on the primary
temperature level, the outdoor temperature has to be
considered. In this case measured values for the
outdoor temperature in Malm have been used, see
Fig. 21.
May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May
-10
-5
0
5
10
15
20
25
30
35
40
T
o
u
t

(
C
)

Fig. 21 Outdoor temperature in Mam
1
st
May 2006 30
th
April 2007.
When calculating the annual gain for a DH-network a
comparison of flow-compensated mean temperature
during the heating season has been made. For the
calculations we assume a maximum heat output (Q
DOT
)
at -15 C and the balance temperature, when no space
heating is needed, +17 C.
Table 2 Reduction in annual primary temperature level
during heating season

Tps -
unchanged
mp -
unchanged

Tpr
2
pr ps
T T A + A

C
o
l
u
m
n

r
a
d
i
a
t
o
r

Pfan= 2.2 W -2.2 C -2.4 C
Pfan= 3.0 W -5.7 C -6.6 C
P
a
n
e
l

r
a
d
i
a
t
o
r

Pfan= 1.9 W -0.8 C -0.9 C
Pfan= 2.7 W -2.5 C -2.7 C
September 5
th
to September 7
th

29

Note that the annual flow-compensated mean
temperature in Table 2 is based on results from the field
study where measured values for low relative heat load
are missing, which makes the values in the table
somewhat underestimated.
CONCLUSION AND DISCUSSION
By installing the add-on-fan blower application on
existing radiators the temperature level in the heating
system can be substantially reduced. This will also have
impacts on the DH network and DH production units.
The impact on the DH network can be applied based on
two principles:
1) DH supply temperature kept at the same level as
without the add-on-fan blowers. This will result in
reduced primary flow rate.
2) Reduced DH supply temperature while primary flow
rate is kept constant.
The first strategy could be applied immediately, since
the primary supply temperature is kept as the same
level as before. This means that not all heating systems
connected to the DH network need to be modified in
order to apply this method. The lowered secondary
temperature level results not only in reduced DH-return
temperature, but also in a reduction of the DH flow rate.
The reduced flow rate could be used to increase the
number of buildings connected to the DH network, or to
avoid bottlenecks in the DH network. The magnitude of
the reduction of the DH supply temperature is between
9 and 12 C at DOT and at the same time the flow rate
is decreased with more than 10 %. On annual basis the
possible reduction of temperature level in the DH
network is in the magnitude of several degrees Celsius.
In order to apply the second strategy the demand for a
high temperature level in the DH network needs to be
reduced for all the connected buildings. Otherwise the
DH flow rate will increase. Calculations based on the
results from the field study in this paper shows that the
DH supply temperature can be reduced with about
10 C at DOT without affecting the DH flow rate. At the
same time the DH return temperature will be reduced
with as much as 10 C at DOT.
The performance of the tested add-on-fan blowers
corresponds to the pattern of theoretical calculations.
However, the results are not comparable since the air
flow in the pilot project has not been measured.
The results presented here are an important part in the
evaluation of effects of improvements in consumer
heating systems on primary energy efficiency in DH
systems including production plants, especially CHP.
ACKNOWLEDGEMENT
This work is part of the Primary Energy Efficiency
project of Nordic Energy Research.
NOMECLATURE
Abbreviations
CHP Combined heat and power station
DH District heating
DHW Domestic hot water
DOT Design outdoor temperature
HEX Heat exchanger (DH substation)
Variables
Heat transfer
coefficient (W/m
2.
K)
Gr Grashof number (-)
coefficient of
expansion (K-1)
h Height (m)
Thickness (m) k Heat transfer
coefficient (W/m
2.
K)
Conductivity (W/m
.
K) L Length (m)
Emissisivity (-)
m mass flow (kg/s)
Kinematic viscosity
(m
2
/s)
Nu Nusselt number (-)
Stephan-Boltzman
constant
P Electric power (W or
kW)
Logarithmic mean
temperature
difference (K)
Pr Prandtl number (-)
A Area (m
2
) Q Heat output (W or kW)
c
p
Heat capacity (J/kgK) Ra Rayleigh number (-)
C Constant Re Reynolds number (-)
g Gravity force (N/s
2
) T Temperature (C or K)
Subscripts
0 Design condition
(without fan)
p Primary (side)
Fan Add-on-fan blower in
operation
r Return
i indoor rad radiation
m Mean rel Relative
out outdoor s Secondary (side) or
Supply
September 5
th
to September 7
th

30

REFERENCES
[1] Svensk Fjrrvrme, District heating substations
design and installation, Technical regulations
F:101, The Swedish District Heating Association,
2008
[2] P. Ljunggren, P-O. Johansson, J. Wollerstrand,
Optimized space heating system operation with the
aim of lowering the primary return temperature, in
Proc. of the 11th International Symposium on
District Heating and Cooling, 2008, Reykjavik
[3] http://www.lenhovdaradiatorfabrik.se/display_sub.a
sp?apid=20, 2010-04-16, Downloaded spread
sheet for calculating heat output.
[4] Swedish Energy Agency, 2008, Energy Indicators
2008, Theme: Renewable energy, 2008
[5] http://a-energi.jetshop.se/, 20010-04-20
[6] EN 15316-4-5:2007, Heating systems in buildings.
Method for calculation of system energy
requirements and system efficiencies, CEN,
Brussels, 2007
[7] J. A. Myhren, S. Holmberg, Design consideration
with ventilation-radiators: Comparisons to traditional
two-panel radiators, Energy and buildings 41, p.
92-100, 2009
[8] A. Trschell, Vrmesystem med luftvrmare och
radiatorer, En analys av funktion och prestanda,
Licentiate Thesis, Chalmers, Gteborg, 1999
[9] S. W. Churchill, Correlating equations for laminar
and turbulent free convection from a vertical plate,
Int. J. Heat Mass Transfer, Vol. 18, p. 1323-1329,
1975
[10] J. P. Holman, Heat transfer, 9th edition, 2002
[11] Discussion with professor B. Sundn, April 2010
[12] P. Selinder, H. Walletun, Modell fr ndrade
frutsttningar i fjrrvrment, Rapport 2009:50,
Svensk Fjrrvrme, 2009
[13] S. Werner, FVB Sverige AB, Nytta med svensk
fjrrvrmeforskning, FoU orientering 2004:9,
Svensk Fjrrvrme, 2004
September 5
th
to September 7
th

31

PRIMARY ENERGY EFFICIENCY AND SYSTEMS ENGINEERING
M.Berner
1
, R. Ulseth
1
, J.Stang
2

1
Norwegian University of Science and Technology (NTNU)
2
SINTEF Energy Research

ABSTRACT
The revised Energy Performance of Building Directive
(EPBD) [1] emphasizes that the energy performance of
a building shall be calculated by use of Primary Energy
Factors (PEF). Calculation of CO
2
emission will not be
mandatory so far. Thus EPBD will reduce the use of
non-renewable energy, incite the use of energy from
combined heat and power generation (CHP) and
reduce the energy consumption in the building sector.
A simplified method that enables comparison of the
PEF from different energy chains is required. However,
calculation of all the parameters affecting the PEF
values like energy used for extraction, transportation,
power and heat generation etc. is time-consuming. The
method described in EN 15603 [2] is rather general l
and provides PEF values for 13 energy carriers and
chains. This is based on average European values. Life
Cycle Assessment methods include several of the
relevant steps, but a complete LCA often imply
collection of more than 6000 parameters.
The systems engineering method used here have
demonstrated the feasibility of developing a generic
method that provides credible data for calculating
primary energy efficiency. It applies the generic method
on energy chains in the Nordic region which is relevant
to CHP plants utilising bio based fuel.
INTRODUCTION
Background
The terms Primary Energy, Primary Energy Efficiency
and Primary Energy Factors (PEF) are introduced [3] -
[8] in order to compare different energy sources and
chains based on losses and a calculated environmental
impact.
Primary energy is energy that has not been subject to
any conversion or transformation process. The use of
primary energy factors takes into account the energy
that are used from the extraction of the energy carrier
and all of the losses until energy is delivered to the end
use in the desired form such as heat, cooling or
electricity .
The primary energy factor (PEF) expresses how much
primary energy is needed to deliver 1 unit of power,
heat or cooling to the end user. The term primary
energy efficiency (PEE) therefore is used to describe
the total use of energy from extraction to the end user.

Figure 1 A Typical energy chain

Methodology
An energy chain might consist of several elements or
processes from extraction, through processes such as
drying, storage, transport, power/heat/cool generation,
and distribution to the end user. In order to ensure that
there is a correct PEF, all elements that influence the
energy flow have to be accounted for.
The energy balance or calculation of the energy
efficiency of a process focuses primary on the energy
input in the form of fuel and the output in kWh, and
lacks information on the energy used to build
infrastructures such as the power plant, distribution net,
transportation and the extraction.
Life Cycle Assessment (LCA) might contribute to
provide such information in a generic method.
However, the number of input parameters, often more
than 6000 in an ordinary LCA analysis demonstrates
the need for an easily accessible method.
Systems engineering is a method that has been
developed gradually with increasing complexity of
projects and systems. Systems engineering is often
considered to have started at Bell Laboratories in the
1940s, later applied in organizations such as NASA
and formalized as a separate engineering field with the
formation of INCOSE [9] in 1990. The benefits of
systems engineering is the possibility to treat complex
systems with several subsystems. Therefore, as a first
step in the development of a method a systems
engineering approach has been chosen. The main
objective is to develop systems and methods that
enable a sufficiently reliable calculation to be made of
the primary energy factor (PEF) in general and for
different energy chains with required level of details.
At present systems engineering approaches have not
been found to have been previously applied on the
development of generic PEF methods for different
energy chains.
Objective
The objective of this paper is to show how systems
engineering can be used as a tool to reveal important
E
x
t
r
a
c
t
i
o
n
P
r
o
c
e
s
s
i
n
g
T
r
a
n
s
p
o
r
t
S
t
o
r
a
g
e
G
e
n
e
r
a
t
i
o
n
T
r
a
n
s
f
o
r
m
a
t
i
o
n
T
r
a
n
s
m
i
s
s
i
o
n
D
i
s
t
r
i
b
u
t
i
o
n
September 5
th
to September 7
th

32

parameters when a model for calculation of the PEE of
different energy chains is developed. The paper will
show an overall approach and will not describe all the
necessary iterations in detail.
SYSTEMS ENGINEERING
The system engineering process
A systematic approach such as systems engineering is
essential to be able to develop a generic model
describing a complex system with several subsystems.
The intention with the systems engineering process is
to analyse and describe complex systems. Often the
method is used in the design process, to make sure
that the subsystems are connected properly, that the
process is optimized and that the different components
are described, implemented and integrated precisely.
A common feature of all systems engineering
processes is an indefinite number of iterations at all
different steps.
Systems engineering principles are often applied when
a new system or products are developed. The
methodology alters slightly between development and
re-engineering.
Re-engineering methods are applied when an existing
system is described. The energy chains considered are
already designed and built, and a re-engineering
technique is selected in order to develop a method that
calculates the PEF for different kind of energy chains.
CHOSEN METHODOLOGY
The system re-engineering process consists of the
following six different tasks according [2]. Some of
them might seem unnecessary, but they all contribute
to the decomposing of a system and development of a
method.
1. Establish problem statement;
This comprises the definition of the problem approach,
which includes development of a problem statement
describing the problem/challenge, its importance and a
state of the art. To be able to establish the problem
statement; four questions must be answered:
What is the problem?
Why is it important
What have others done?
What must be done?
2. Assess available information assessment
Provide available information including an overview of
possible stakeholders.
3. Measures of effectiveness (MOE)
The definition of MOE are: A small subset of the
requirements that are so important that the system will
fail if they are not met and will be a huge success if
they are met [11].
4. Development of information models
The different information models describe the observed
system in relation to legislation, physical architecture
and a system interface model. Four separate models
are developed
- Requirement traceability model
- System architecture model
- Behaviour model
- System interface model
5. Trade-offs
The trade-off phase is essential in the development of
a method. Each of the steps is carried out in iterative
loops gradually increasing detailing level. After
satisfactory trade-offs have been performed and
consistent information models obtained, a theoretical
method is developed. Real data are collected and trade
-off between the model and the gathered data are
performed.
6. Documentation
The developed method will be then documented by
actual case studies before a final reporting.

Figure 2 The system re-engineering process described
as a functional block diagram (FFBD), ref. [10]
ESTABLISH PROBLEM STATEMENT
What is the problem?
Use of Primary Energy Factor (PEF) will provide
information on the energy losses and consequently the
environmental impact of different kind of energy
sources, power production processes and energy
1
Establish
problem
statement
2
Asssess
available
inform-
ation
3
Define
effective-
ness
measures
6
Document
current
system
design
5
Trade-
offs
Create
requirement
traceability model
4
Create context
model
4
Create system
architect. model
4
Create behaviour
model
4
and
Feasible
solution
No
feasible
solution
Iterate to find feasible
solution
1
Establish
problem
statement
2
Asssess
available
inform-
ation
3
Define
effective-
ness
measures
6
Document
current
system
design
5
Trade-
offs
Create
requirement
traceability model
4
Create context
model
4
Create context
model
4
Create system
architect. model
4
Create system
architect. model
4
Create behaviour
model
4
Create behaviour
model
4
and and
Feasible
solution
No
feasible
solution
Iterate to find feasible
solution
September 5
th
to September 7
th

33

transport systems. At preset there exists no easy
accessible calculation method.
Different countries have different energy chains and
energy supply systems. Analysis of even the most
actual processes and process lines does not exist
neither for Norway or Europe [12] [13]. In order to
compare and choose different energy chain there is a
need for standardized methods. The lack of objective
and reliable data of the different elements in the energy
chain might prevent an efficient use of energy, and
contributes to wrong choices and unnecessary CO
2

emissions. [14] [15].
The method is principally described in EN 15603 [2]
and provides only single PEF values for 9 energy
carriers and 4 energy chains, and is based on average
European values. Without an easy accessible method
or methods is it not possible to compare PEF values
and calculate the actual environmental impact of
different energy chains. Some studies [16] -[21] have
described parts of this topic, but they lack a holistic
view of the energy chains from cradle to grave, often
the chosen system boundaries are different, time scale
varies, detailing level different and the, approach/
method varies. Results from different studies therefore
are not comparable.
Why is it important
PEF is a key indicator to be able to evaluate energy
use (for different purposes) especially with regards to
the goals of the EPBD [1]. PEF is an over all energy
efficiency indicator which makes it possible to compare
and collocate different energy sources and energy
carriers by a single number. The same method can be
used to calculate the CO
2
emission.
What have others done?
Different CEN standards describe, and partly discuss,
the theory. In the EC-mandated CEN standards related
to EPBD mainly one single reference are referred [14]
whilst the PEF values have been gradually changed
over time. An extended literature survey has showed
discontinuity between some of the studies performed
and lack of details in the calculations.
Methods developed to provide PEF values for heating
systems in buildings might be useful, but they will not
totally comply with a whole energy chain approach. Life
Cycle Assessment (LCA) might also contribute to a
generic method, but the vast number of input
parameters, often more than 6000 in a traditionally LCA
demonstrates the need for a more easy accessible
method.
What must be done?
In order to develop a method a systems engineering
approach will be used. The most important task in this
context is the identification of relevant energy systems
and process lines (chains) primary in Norway and the
Nordic countries. Detailed data must be provided such
as efficiency and loss from the different systems and
mix of systems, or at least provide the necessary
parameters. Since the systems engineering approach
is chosen, the problem approach must be defined, a
theoretical method developed and data collect. This
includes performing of a trade-off between the
theoretical model and available information. The
method shall be tested by selected case studies and
finally adjusted.
Main hypothesis
As a part of the systems engineering process, one or
several (systems engineering) hypothesis is developed.
The success of a system engineering process is
related to the fulfilment of the hypothesis. In this project
the system engineering method must prove two main
hypotheses;
1. It is possible to develop a generic method that
provides credible data for calculating primary
energy use by use of PEF values.
2. It is possible to apply the generic method on
energy systems in the Nordic region for CHP
plants utilising bio based fuel.
Stakeholder analysis
A stakeholder is a party having a right, share or claim
in the system [16]. The intention with the stakeholder
analysis is to reveal the different kinds of stakeholders
since they might have requirements influencing a
possible method in a legal way. Stakeholders with
mutual interest are aggregated in groups; some of
them might not be in incompliance with each other.
- Energy producer, distributors, energy companies;
Business profitability is the main issue by
optimizing production from different energy carriers
according to cost-benefit
- Investors (energy and building); The electricity
markets are opening gradually throughout Europe,
e.g. Nord Pool Financial Marked and investments
in Power production and the introduction of so
called Green Electricity Certificates might be a new
or extended business area. Investors in the
building marked might be interested in the actual
PEF values and primary energy use when
choosing between different investment objects.
- Building owners, end user; Correct calculation of
PEF values and primary energy use is supposed to
have significant importance for the choice of
energy supply system, building services, insulation
level, especially for new buildings and major
rehabilitation projects. Future operating cost might
September 5
th
to September 7
th

34

depend of PEF since taxes might be dependant of
the primary energy use and/or the CO
2
emission.
- Developers& Building and construction industry;
Technical equipment in the building and design
strategies depends on the actual use of the
specific energy carrier. The use of PEF values in
the primary energy calculations might change the
value of traditionally installed equipment due to
overall energy costs and also create a demand for
more energy flexible solutions.
- Politicians, government, Regulators, Community
planning; Most European countries have affiliated
the Kyoto Protocol, and a possible method to
increase the use of renewable energy policy tools
and subsidy schemes might be based on the use
of PEF values for the different solutions, besides
possible tax on systems with high primary energy
use.
National regulators mandatory monitor and report
emission levels and this influences national
legislation, local and urban planning
- Research groups, Universities; Different research
communities might be interested in development of
other PEF calculation methods or adjustments of
methods and development of new solutions and
systems
DESCRIPTION OF MEASURES OF EFFECTIVENESS
As earlier stated the measures of effectiveness (MOE)
should be independent of any solutions and not
concerned with internal details [22]. It might also be
fruitful to develop MOE for the different kind of
stakeholders since they often might have a different
opinion regarding MOE.
In this context MOE are primarily described for the
ongoing Nordic PhD-project Primary Energy Efficiency
(PEE). A further detailing level, by including the
stakeholders, might provide valuable information, but
that is considered to lie outside the scope of this work.
The methods (tools) developed during the project
should be suitable for different kind of energy chains.
- The results should be utilised by the different
kinds of stakeholders e.g. the building owner, the
architect/designers of the building, the energy
supplier and producer and finally politicians and
governments.
- The methods will enable the different stakeholders
to choose between different energy systems and
furthermore be able to reduce primary energy use
and CO
2
emissions from stationary energy
purposes.
INFORMATION MODELS
The requirement traceability information model
In systems engineering shall the requirement
traceability information model aim to show the break
down of requirements from source documents to final
allocation functions to stakeholders [2]
This model is an important tool to keep track of the
different requirements, source documents and
eventually what the system accomplishes and who or
what are in charge. Usually an Entity-Relationship-
Attribute method is used [23], where the entities
(objects) represent the legislation, requirements, etc.
whilst the relationship shows the association between
the system/process.
Planning and
Building Act
Source
Planning and
Building Act
Source
Pollution Control
Act
Requirements
Pollution Control
Act
Requirements
96/62/EC
Ambient Air
Quality
Source
96/62/EC
Ambient Air
Quality
Source
1999/30/EC Limit
values
NOx
Requirements
1999/30/EC Limit
values
NOx
Requirements
Discharge permit
Requirements
Discharge permit
Requirements
Internal control
system
function
Internal control
system
function
Building
regulation TEK
Source
Building
regulation TEK
Source
NS3031
function
NS3031
function
Building Guide
REN
Source
Building Guide
REN
Source
Building Permit
stakeholder
Building Permit
stakeholder
Documents
Incorporates
Specifies
Allocated to
Documents
Incorporates
Specifie
s
Allocated
to
Working Environm. Act
Source
Working Environm. Act
Source
EPBD Energy
Performance.
Source
EPBD Energy
Performance.
Source
EN 15316-4-
4:2007)
Source
EN 15316-4-
4:2007)
Source
N 15603:2007
Source
N 15603:2007
Source
Documents
NOx emissions
Boiler
NOx emissions
Boiler
Disharge permit
Requirements
Disharge permit
Requirements
Specifie
s
NOx emission
1.2.2.1
Stakeholder
NOx emission
1.2.2.1
Stakeholder
Specifies Allocated to
Landfill Directive
Source
Landfill Directive
Source
Waste regulations
Source
Waste regulations
Source

Figure 3 Selected part of the requirement traceability
model, case Norway

Most EC directives are enforced and implemented in
laws, directions, regulations and guidelines, both in the
EU and associated EEC countries, hence the order of
entities in Figure 3. Several directives influence the
national laws and regulations. Since the directives
usually are enforced through national laws, the law
includes requirements from more than one directive like
the Norwegian Planning and Building Act [26], which
includes requirements from EPBD [27], Directive on the
promotion of the use of energy from renewable sources
[30], Directive on the landfill of waste [29], The
Pollution Control Act [30] amongst others.
The requirements traceability model provides important
information about constraints regarding an energy
chain. Some of the elements such as the the Working
Environment Act [31] might seem irrelevant, but the
regulations set limits for the pollutant inside the working
area, introducing need for e.g. conveyor belts.
Each function consists of several entities for instance
the discharge permit from regulators like The
Norwegian Climate and Pollution Agency will set
restrictions on the authorised discharge levels of
different gasses not only NOx as illustrated in the
Figure 3.
September 5
th
to September 7
th

35

The requirements affect most of the stakeholders, for
instance the Planning and Building Act will affect both
end users, construction industry and energy
distributors.
Architecture information model
The architecture information model shows the physical
components of a system with subsystems.
In order to describe the possible physical systems a
generic model is developed [11], detailed description of
some of the most relevant energy chains are carried
out in the actual PhD thesis, Figure 1 shows a principal
descriptions of a single energy chain. An end user will
typically be supplied with energy from a various
number of energy chains, and each element might
represent parallel processes.
A CHP utilizing biomass might consist of the following
elements:
- Fertilizing, cultivation, logging, logging track, loop
of twigs, trimming, transport
- Chipping, packing, transport, local roads
- Intermediate storage, transport regional roads
- Transport central and regional roads
- Building, operation demolition of power plant,
technology, efficiency, part-load, size, Lifetime,
waste treatment, gas cleaning supply of additives,
internal transport
- Transformation to central net, building, operation,
demolition of infrastructure, heat/power loss
- Transmission to local net, building, operation,
demolition of infrastructure (pipes, high-tension
lines heat/power loss (insulation, temperature
levels (supply, return, ground), twin/single pipes,
length)
- Distribution to end user, building, operation,
demolition of infrastructure (pipes, lines,
substations) heat/power loss

Figure 4 Architecture information model for a part of the
energy chain from generation including distribution, based
on [25]
An energy chain consists of several sub systems as
described in Figure 4. A more detailed architecture
information model is also developed.
Each of those sub elements can be spilt up into sub
elements as shown in Figure 5. The final or basis
element can be described as Figure 6.

Energy
Transformatio
n system
Component
Energy
Feeding
system
Component
Waste
handling
system
Component
Combustion
process
Component
Energy
Production
system
Component
Combustion
camber
Component
Purification
system
Component
Energy
Transformatio
n system
Component
Electricity
production unit
Component
Heat Storage
system
Component
Heat
production unit
Component
Heat
Transport
system
Component
Consist of
Filter
Component
Consist of Consist of
Consist of
Built of
Consist of
Consist of
Consist of
Internal
Control
System
Stakeholder
Energy
Transformatio
n system
Component
Energy
Transformatio
n system
Component
Energy
Feeding
system
Component
Energy
Feeding
system
Component
Waste
handling
system
Component
Waste
handling
system
Component
Combustion
process
Component
Combustion
process
Component
Energy
Production
system
Component
Energy
Production
system
Component
Combustion
camber
Component
Combustion
camber
Component
Purification
system
Component
Purification
system
Component
Energy
Transformatio
n system
Component
Energy
Transformatio
n system
Component
Electricity
production unit
Component
Electricity
production unit
Component
Heat Storage
system
Component
Heat Storage
system
Component
Heat
production unit
Component
Heat
production unit
Component
Heat
Transport
system
Component
Heat
Transport
system
Component
Consist of
Filter
Component
Filter
Component
Consist of
Built of
Consist of
Consist of
Consist of
Internal
Control
System
Stakeholder
Internal
Control
System
Stakeholder

Figure 5 Segment/selection of part of the architecture
information model.

PEF
in
Loss
PEF
out
Infrastructure,
buildings,
machinery etc.
Additional PEF
Operation and
maintenance
Demolition
PEF
in
Loss
PEF
out
Infrastructure,
buildings,
machinery etc.
Additional PEF
Operation and
maintenance
Demolition
Figure 6 Architecture information model, basis element

Since the Primary Energy Efficiency (PEE) of an
energy chain consist of all of the elements from
extraction to delivery the PEF for a chain can be
calculated by
on Distributi
on Transmissi tion Transforma
Generation Transpor Storage
Processing Extraction Fuel Chain
E
E E
E t E E
E E E E
A +
A + A +
A + A + A +
A + A =
+

Where E is the primary energy input to the
system
(1)

The Power Bonus Method
In [13] the power bonus method is applied to calculate
the PEF value for a district heating system with CHP.

= ) ( ) (
exp, , exp, , , , i P i i del P i del P
f E f E E (2)
where
E
P
Primary energy input to the system
E
del,I
Delivered energy, energy carrier i
(1)

Fuel
Heat
prod
Cooling
prod
Dist.
net
Sub
station
Construction
Dismantlin
g
Construction
Construction
Construction
Dismantling
Dismantling
Dismantling
Waste
handlin
g
Other
(chem
etc)
Cold
Heat
Hea
t
Col
d
Elec
prod
Electricit
y
Fuel
in
Heat
Cool
Construction
Dismantling
September 5
th
to September 7
th

36

f
P,del,i
Primary energy factor, delivered energy carrier i
E
exp,
Exported energy, energy carrier i
f
P,expl,i
Primary energy factor, exported energy carrier i
Power exported from the CHP plant multiplied with the
PEF value for the replaced power shall be subtracted
from the delivered primary energy to the buildings
when calculating the PEF-value for the for the district
heating system [7]. The power bonus method is
enforced in order to promote CHP, and the subtraction
of power produced and delivered outside the system
boundary significantly reduces the PEF value for the
energy chain. This implies that the PEF value for a
CHP will be dependent on the power to heat ratio.
Behavioural model
The behavioural model is another information model
the what it does[10], but also described as the way in
which an organism, organ, or substance acts,
especially in response to a stimulus [23].
A behavioural model consists of functions, inputs and
outputs and control operators. This implies that it is
supposed to provide information on what is happening,
in which order and what kind of iterations are
performed.
Establish problem
statement
Assess available
information
Define MOE
Energy source
Energy transport
system
Energy
production
Develop generic
model
Trade-off
Collect data
case study
Simplify method
Publish model
Test, evaluate
method
And
And
Establish problem
statement
Assess available
information
Define MOE
Energy source
Energy transport
system
Energy
production
Develop generic
model
Trade-off
Collect data
case study
Simplify method
Publish model
Test, evaluate
method
And
And

Figure 7 Simplified behaviour information model of the
model developing process.

A more detailed partition of the energy chains have
been applied in the development of the method. The
energy chain is divided in subsystems as shown in
Figure 4 and Figure 5. This is an iterative process and
the detailing level is the first steps gradually increasing,
until the analysis (trade-off) of the different factors
influencing the PEF value enables a removal of factors
with an impact of 1% or less.
System interface information model
The system interface model also denoted the context
information model shows the systems interface with its
surroundings and the environment. The model provides
information on the core system and other
interconnecting systems; this means a description on
how things relate to each other.
The context is according to [10] the interrelated
conditions in which something exists or occurs.

Figure 8 A simplified context information model.

The system boundary is drawn with a dashed line, and
the system assessed lies within. Since this is a
simplified model the relation towards investors, national
regulators, constructors etc. are not shown. In this
system energy source/carrier is closely connected to
Extraction, Energy Source consists of storage and
transport, Energy production corresponds with
Generation and Energy transport system to Trans-
formation, transmission and distribution in Figure 1.
The main issue has been to show the connection
between the energy chains from production to end use.
Politicians and national regulators might have specific
requirements on each level. The building industry,
constructors may likewise have interest on several of
the levels, but a final listing is not possible to provide
within this paper.
Another important question is the definition of the
system boundary. Precise definitions of the system
boundaries are essential in order to be able to compare
different studies. The system boundaries must
distinguish between what is included and what lies
outside of the task of the LCA, since the method must
also rely on data collected by other parties and the use
of different constraints might influence the quality of the
method.
In order to provide information of the whole energy
chain, all major elements have to be included i.e. the
extraction phase is an integrated part of the chain.
Energy
prod. Raw
material
Energy
source
carrier
Energy
transport
system
Energy
productio
n
End user
Energy
Requirements
Requirements
Requirements
Requirements
Energy
Energy
Energy
September 5
th
to September 7
th

37

Trade-offs
In the ongoing PhD-project of the first author a cut-off
rule is set to 1%. That implies that factors with less
than 1% impact on the final result can be removed.
The trade-off considerations are still an ongoing
process, and it will be presented and documented in a
later paper. According to the main hypothesis this is not
meant be developed as an optimization tool, since the
intention of the ongoing PhD-project is to develop a
method describing different energy chains.
A complete trade-off could preferably [32] be performed
by use of computerizes programme like CORE.
[http://www.vitechcorp.com/solutions/]. The complexity
of the different kind of systems shows the utility value
of more than manual tools, which has been applied.
Document current system design
The results of the iterative process are described in the
figures mentioned above. Only selected parts of the
chosen design are illustrated in this document due to
limitation in size.
The final system design is carried out according to
Figure 6 for each element.
CONCLUSION
By performing a system engineering process
describing different energy chains an outline of the
model have been developed. The method has proven
to be efficient in structuring the thoughts and will
hopefully reduce mistakes in the future development of
the model.
The decomposition process in different subsystems is
valuable, and the generic model will be able to treat
different kind of energy systems and chains.
The systems engineering process have demonstrated
that;
1. It is possible to develop a generic method that
provides credible data for calculating PEF-values
and the primary energy efficiency.
2. It is e.g. possible to apply the generic method on
energy systems in the Nordic region with CHP
plants utilising bio based fuel
The system engineering process provides a new
approach to the design and development of a generic
model describing PEF-values for energy systems with
different kind of energy carriers. The method might be
used for more than systems using CHP-technology
since the model development are generic and thereby
utilizes different kind of energy carriers.
The method can provide detailed data (e.g. efficiency,
loss etc) from the different energy chains and mix of
chains. A major challenge is the data collection, some
of the parameters lack standardization. The life time of
different equipment varies, the economical lifetime is
often significant lower than the actual exchange rate
e.g. pipelines might have a twice times higher - more
than 60 years. The use of yearly average efficiency and
appurtenant power-to-heat ratio will often increase the
PEF value for the whole system due to the impact of
the power bonus method.
The reliability of the method will be influenced by
possible lack of detailed data, but based of average
data a reliable comparison of different energy chains
might be performed.
More standardized values for some the different
parameters needs to be developed, like lifetime, heat
load curves and extraction of biomass. Some
adjustment will be necessary for instance for extraction
where the transport distances are an important
parameter. The resulting model can form a basis for
future optimization tools, since only elements with
major influence on the PEF-values are included.
ACKNOWLEDGEMENT
This paper is developed as a part of the PhD-project
Primary Energy Efficiency (PEE) and the work title for
thee PhD-Theses is "System, methods and credible
data for calculating primary energy efficiency in general
and for energy systems in the Nordic region with special
focus on energy systems applying CHP-technology with
bio based fuel in particular".
The project is financed by Nordic Energy Research
with financial support from the industry and includes six
PhD-studies carried out in the respective countries;
Estonia, Finland, Sweden, Iceland and Norway. The
projects objective is to contribute to the effort of
enhancing the primary energy efficiency (PEE) and
reducing CO2-emissions in the energy sector.
FURTHER INFORMATION
PhD.student Monica Berner, Norwegian University of
Science and Technology (NTNU).
Address: Monica.Berner@ntnu.no
REFERENCES
[1] Proposal for a Directive of the European
Parliament and of the Council on the Energy
Performance of Buildings Recast SEC (2008)
2820, SEC (2008) 2821)
[2] EN 15603 Energy performance of buildings
Overall energy use and definition of energy ratings
[3] EN 15603:2007 Energy performance of buildings
overall energy use and definition of energy ratings
[4] EN 15316-1: 2007 Heating systems in buildings
requirements and system efficiencies Part 1:
General
September 5
th
to September 7
th

38

[5] EN 15316-2-1:2007 Heating systems in buildings
method for calculation of system energy
requirements and system efficiencies part 2-1
space heating emission systems
requirements and system efficiencies Part 2-3
Space heating distribution systems:
requirements and system efficiencies Part4-4
Heat generation systems, building-integrated
cogeneration systems
requirements and system efficiencies Part 4-5
Space heating generation systems, the
performance and quality of district heating and
large volume systems
[9] INCOSE, International Council on Systems
Engineering, A Consensus of the INCOSE Fellows,
www.incose.org
[10] Dahl H J, Information modelling and systems re-
engineering an efficient approach to assessing
complex current Norwegian natural gas transport
operations, Proceedings of the Tenth Annual
International Symposium of the International
Council on Systems Engineering (INCOCE), July
2000
[11] Olivier DW, Kelliher TP, Keegan JG, Engineering
complex systems with models and objects, ISBN
048188-1, McGraw-Hill, 1997
[12] Joelsson. A. Primary Energy efficiency and CO2
mitigation in Residential buildings, Doctoral Thesis
58/2008, Mid Sweden University (Dissertation
3.October 2008)
[13] Berner M., Ulseth R., The Primary Energy
Concept, The 11th International Symposium on
District Heating and Cooling, August 31 to
September 2, 2008, Reykjavik, ICELAND
[14] Frischknecht, R, Jungbluth et al, 2007,
koinventare fr energiesysteme Grundlagen fr
den kologishen Vergleich von Energiensystemen
und den Einbezug von Energiesystemen in
kobilanzen fr die Schweiz , ETH, Zrich 1996
[15] CEN/ CLC BT JWG, Energy Management, 2005)
[16] Nrsteb V., Application of systems engineering
and information models to optimize operation of
gas export systems, Systems Engineering archive,
Volume 11 , Issue 4 (November 2008), p: 329-
342, 2008, ISSN:1098-1241
[17] Sther S, Thermal Heat and Power Production
with models for local and Regional energy
Systems, ITEV-Report 1999:06, Dr.ing Thesis
1999:117, NTNU
[18] Sarigiannis D.A., Triacchini G., Meso-scale life-
cycle impact assessment of novel technology
policies: The case of renewable energy, Journal of
Hazardous Materials 78, 2000 p. 145-171
[19] Alanne K., Salo A., Saari A., Gustafsson S., Multi-
criteria evaluation of residential energy supply
systems, Energy and buildings 39, 2007 p 1218-
1226.
[20] Eriksson O, Finnveden G, Ekvall T, Bjorklund A,
Life cycle assessment of fuels for district heating: A
comparison of waste incineration, biomass- and
natural gas combustion, energy Policy 35, 2007
p.1346-1362.
[21] Mnster M., Lund H., Use of waste for heat,
electricity and transport Challenges when
performing energy system analysis. Energy 34,
2009 p. 636-644
[22] Lenzen M., Life cycle energy and greenhouse gas
emissions of nuclear energy: A review, energy
Conversion &Management 49, 2008 p.2178-2199
[23] Sproles N, Coming to Grips with Measures of
Effectiveness, John Wiley & Sons, Inc Syst Eng.
3:50-58, 2000
[24] Olivier, Merrian Webster 1981
[25] Berner M, Primary Energy Concept and Life Cycle
Assessment (LCA), Report no: 2009/001, June
2010, The Norwegian University of Science and
Technology
[26] Act of 14 June 1985 No. 77 the Planning and
Building Act, The Ministry of the Environment and
the Ministry of Local Government and Regional
Development
[27] Directive 2002/91/EC of the European Parliament
and of the Council of 16 December 2002 on the
energy performance of buildings.
[28] Directive 2009/28/EC on the promotion of the use
of energy from renewable sources and amending
and subsequently repealing Directives 2001/77/EC
and 2003/30/EC
[29] Council Directive 1999/31/EC of 26 April 1999 on
the landfill of waste
[30] Act of 13 March 1981 No.6 Concerning Protection
Against Pollution and Concerning Waste, [The
Pollution Control Act]
[31] Act of 17 June 2005 No. 62 relating to working
environment, working hours and employment
protection, etc. as subsequently amended, last by
Act of 23 February 2007 No. 10, (The Working
Environment Act)
[32] Purves B, Information Models as a Prerequisite to
Software Tool Interoperability, Incose Insight, 1998
September 5
th
to September 7
th

39

ENHANCED DISTRICT HEATING AND COOLING SYSTEMS
REALISATION OF THE LOW-EX CONCEPT
Stefan Bargel
1
, Clemens Pollerberg
1
, Armin Knels
2
, Li Huang
1
, Dirk Mller
2
and Christian Dtsch
1
1
Fraunhofer Institute for Environmental, Safety, and Energy Technology UMSICHT,
Osterfelder Strasse 3, 46047 Oberhausen, Germany,
Phone: +49 (0) 208-8598-1276, Fax: +49 (0) 208-8598-1423,
stefan.bargel@umsicht.fraunhofer.de, clemens.pollerberg@umsicht.fraunhofer.de
2
RWTH Aachen University, E.ON Energy Research Center - EBC,
Mathieustr. 6, 52074 Aachen, Germany,
Phone: +49 (0) 241-8049-780, Fax: +49 (0) 241-8049-769

ABSTRACT
Since heating and cooling represent low-exergy energy
streams, high efficiencies can be obtained, if the
energy demand is covered by appropriate meaning
also low-exergy (low-ex) input energy flows.
In order to be able to employ great potentials of low-
exergy heat from many different sources, it is important
to develop technologies for the supply and the use of
energy that allow network temperatures close to
ambient temperature in return as well as in supply
pipes. Two possible technologies are phase change
slurries (PCS) and capillary tube mats (CTM).
PCS are discussed as heat transfer fluid, which has an
increased heat capacity compared to water. The use of
PCS in energy supply networks instead of water leads
to an improved energy transport capacity, which results
in a reduction of the necessary temperature difference
of the transfer fluid. To ensure the transfer of energy
from the supply network into the building while the
temperature difference between network and building is
low, large heat transfer areas are required, which can
be achieved by the use of CTM.
This paper discusses opportunities for the realisation of
cold supply networks and low-ex systems and presents
exemplary technologies for their realisation.
INTRODUCTION
Temperature levels in district heating and cooling
networks have long been discussed. During the last
years a tendency towards low temperature networks
can be observed. From a scientific point of view
answers to the question for the optimal temperature
levels can be given using exergy efficiencies as for
example discussed in [1]. The main advantage of this
evaluation parameter is the thermodynamically correct
distinction of thermal (low-exergy) and non-thermal
(high-exergy) energy flows.
Since heating and cooling represent low-exergy flows,
it is of uttermost importance to cover these demands by
appropriate meaning also low-exergy input energy
flows. For example a heating system based on a
domestic gas boiler used to provide space heating
wastes a huge amount of exergy, since the exergy
efficiency of such a system reaches only approximately
5%! This result is valid for arbitrary heating systems in
the supply target (room) itself. Therefore it is
mandatory to use an integral evaluation approach to
decide whether an energy system is efficient or not.
With respect to district heating and cooling networks as
energy supply systems two findings are important.
First, it can be shown that the network subsystem itself
as depicted in figure 2 reaches optimal exergetic
efficiency at quite low temperatures since the heat
losses dominate the pumping electricity effort.
Secondly the overall energy supply system efficiency
can be greatly enhanced by utilising low-exergy input
energy flows such as industrial waste heat.
In order to be able to employ great potentials of low-
temperature waste heat from many different sources, it
is important to develop technologies for the supply and
the use of energy that allow network temperatures
close to ambient temperature in return as well as in
supply pipes.
Today, district heating and cooling networks use water
as heat transfer fluid. The heat is transported as
sensible heat and the transport capacity of the
networks is determined by the heat capacity of water
and the temperature difference between forward and
backward flow. In cold supply networks as well as in
low temperature heating networks, high volumetric flow
rates are necessary to provide the required transport
capacity due to the comparably small temperature
difference between forward and backward flow. To
overcome these restrictions, a new heat transfer fluid
with an increased heat capacity is under development
as an alternative to water, phase change slurries. PCS
are mixtures of dispersed phase change material and a
continuous liquid phase, which can be used as heat
transfer fluid in district heating and cooling networks.
PCS possess an increased heat capacity due to
additional latent heat of fusion occurring during the
phase transition of the phase change material. The use
of such a dispersion in energy supply networks leads to
an improved energy transport capacity, which in turn
September 5
th
to September 7
th

40

results in a reduced temperature difference or
volumetric flow rate of the transfer fluid needed to
transfer a given amount of heat. The application of
PCS for thermal energy transportation is investigated
and discussed for example in [2].
An improved transport capacity is one important point
for the realisation of the low-ex concept; another
important point is the use of the energy on the
consumer side. To ensure the transfer of the energy
from the supply network into the building while the
temperature difference between network and building is
low, large heat transfer areas are necessary. These
heat transfer areas can be realised by using capillary
tube mats integrated into the walls, the floors and the
ceilings of buildings.
The E.ON Energy Research Center of the RWTH
Aachen and Fraunhofer UMSICHT investigated the
possibilities to realise district heating and cooling
networks as low-ex systems. These investigations
include system modelling and analysing as well as the
development and testing of technologies.
1. Exergy as evaluation parameter
1.1. The low-ex concept
Exergy can be understood as the theoretical maximum
of mechanical work that can be utilised by equilibrating
an energy flow whilst considering its ambient
conditions.
Consequently this property distinguishes between
types of energy that can theoretically be transformed
into each other without any losses - like mechanical
work, electrical energy or combustible fuels - and
thermal energy. The possibility to transform the latter
into any other type of energy is limited by the second
law of thermodynamics and therefore inevitably
connected to losses.
This distinction is of importance if one analyses a
system where both types of energy flows (thermal and
non-thermal) occur and have to be related to each
other as is the case with heating and cooling
applications.
The ultimate goal of heating and cooling is to keep a
target (room) at a constant temperature of e.g. 20 C.
As the outdoor temperature varies additional heat has
to be supplied or excess heat has to be disposed of to
fulfil this task.
Theoretically the supplied energy flow could be
transferred to the room using infinitesimal small
temperature differences between supply flow and
target
2
. The real temperature differences occur due to
heating and cooling techniques applied which are
mainly limited by finite heat transfer areas. Keeping in

2
This statement is analogously true for cooling applications.
mind that the annual average outdoor temperature for
the heating period e.g. in Germany is about 3.5 C, it
becomes apparent that the exergy to energy ratio of
the target energy flows - passing the building envelope
at 20 C is very small (approx. 7%). On the other
hand, exergy to energy ratios of conventional input
energy flows are usually 100% as combustible fuels or
electricity is used.
The low-ex concept acknowledges the fact that
demand flows are low-ex - meaning that they possess
small exergy to energy ratios. Hence the concept
demands to supply energy on appropriate exergy
levels, instead of wasting exergy by transforming high
exergy flows into low exergy ones. In doing so, this
approach opens up a totally new dimension of
enhancement potential since it deals with the quality
aspect of the energy flows under consideration.
Therefore, within the low-ex concept energy is no
longer one-dimensional. In addition to decreasing the
amount of energy demanded by the consumers
leading to insulation efforts a kind of exergetic
suitability has to be taken into account and the task at
hand becomes a two-dimensional problem (cf. fig. 1).
Consequently, the concept aims at maximizing the
exergy efficiency of an energy supply system, which
allows to utilize potentials in both dimensions, quantity
AND quality.
The exergy efficiency can be defined as:
=
) (
) (
supply exergy
demand exergy
ex
c (1)
In applying this efficiency the demand flows and
particularly the supply flows have to be defined
carefully (cf. chapter 1.2.).
insulat ion
energy demand (quant it y)
e
x
e
r
g
e
t
i
c

q
u
a
l
i
t
y
exerget ic suit abilit y
enhancement
low-ex
concept
insulat ion
energy demand (quant it y)
e
x
e
r
g
e
t
i
c

q
u
a
l
i
t
y
exerget ic suit abilit y
enhancement
low-ex
concept

Figure 1. Energy as two-dimensional concept. Orange (light
grey): conventional system, green (dark grey): optimal
system
September 5
th
to September 7
th

41

Figure 2. Evaluation boundaries of an energy supply system. The blue (outer) dashed line marks the complete system; the
black (inner) dashed line marks the network subsystem

1.2. Integrated system evaluation
When evaluating a system it is important to specify the
evaluation boundaries (cf. fig. 2). It has to be pointed
out that an integrated system evaluation is mandatory
since otherwise results are ambiguous and misleading.
This can be demonstrated by assuming e.g. evaluation
of the building subsystem only. If two systems are
compared, one consisting of a target room equipped
with space heating and the other one with a target
room equipped with conventional heating, one could
arrive at the conclusion, that the system utilizing space
heating is more efficient. However, assuming both
systems are also equipped with an identical
condensing gas boiler providing the heat, an evaluation
comprising the total system (consisting of heat
generation and heat transfer to the target) would arrive
at a totally different conclusion. In this case, both
systems possess the same exergy efficiency, which is
approximately 5% for the outlined case. This is
because a potentially more efficient heating system is
not put to use as the same input and supply flows
occur in both cases.
1.3 Efficiency enhancement potentials
The complete energy supply system can be divided
into three subsystems generation, distribution and
building (representing the consumption). These
subsystems possess different potentials to enhance
overall system efficiency.
Currently heating demands are met by burning high-
exergy fuels, great enhancement potentials are
available within the generation subsystem. Firstly, fuels
should not be used to directly satisfy thermal demands
at all since this embodies pure exergy destruction.
Instead thermal input flows as industrial waste heat3 or
geothermal energy should be applied. On the other
hand, if combustible fuels are used to meet thermal
demands, at least Combined Heat and
Power generation (CHP) with a maximum electrical
degree of efficiency should be utilized. This allows
transforming part of the high-exergy fuel into high-
exergy electric current. Heat is produced as waste
product of this conversion.
Optimization potentials within the distribution
subsystem are basically indirect. At first glance, the
distribution system has no influence at all since the
network acts as connection between heat generation
and heat consumption. Consequently, no thermal flows
exist that pass the overall system evaluation
boundaries. However, two aspects remain and need to
be accounted for. One is heat losses occurring
throughout the network that have to be compensated
by additional heat generation. The other is pumping to
maintain the heat transfer medium circulation, which is
met by an unalterable high-exergy input (electricity).
The main problem is that concepts, which lead to
decreasing heat losses cause increasing pumping
efforts and vice versa. Nevertheless, heat losses are
the exergetically dominant influence, therefore the
focus should be to confine these losses. Heat losses
depend on the driving temperature difference between
medium and surrounding ground and on surface area.
Minimization of the losses can most easily be achieved
by reducing the network temperatures since pipe
dimensions are affixed due to demands so that surface
areas are not a modifiable parameter. This approach is
even more rewarding since it allows employing low

3
Industrial waste heat in this sense is heat that can no more be
put to any use within the industrial production process.
heat
generat or
heat ing
surf ace
T = 20C
dist rict
heat ing
net work
dist rict
heat ing
net work
power plant power plant
heat
f low
heat
f low
pumping
elect ricit y
net work
heat
losses
primary energy
conversion losses
building
heat
losses
heat
f low
primary
energy
conversion
losses
September 5
th
to September 7
th

42

temperature thermal input flows and therefore
represents the prerequisite for an efficient generation
subsystem.
The last subsystem possessing enhancement potential
is the consumer. Since the target temperature
determines the exergetic quality of the thermal
demand, therein lays no significant optimization
potential. However, as decreasing the amount of
energy that has to be supplied is also part of the low-ex
concept insulation can help to improve the system. On
the other hand, benefits similar to those already
discussed for the distribution subsystem can be
identified for the consumer system as well. By choosing
appropriate heating and cooling technologies, as e.g.
investigated in [3], the exergy destruction during heat
transfer to the room air can be minimized. This is
achieved by applying low-temperature heating and
high-temperature cooling devices. Inlet and outlet
temperatures of the heating/cooling device
simultaneously define constraints for the distribution
network subsystem, which in turn set constraints for the
generation. In the end supply temperatures close to the
target temperature form the basis for a low-ex ready
consumer. Without this step an exergetically optimal
energy supply system would be greatly hindered.
2. Applicable technologies for the realisation
2.1. Phase Change Slurries
The most used heat transfer fluid in district heating and
cooling networks is water. In supply networks, the heat
is transferred as sensible heat with a temperature
difference between forward and backward flow. The
heat transfer capacity of a network is determined by the
temperature difference, the mass flow and the heat
capacity of the heat transfer fluid. The temperature
difference and the temperature level of the network are
limited by technical restrictions and determine the
necessary mass flow of the heat transfer fluid. To
overcome these restrictions, fluids with higher heat
capacities than the heat capacity of water are under
development. An alternative to water could be PCS.
PCS are mixtures of dispersed phase change material
and a continuous liquid phase, which possess an
increased heat capacity due to the additional latent
heat of fusion occurring during the phase transition of
the phase change material. The PCS remains
pumpable even when the phase change material is
frozen. Thus, the PCS can be used as heat transfer
fluid in supply networks. A promising PCS for heat or
cold supply networks is paraffin/water dispersion.
Figure 3 is a photograph of a paraffin/water dispersion.
Paraffin is the phase change material, which can be
chosen according to the desired temperature of the
phase transition, and water is the continuous phase of
the dispersion. In [4] paraffin/water dispersions are
investigated and their properties presented

Figure3. Photograph of a paraffin/water dispersion

The increase of the heat transport capacity of a supply
network using a PCS instead of water can be described
by a thermal capacity enhancement factor (TCEF),
which is calculated according to equation (2).

( ) ( ) ( )
T c
T c w T c h w
TCEF
w p w
w p PCM p PCM f PCS
A
A A A
p
+ + p
=
,
, , ,
1
(2)

The TCEF is a function of the densities of the PCS
PCS

and water
w
, the mass concentration of the PCM w,
the specific heat capacity of PCM c
p,PCM
and water c
p,w

the heat of fusion of the PCM h
f,PCM
and the
temperature change T of the fluids. The TCEF is
calculated and plotted in the diagram figure 4 for
temperature differences T between the forward and
backward flow of 10 and 15 K as function of the mass
concentration w.
1
1.5
2
2.5
3
3.5
0 0.2 0.4 0.6 0.8 1
w [-]
TCEF [-]
delta T = 10 K
delta T = 15 K

Figure 4. TCEF PCS compared to water for temperature
differences 10 and 15 K, diagram calculated with the
properties of water and RT-42 of the company
Rubitherm [5]
Using PCS with a mass concentration w of 0.4 would
increase the heat transport capacity of the supply
network to 1.5 times of the value compared to water, if
September 5
th
to September 7
th

43

the temperature difference of the supply network is
15 K, and even 2 times, if the temperature difference is
only 10 K. Furthermore, the diagram shows that with
rising temperature difference the gradient of the TCEF
is lower, which means that the advantage of the PCS
compared to water disappear at higher temperature
differences. At the point where the gradient of the
TCEF is 0, the water system and the PCS system have
the same transport capacity. At that point, the mass
concentration of paraffin w has no influence on the
TCEF.
The use of PCS in energy systems leads to an
improved energy transport capacity, which results in a
reduction of the necessary temperature difference or
volumetric flow rate of the transfer fluid needed to
transfer a given amount of heat.
Another technical issue of PCS systems is the
increased pressure drop in the pipes due to the higher
viscosity of the PCS. A calculation methods and
measurement data can be found in [6, 7 and 8]. The
viscosity of PCS is related to several influence
quantities and can cause an incensement of the
pressure drop up 100%. PCS are non-newtonian fluids.
2.2. Capillary Tube Mats
The most often used heat exchanger type in heating
systems is a convective radiator, which is installed in
rooms close to the window. The size of a radiator
should be small, so that also the heat exchange
surface is small and the heating system must be
operated on a high temperature level to ensure the
heat transfer from the heating system into the room. An
alternative to convective radiators are floor heating
systems. Floor heating systems consist of a capillary
tube mat, which is installed in the upper layer of the
floor. Because of the bigger heat exchange surface
compared to the convective radiator, the temperature
level of the heating system is lower. A new approach to
realise heating and cooling of buildings is via CTM,
which are integrated in the floors of the building, as well
as in the walls and ceilings. This system offers a big
heat exchange area and allows the heating and the
passive cooling of the building. Due to the increased
heat exchanger area, a low temperature difference
between the heating system and room is possible. For
the further discussion, the following simple model is
used to describe the heat release of the heating system
in the building. The heating release system is
evaluated by the number of transfer units (NTU). The
heat capacity provided by the heating network Q
is
calculated by equation (3) with the inlet and outlet
temperature T
in/out
of the supply network, the mass flow
m and heat capacity c
p
of the heat transfer fluid.
( )
out in p
T T c m Q =
(3)

In view of the heat release in the room, the heat
capacity Q
can also be described by equation (4) and

is related to the heat transfer coefficient U, the heat
exchange area A and the temperature difference
between the mean temperature of the heat release T
m

as well as the room temperature T
r
.
( )
r m
T T A U Q =
(4)

The mean temperature of the heat release T
m
is
calculated by equation (5).
out
in
out in
m
T
T
T T
T
ln
= (5)

Based on the equations (3) to (5), it is possible to
calculate the NTU, which characterizes the heat
release in the room, according to equation (6), which is
only a function of the inlet and outlet temperature T
in/out

of the heat supply, the mean temperature T
m
of the
heat release and the room temperature T
r
.
( )
( )
r m
out in
p
T T
T T
c m
A U
NTU
(6)

The NTU values have been calculated for a convective
radiator system and a CTM system. The assumed
temperatures for the calculation and the results are
given in table I.
Table I. NTU for both heat release systems:
conventional radiator and CTM
parameter convective radiator CTM system
Tin [C] 80 37
Tout [C] 60 31
Tr [C] 20 20
NTU [-] 0.4 0.43

The NTU value of the CTM system is 0.43 and as high
as the NTU value of the convective radiator. This
means that both systems have the same heat release
capacity, although the inlet temperature T
in
of the CTM
system is lower and the temperature difference
between inlet T
in
and outlet T
out
of the CTM system is
smaller.
CONCLUSION
From the point of view of the low-ex concept the major
task en route to an exergetically efficient energy supply
system is the replacement of the combustible fuel
boiler by utilization of low temperature thermal input
September 5
th
to September 7
th

44

flows such as industrial waste heat or geothermal
energy. To achieve this goal it is necessary to
decrease the medium temperatures within the
distribution networks first. A prerequisite is a low-ex
ready consumer that allows meeting the thermal
demands applying low temperatures.
A possible realisation employs CTM in the heating
system of the building that allows applying inlet and
outlet temperatures of approximately 37 C and 31 C,
respectively. Within the district heating or cooling
network, the utilization of PCS instead of pure water
enables the application of small temperature
differences between forward and backward flow while
retaining the pipe dimensions. Since the backward flow
temperature mainly depends on the outlet temperature
of the consumer system, small temperature differences
within the network automatically lead to low forward
flow temperatures. Consequently, the exploitation of
low temperature heat sources as input flows for the
energy supply system is rendered possible.
Moreover, the decreasing temperatures in both forward
and backward flows of the network reduce the
transportation heat losses. This leads in the end to a
reduction of energy input (quantitative aspect of the
low-ex concept) into the supply system.
The only drawback suffered occurs in terms of an
increased pumping effort caused by a higher viscosity
of the PCS in comparison with water. But, since heat
losses are the predominant factor over circulation
pump energy, an overall benefit should be
accomplishable.
Summarizing it should be pointed out that applying
technologies such as CTM in the building heating or
cooling system and PCS as alternate heat transfer
medium for the distribution networks the low-ex
concept can be realised, thus greatly enhancing the
efficiency of energy supply systems.
ACKNOWLEDGEMENT
This study was supported by the Project Management
Juelich (PTJ) and the Federal Ministry of Economics
and Technology (BMWi) under 0327471A.
Comments of a highly constructive nature were
received from Daniel Wolf, Jorrit Wronski and Astrid
Pohlig.
REFERENCES
[1] C. Kemal et al., Evaluation of energy and exergy
losses in district heating network, Applied Thermal
Engineering, 24 (2004), pp. 1009-1017.
[2] H. Inaba, New challenge in advanced thermal
energy transportation using functionally thermal
fluids, International Journal of Thermal Sciences,
39 (2000), pp. 991-1003.
[3] M. Ala-Juusela et al., LowExergy Systems for
Heating and Cooling of Buildings, final report of the
IEA ECBCS Annex 37.
[4] L. Huang et al., Evaluation of paraffin/water
emulsion as a phase change slurry for cooling
applications, Energy, 34 (2009), pp. 1145-1155.
[5] Rubitherm RT-42, datasheet 08/20/2009,
http://www.rubitherm.de, Rubitherm Technologies
GmbH, Berlin (2010).
[6] Yinping Zhang, et al., Experimental research on
laminar flow performance of phase change
emulsion, Applied Thermal Engineering, 26 (2006),
pp. 1238-1245.
[7] A., B. Metzner et al., Flow of Non-Newtonian Fluids
Correlation of the Laminar, Transition, and
Turbulent-flow Regions, American Institute of
Chemical Engineers Journal, Vol. 1, No. 4 (1955),
pp. 434-440.
[8] R. Rautenbach, Kennzeichnung nicht-Newtonscher
Flssigkeiten durch zwei Stoffkonstanten, Chemie-
Ingenieur-Technik, 36 No. 3 (1964), pp. 277-282.

September 5
th
to September 7
th

45

APPLICATION OF EXERGOECONOMICS TO THE OPTIMIZATION OF BUILDING
HEATING SYSTEMS CONNECTED TO DISTRICT HEATING NETWORKS
C. W. Snoek and S. C. Kluiters
Renewables and Integrated Energy Systems, CanmetENERGY, Natural Resources Canada,
1 Haanel Dr, Ottawa, K1A 1M1, Canada

ABSTRACT
The concept of energy efficiency, defined as useful
energy output as fraction of required energy input, has
been used for years in technical systems assessments.
In addition to energy efficiency, there are benefits to
using exergy efficiency to assess system performance.
Whether systems will be installed or not is ultimately
determined by their economic performance. This
performance is usually determined by comparing initial
investment cost and operational cost with revenues
throughout a systems lifetime in terms of payback time
or net present value.
This paper describes a novel methodology that uses
the concept of exergy and the thermoeconomic factor,
a ratio that compares investment-related cost and
exergy destruction cost, for the economic optimization
of a community energy system. It compares the cost of
exergy and the required capital and operational costs
including carbon taxes to accommodate this low quality
energy. In doing so it enables a quick way to properly
assess the value of a systems ability to use low exergy
energy inputs. The method is compared to a more
traditional economic analysis.
INTRODUCTION
In the last few years, we have become painfully aware
of the effects of climate change. The burning of fossil
fuels and the resulting emissions are thought to be a
major contributor to the apparent increase of adverse
weather events. While people need energy for comfort,
in some cases there may be a choice in the source and
nature of that energy. In addition to climate change,
there is also a concern about the rapid depletion of the
more valuable of fossil fuels, natural gas and oil. For
these reasons it makes much sense to re-evaluate the
sources of the energy we use and the effect of using
them has on the environment.
To lower energy requirements, energy efficiency has
been practiced for many years. In terms of comfort
heating in houses, most of the effort has gone into
improving building insulation, better windows, building
orientation with respect to the sun, shading from solar
energy etc. In terms of energy conversion equipment,
improving the efficiency often meets natural limits,
such as those expressed by Carnots Law.
Often, omitted from consideration is the quality of the
energy that is needed to provide comfort to the
occupants of a building. While the heating
requirements of a building can be determined (in GJ or
TJ), the nature or origin of this energy is not addressed
in energy efficiency calculations. The total amount of
Joules can be provided by oil, natural gas, electricity or
low temperature waste heat. While the first three
energy sources are considered high quality, and can be
used to generate very high temperatures (over
1000 C), run equipment such as computers, radio and
TV transmitters and receivers, waste heat is of low
quality and has no other use. Comfort heating does not
require high temperatures and therefore using high
quality fuel for low quality applications is considered
wasteful.
Energy quality is often expressed as exergy. Exergy is
defined as the maximum useful work possible during a
process that brings the system into equilibrium with a
heat reservoir. To illustrate the concept of exergy one
can compare two different forms of the same amount of
energy: 100 kJ of energy is equivalent to:
12 V/2.3 Ah stored in a car battery, or
1 kg of water at 43 C in a room with an ambient
temperature of 20 C.
Obviously, the energy contained in the battery is
considered more useful and therefore has the higher
quality or exergy.
The ratio of Exergy (E) to Energy (Q) can be expressed
as:
ply
ambient
T
T
Q
E
sup
1 = (1)
where T is given in K.
Equation 1 shows that when the supply temperature of
an energy source is high, the exergy converges to the
energy value. Electricity and mechanical work are
(nearly) perfectly convertible and the exergy content is
therefore equal to the energy content. Conversely,
when the supply temperature is closer to the
environmental temperature, the value of the exergy
becomes (much) smaller than that of the energy.
Wall [1], in his paper on Exergy and Morals quotes
Alfven who claimed that energy accounting based on
energy only is like a bank teller counting by the amount
September 5
th
to September 7
th

46

of coins and bills, but neglecting their value. In an
ethical society the value, worth and quality of different
energy supplies should, as a minimum, be matched to
the requirements of the different energy applications.
Methods to design low exergy buildings are available
today. For instance, Schmidt [2] developed a method
and pre-design tool for low exergy buildings in which he
compared different heating systems, such as boilers,
condensing boilers, electric heating, GSHP and low
temperature under-floor heating. However, this method
does not directly address the effect of system heat
transfer surface area on the overall economics.
Also, there is an additional benefit realizing that a
building that can accommodate low exergy streams is
ready for future hook-up to other, perhaps renewable
energy sources: GSHP, solar, waste heat from
industry, energy from thermal storage to name a few.
This is a distinct advantage when the move to a
sustainable society gains momentum, and the concept
of low-temperature heating should be incorporated in
building codes.
This paper considers the cost of using the low quality
part of the energy source and the (increased) capital
cost and operating cost that are required to
accommodate low quality energy. A methodology has
been developed to determine the optimal cost of
operation, based on the capital cost, operational cost
and the cost of the exergy.
This type of analysis is considered part of the field of
thermoeconomics, more in particular exergoeconomics.
Wikipedia defines thermoeconomics in a very
theoretical way as a school of economics that applies
laws of thermodynamics to economy. Valero et al. [3]
operationalize this definition by describing two aims of
thermoeconomics, (1) optimization to minimize cost of
a system or component, and (2) cost allocation of
individual outputs of a plant producing a number of
outputs.
Valero and coworkers [3] date this research field back
as far as 1932, when Keenan apportioned cost of heat
and work taking into account irreversibility and
thermodynamic efficiency instead of enthalpy only.
However, they go on to say that Gaggioli, and Tribus
and Evans in the early 1960s started off real
development in thermoeconomics. Ever since, these
fields have received tremendous attention. Valero and
coworkers identify that an important problem in this
body of research is the variety of methodologies used
with accompanying nomenclature. Between them and
Tsatsaronis [4] they already name a fair amount of
methods. In doing so, Tsatsaronis introduces the
exergoeconomic factor f, as a fraction that compares
two sources contributing to cost increases, investment-
related cost and exergy destruction cost. This
exergoeconomic factor is also found in other sources,
such as Temir & Bilge [5].
It is beyond the scope of this paper to provide a
comprehensive literature overview of thermoeconomic
publications or even of the methods used in these
publications. The aim of this paper is to apply one of
these methods, using the above-mentioned
exergoeconomic factor to optimize building heating
systems connected to a district heating system. To the
best of the authors knowledge, so far this method has
only been applied to optimize individual components.
This work ties in with research into advanced low-
temperature district energy systems currently carried
out at the CanmetENERGY laboratories of Natural
Resources Canada in Ottawa, Canada.
The system considered consists of buildings with their
heating system (radiators and cross-flow heat
exchangers are considered), the energy centre with
boilers and pumps and the pipeline to move the energy
in the form of hot water to the community. The
development of the methodology was the main object
of the study, not the optimization itself.
While the development of the optimization was related
to economics, in other words, the least costly option, it
should be noted that the concept of exergy opens up
the notion of morals and ethics. For new
developments, the costs of resource depletion and
environmental destruction should be considered as
well. Just because a certain system is economic, it is
not necessarily the best moral or ethical choice. Just
because a certain system does not cause local
problems, that does not mean that (environmental or
other) problems caused by this system elsewhere can
be ignored.
Traditional Optimizations
System optimization is often done by optimizing
systems separately, and not by considering the overall
efficiency of integrated systems. Often, an integrated
approach leads to optimal solutions, as in electricity
generation using a back pressure steam turbine.
Accepting a lower efficiency of the turbine may lead to
the residual energy in the condenser being useful in
other applications, whereas in the separately optimized
version this thermal energy would be useless. In the
latter case, the turbine back pressure is kept as low as
possible, to extract the maximum electrical power. This
makes the condensate of too low a temperature to be
useful in other applications. Optimizing integrated
systems as a whole avoids this problem.
Exergoeconomic Optimization
In an exergoeconomic optimization, the concept of
exergy is used to determine the best and most
economic solution to an energy conversion process or
September 5
th
to September 7
th

47

system. While the total quantity of transferred energy
remains the same, the exergy that delivers this energy
may vary. Analyzing the required exergy with respect to
the energy transfer equipment will result in an optimum
economic solution allowing for integration of the system
with other systems.
In this exergoeconomic optimization, the system is
treated as an integrated whole together with other
systems. While the (comfort) energy supplied remains
the same in any given scenario, the exergy required for
this scenario varies and the cost implications of this
variation are included in the analysis. Therefore, in this
analysis the consumer of energy does not pay for the
energy, but for the exergy, the real value of the energy
supplied.
SYSTEM DESCRIPTION
The system considered here to develop the
methodology is a district heating system supplying hot
water for space heating to a 1000 home community in
the Ottawa area in Canada. It was modelled using the
RETScreen clean energy project analysis software tool
[6] and in-house spreadsheet based models. The hot
water is transported from an energy centre located
centrally in the community to the 1000 detached
homes. Inside the buildings, radiators or cross-flow
heat exchangers (water-to-air fan coils) are employed
to provide space heating.
Energy Supply
The building temperature set point is kept constant at
20C. The hot water supply temperature is determined
by the outdoor temperature. If the outdoor temperature
is above 5 C, the supply temperature is 70 C. When
the outdoor temperature drops below -15 C, the
supply temperature equals 90 C. Between 5 C and
-15 C, the supply increases linearly from 70 C to
90 C. This is a common supply temperature profile
used in many European district heating systems. It
prevents excessive flows in the pipes at high loads and
permits smaller heat transfer surfaces in the buildings
due to the higher temperature difference between
water and building air. When the heat transfer surface
area was varied to reach an optimum solution the
supply temperature was adjusted by a constant value
over the entire load range. The water return
temperature was set at 30 C in all design calculations,
but varied throughout the year according to the off-
design characteristics of the heating equipment used.
The load of the buildings is related to the outdoor
temperature. The annual heat consumption was set at
100 GJ per house, a typical value for detached homes
in this area. The instantaneous load throughout the
year is simply calculated as a linear relationship
between zero and the maximum capacity, when the
outdoor temperature varies between 20 C and -28 C,
the design temperature for Ottawa. While this is an
over-simplification of reality, it neither hinders the
development of the methodology nor introduces
serious errors of consequence.
Energy Transmission
The pipe diameters were estimated using the
RETScreen software tool. This means for diameters
under 400 mm the pressure drop is kept below 200 Pa
per meter of pipe and for larger diameters flow velocity
is maximized at 3 m/s [6]. As RETScreen has a limit of
13 sections for district heating systems, the 1000
homes were assumed to be located along twelve 80-
home streets. The final 40 homes were located in a
separate street.
The energy transfer fluid is water. The pipes are
preinsulated steel or cross-linked polyethylene (PEX)
pipes. The first iteration of the methodology accounted
for heat losses from the pipes. Since it was found that
this heat loss was a negligible fraction of the
transmitted energy, it was omitted in subsequent
versions. For a thorough analysis, it is recommended to
include heat losses, especially if the piping system is
extensive and the supply temperatures reach high
levels.
A pressure drop analysis was used to determine the
required pump energy. The electric motor driving the
pump was estimated to have 90% efficiency while the
pump was assigned an efficiency of 85%.
End-Use
The energy supplied to the pipeline was used to keep
the building temperatures at set point. Therefore,
regardless of the (size of) building heating system
used, the same energy was used to keep the buildings
warm. However, the exergy used was dependent of the
system in place and of its size. The larger the size, the
lower the required water temperature and hence less
exergy was required to achieve the same end result.
Two different technologies were used to model the
transfer of energy into the building space: cross-flow
heat exchangers, and radiators. Simulations were done
for both technologies separately, and the technologies
were never mixed. This was done to simplify the
analysis. In reality, mixed systems will occur and
should be analysed as such. While this will increase
the level of modelling complexity, it is not difficult to do.
DESCRIPTION OF MODELING
Climate
The local climate has a significant effect on the design
of a building heating system. A maritime climate may
have many degree days but not show the variability in
demand that a building in a continental climate with an
equal amount of degree days experiences. Even if both
September 5
th
to September 7
th

48

buildings use the same amount of energy per year, the
demand load in the building with the continental climate
may be far greater. Therefore, the climate plays an
important role in the design of a heating system and
should, therefore, be considered in this analysis.
Supply Temperature
To include the effect of the variability of the energy
demand with time-of-day and the seasons, the
statistical average hourly temperatures for the city of
Ottawa were used. These temperature values are real
values, with realistic variability (high and low
temperatures), using time-periods from different years
to provide for a correct average. In total, 8760 hourly
values of temperature were used in the spreadsheet,
as shown in Figure 1.
-40
-30
-20
-10
0
10
20
30
40
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Hour (-)
T
e
m
p
e
r
a
t
u
r
e

(
C
)
Fig.1. Average hourly temperatures in Ottawa
Solar Radiation
To simplify the spreadsheet calculations, the effects of
solar radiation, plug loads and occupancy gains were
neglected. When performing the optimization, these
effects remain constant and so have little effect on the
final outcome. When using this method for design,
these contributors to the exergy balance should be
considered.
Determination of Instantaneous Load
The maximum thermal load of the community for space
heating was determined using the average hourly
temperatures and the annual heat consumption per
house of 100 GJ. It turned out to be 10,640 kW. When
the outdoor temperature reaches -28 C, the Ottawa
design temperature, the community requires the
maximum thermal load. At the ambient temperature of
20 C, the load is nil. The modelling is set up so that
between these ambient temperatures, the load varies
linearly. For instance, at -4 C, the load equals 5.32
MW.
Pumping Power and Exergy
To meet the load, the water had to be pumped from the
supply source to the load. The amount of water
pumped varied with the load and the supply
temperature. The pumping energy required was
included in the modeling. Since electrical energy is
equivalent to exergy, the pumping energy calculated
from the pressure drop calculations (including
efficiencies), was numerically counted as exergy.
Normally, during periods of no-load, the pumps keep
operating to keep a supply of design temperature water
close to the load. This is done with a thermostatically
operated by-pass valve. Since this valve represents a
constant effect which does not affect the optimization, it
was not modelled for simplicity.
Design of Cross-Flow Heat Exchanger
The design of the cross-flow (or fan-coil) heat
exchanger was based on the assumption that the
overall heat transfer coefficient U was 25 W/(m2K).
The F-factor was set at 0.94. To meet the design load,
heat exchangers with a combined area of 17,152 m2
were required.
Design of Radiator Heating System
The design of the radiators was done in a very simple
manner. It is acknowledged that better methods exist,
but the development of the methodology did not suffer
because of this simplification. For any optimization,
actual modelling of the equipment should take place.
To determine the heat transfer from the panels, the
general radiation equation
( )
4 4
room panel
T T Q =oc (2)
was used with the average panel temperature. The
surface emissivity was estimated at 0.9. Convection
from the surfaces was not separately considered.
To meet the design load, 42,271 m2 of radiative
surface was required to meet the design load.
Determination of Exergy Use
As indicated in Section 1, the ratio of Energy to Exergy
can be expressed as:
ply
ambient
T
T
Q
E
sup
1 = (1)
Where E is exergy, Q is energy and T is the
temperature given in K.
For this study, knowing the energy supplied, Equation
(1) was used to calculate the supplied exergy for each
hour interval. For each of the heating systems used
and each of their variations is size, the amount of
energy supplied to the heated space remained the
same. However, due to the supply temperature
September 5
th
to September 7
th

49

requirements, and the different flow requirements, the
exergy used by each system was unique.
Capital and Exergy Cost Assessments
All cost numbers reported in this paper are in 2009
Canadian Dollars. An in-house costing tool was used to
estimate cost for the district heating energy centre,
containing the pumps and boilers, and the buried
distribution piping. For the two heating systems
considered, the costs were assigned as shown in
Table I. For water-to-air fan coils (cross-flow heat
exchangers) an installed cost of $250/m2 was
considered representative, and for radiators $200/m2
was selected as a typical value.
Table I. Cost for heating technologies
Cross-flow heat
exchanger
$250/m
2
of heat transfer surface
Radiative system $200/m
2
of exposed panel

Future cash flows were discounted at a rate of 8% and
system lifetime was set at 40 years. Annual operating
and maintenance (O&M) cost other than cost for heat
and electricity were set at a fixed fraction of 1% of total
investment cost.
To compare traditional optimization with exergo-
economic optimization three types of analyses were
performed. The classical analysis applies the
traditional optimization where heat is valued based on
energy content, at a rate of $5/GJ, which is considered
representative for heat from natural gas combustion.
Electricity cost has been set at $17/GJ (just over
$60/MWh).
In the exergoeconomic analysis heat and electricity are
priced based on the exergy content. The exergy charge
was determined at $30/GJ for thermal energy, based
on the above mentioned $5/GJ for heat, assuming a 1
to 6 ratio of exergy to energy content (applies to a
temperature around 80 C). The electrical energy to
exergy ratio was taken as one, resulting in an exergy
charge of $17/GJ for electricity. At first glace it may
seem erroneous to charge more for exergy from the
thermal source than that for the electricity for the pump,
but it must be remembered that the (thermal) exergy is
a fraction of the thermal energy.
The third type of analysis is a classical analysis
corrected for the difference in value of low- and high-
temperature heat, by assuming energy under 60C is
available free of charge (as waste heat from a nearby
process). For energy over 60 C the charge is still
$5/GJ.
To assess the influence of carbon taxes, two sets of
results are presented. One assumes no carbon taxes
are in place and the other assumes a carbon tax of $30
per ton CO2eq. Carbon intensity factors of 0.050 ton
CO2eq/GJ were used for natural gas and 0.054 ton
CO2eq/GJ for electricity (taken from RETScreen [6] as
representative for Canada). This results in a $6.5/GJ
energy charge for heat, a $38.9/GJ exergy charge for
heat and an $18.6/GJ energy (or exergy) charge for
electricity.
Thermoeconomic Factor
The exergoeconomic or thermoeconomic factor f
compares two sources contributing to cost, investment-
related cost and exergy destruction cost. It is defined
here as the ratio of Capital Cost Rate (CCR, which
includes O&M cost, but excludes heat and electricity
cost) and the sum of Exergy Destruction Cost Rate
(EDCR) and CCR. The CCR equals the cost per unit
time for the installation, depreciation, maintenance, etc,
while EDCR is the cost of exergy.
CCR EDCR
CCR
f
+
= (4)
Since CCR and EDCR have the dimensions of $/time,
f is dimensionless.
A high value for f indicates that the capital and
maintenance costs are dominant. Also, a high f value
indicates good use of the exergy in the fuel. On the
other hand, a low value for f indicates an inefficient
use of fuel resources. For each heating system
variation, the average annual thermoeconomic factor
was calculated.
MODELLING RESULTS
Base case design
Table II shows the main information for the base case
designs for both the radiator and cross-flow heat
exchanger systems. As expected, the distribution pipe
diameters, required pump capacity, annual space heat
consumption and annual heat cost are the same for
both systems.
As the water return temperatures throughout the year
are generally lower for the radiator system, the required
water flows and consequently the annual electricity
consumption are lower for the radiator system. As both
systems have a design supply temperature of 90 C
(and thus also the same off-design supply
temperatures throughout the year), the annual exergy
consumption is the same for both. The lower return
temperatures for the radiator system also show in the
higher fraction of energy provided under 60 C. In
terms of cost, the radiators are clearly more expensive
resulting in higher annual investment and O&M cost,
which is not offset by the somewhat lower electricity
cost. Overall the more capital intensive radiator system
has a higher f-factor than the cross-flow heat
September 5
th
to September 7
th

50

exchanger system. In a comparison between the two
systems, the cross-flow heat exchanger system works
out cheaper using all three types of analysis due to the
large difference in investment cost.

Table II. Main information base case designs without
CO2 tax.
Radiator
Cross-flow
heat
exchanger
Surface area (m
2
) 42,271 17,152
Distribution pipe
diameters (mm) DN80/DN65 DN80/DN65
Required pump capacity
(kW) 32.2 32.2
Annual electricity
consumption (GJ) 115.2 147.6
Annual exergy
consumption (GJ) 16,810 16,810
Annual space heat
consumption (GJ) 100,000 100,000
Fraction of energy < 60 C 68.4% 64.4%
Installed cost heaters $8,454,298 $4,288,066
Investment cost district
heating system $11,556,386 $11,556,386
Annual O&M cost $200,107 $158,445
Annual charge investment
and O&M cost $1,878,206 $1,487,163
Annual heat (energy) cost $500,000 $500,000
Annual heat (exergy) cost $504,193 $504,193
Annual electricity cost $1,958 $2,509
Total annual cost classical
analysis $2,380,164 $1,989,672
Total annual cost
exergoeconomic analysis $2,384,357 $1,993,865
Total annual cost heat
under 60 C free analysis $2,038,164 $1,667,672
f-factor 0.814 0.782
Alternative designs radiator system
For both the radiator and the cross-flow heat
exchanger system alternative designs with increased
and decreased surface areas were costed. The district
heat supply temperatures were modified accordingly,
and as noted before, the required water flows and thus
distribution pipe diameters and pumping power
requirements were modified too. The effects of these
variations on cost were taken into account.
The results of all the modelling runs are shown in the
figures below in the form of the relationship between
the annual cost (the sum of capital investment, O&M
cost and energy or exergy costs) and the f-factor. An
increasing f-factor means increasing surface areas
(and thus increasing capital and operating and
maintenance cost) and decreasing heat supply
temperatures (and thus decreasing exergy cost).
Figure 2 shows results for the radiator based heating
system with no carbon taxes in place. The slight jump
in annual cost around an f-factor of 0.81-0.82 is caused
by an increase in district heating piping diameter from
DN65 to DN80 for the 80-house streets and from DN50
to DN65 for the 40-house. All lower f-factors shown
have piping diameters of DN65 and DN50 and all
higher f-factors shown have DN80 and DN65
respectively.
$1,500,000
$1,700,000
$1,900,000
$2,100,000
$2,300,000
$2,500,000
$2,700,000
$2,900,000
0.68 0.70 0.72 0.74 0.76 0.78 0.80 0.82 0.84 0.86 0.88
f-factor (-)
A
n
n
u
a
l

c
o
s
t

(
$
)
Classical analysis Exergo-economic analysis Heat under 60C free analysis

Fig. 2. Relation between f-factor and annual cost radiator
system, no carbon tax.
The classical analysis shows a continuous increase in
annual cost with increasing f-factor.
1
This makes sense
because cost is not based on exergy but on energy.
Therefore, an increasing surface area means
increasing capital cost, but constant energy cost, so the
lower exergy requirement does not offset the increase
in capital cost. The classical analysis would tell us to
optimize the system with minimum capital expenses. In
reality there would be a limit as ever increasing
temperatures will mean that we are dealing with more
expensive materials and at a certain stage steam
instead of hot water, requiring a more expensive district
heating system. Also, heat losses to the environment

1
Although exergy is not explicitly costed in the classical analysis,
we can still calculate an f-factor.
September 5
th
to September 7
th

51

will increase with increasing supply temperatures,
which is not modelled here.
The exergoeconomic analysis does take into account
the increasing exergy requirements for systems with
lower surface areas. Consequently, there is a minimum
annual cost (for an f-factor in the range 0.710.73).
Going to lower f-factors, exergy cost significantly
increase, which results in increasing annual overall
cost. Going to higher f-factors, the annual cost increase
again because the decreasing exergy cost are more
than offset by the increase in capital and O&M cost.
The analysis discounting heat under 60 C does take
the temperature level of energy supplied into account
while determining costs, though there is no explicit
price for exergy in the calculations. As a result, the line
does not slope up as strongly with increasing f-factor
as the line pertaining to the classical analysis. Going to
higher f-factors, eventually all heat will be delivered
under 60 C, and all heat provided will be free. Going to
lower f-factors, eventually all heat will be supplied at
temperatures over 60 C and the green line will
coincide with the blue classical analysis line. The heat
under 60 C free analysis does not show an optimum
and would suggest minimizing the f-factor. Like the
classical analysis the economic analysis suggests that
capital cost are dominant.
Figure 3 shows the effect of introducing a carbon tax of
$30/tonCO2eq on the exergoeconomic analysis. It is
clear that the carbon tax leads to higher annual cost
and lower f-factors for the same systems, both caused
by the increased exergy cost. Both lines show an
optimum for an f-factor in the range 0.710.73, but for
the case without carbon tax the corresponding surface
area is lower than for the case with carbon tax. This
makes sense as increasing heat and exergy cost mean
a shift to a system with higher surface areas and lower
heat and exergy requirements. As figure 3 shows, the
capital and O&M cost as a fraction of total cost (and
consequently also the exergy cost as a fraction of total
cost) remain in the same range.

$2,200,000
$2,250,000
$2,300,000
$2,350,000
$2,400,000
$2,450,000
$2,500,000
$2,550,000
$2,600,000
$2,650,000
$2,700,000
$2,750,000
0.63 0.65 0.67 0.69 0.71 0.73 0.75 0.77 0.79 0.81 0.83 0.85 0.87
f-factor (-)
A
n
n
u
a
l

c
o
s
t

(
$
)
No carbon tax Carbon tax $30/tCO2

Fig. 3. Relation between f-factor and annual cost radiator
system, with and without carbon tax.
Alternative designs cross-flow heat exchanger
system
Figure 4 shows the results for the system with cross-
flow heat exchangers. Note again that the jump in
annual cost at an f-factor around 0.78 is due to the
increase in district heating pipe diameter. As for the
radiator system, the classical analysis shows a steep
slope with increasing f-factors as capital cost are
dominant and lower exergy requirements do not
translate into cost savings. Again the classical analysis
would lead us to minimize the surface area (with the
same limitations as applied to the radiator).

$1,400,000
$1,500,000
$1,600,000
$1,700,000
$1,800,000
$1,900,000
$2,000,000
$2,100,000
$2,200,000
0.65 0.67 0.69 0.71 0.73 0.75 0.77 0.79 0.81 0.83
f-factor (-)
A
n
n
u
a
l

c
o
s
t

(
$
)
Classical analysis Exergo-economic analysis Heat under 60C free analysis

Fig. 4. Relation between f-factor and annual cost cross-
flow heat exchanger system, no carbon tax.

The exergoeconomic analysis shows a downward
sloping line. This is caused by the reduced capital cost
and O&M cost compared to the radiator system and
thus increased importance of exergy cost as fraction of
the total cost. An increase in cost due to surface area is
more than offset by a decrease in exergy cost.
Contrary to the radiator system, though, the
exergoeconomic analysis does not show a clear
optimum, although it clearly levels off at higher
f-factors. It is interesting to note here that the classical
analysis and the exergoeconomic analysis lead to
contradictory recommendations as to optimization.
The analysis with free heat under 60 C shows a line
gradually sloping up, though far less pronounced than
the classical analysis line. Like the classical analysis
line it would indicate that lower surface areas would
optimize this system.
Figure 5 shows the effect of a carbon tax on the
exergoeconomic analysis. As for the radiator system
the carbon tax means higher annual cost and lower
f-factors for the same system due to increased exergy
cost. As there is not a clear optimum in either line, we
can not conclude that the optimum f-factor is the same
for both. However, it is clear that both level off in the
higher f-factors range.
September 5
th
to September 7
th

52

$1,600,000
$1,700,000
$1,800,000
$1,900,000
$2,000,000
$2,100,000
$2,200,000
$2,300,000
$2,400,000
$2,500,000
0.60 0.62 0.64 0.66 0.68 0.70 0.72 0.74 0.76 0.78 0.80 0.82
f-factor (-)
A
n
n
u
a
l

c
o
s
t

(
$
)
No carbon tax Carbon tax $30/tCO2

Fig.5. Relation between f-factor and annual cost cross-
flow heat exchanger system, with and without carbon tax.

From the foregoing, it is clear that useful comparisons
can be made using this methodology. The results from
exergoeconomic analyses can significantly deviate
from those obtained with a classical analysis. Which of
the two is the more relevant one will depend on the
situation. For non-integrated systems, the classical
analysis may be the one to follow, but for integrated
energy systems, which are expected to become more
and more important, the temperature level of heat
becomes important, and the exergoeconomic analysis
seems more appropriate. Using the f-factor will help in
finding optimum solutions, especially for exergo-
economic analyses.
Variations in external factors, such as fuel costs or
Government / utility incentives could change the shape
of the curves to make the minimum more pronounced.
CONCLUSION AND SUGGESTIONS FOR FURTHER
WORK
From the results of testing the methodology of
exergoeconomic optimization using the f-factor, it is
clear that it is a useful tool to determine the effects of
different heating technologies and heat transfer surface
sizes of these technologies on the annual overall
operational costs. This is especially true if the heating
system is integrated with other energy systems. It is
also true if the temperature level of the heat is
important for another reason. The methodology can be
used to make informed choices regarding technologies
to be used for heating homes or buildings and
regarding the size of these technologies.
To continue this development work, it is recommended
that more practical considerations will be incorporated
into the models and analyses. Increasing temperatures
do not just cost more in terms of exergy but also in
more expensive materials, and steam based district
heating systems are considerably more expensive than
hot water based systems. Heat losses from the pipeline
were small but may need to be considered in a follow-
up study. Including passive heating of houses by solar
radiation, plug loads and occupancy gains will also
improve model predictions. Also mixed systems
combining heating technologies and possibly including
other technologies such as under-floor heating provide
further opportunities to optimize system cost.
In addition, the application of the methodology
developed in this study should be applied to a heat
pump, where the variations in COP with supply
temperature would be included. This would result in the
ability to match the heating equipment to the heat
pump, resulting in an optimum operation.
ACKNOWLEDGEMENT
During this work the authors have had very fruitful
conversations with many colleagues: Mikhail Sorin,
Evgueniy Entchev, Libing Yang, Ibrahim Dincer, Hajo
Ribberink and Kirby Wittich. These discussions helped
focus the work and stimulated further thinking in this
interesting area of science. This is to thank all those
who spent their valuable time listening and providing
valuable comments.
REFERENCES
[1] G. Wall, Exergy and Morals, in Second law
analysis of energy systems: towards the 21st
century, E. Sciubba, M.J. Moran Eds, Circus,
Roma (1995), ISBN 88-86662-0-9, pp. 21-29.
[2] D. Schmidt, Design of Low Exergy Buildings
Method and a Pre-Design Tool, in International
Journal of Low Exergy and Sustainable Buildings,
Vol. 3 (2003), pp. 120-126.
[3] A. Valero, L. Serra & J. Uche, Fundamentals of
Exergy Cost Accounting and Thermoeconomics.
Part I: Theory, in Journal of Energy Resources
Technology, Vol. 128 (2006), pp. 1-8.
[4] G. Tsatsaronis, Application of Thermoeconomics
to the Design and Synthesis of Energy Plants, in
Exergy, Energy System Analysis, and
Optimization, [ed. Christos A. Frongopoulos], in
Encyclopedia of Life Support Systems (EOLSS),
developed under auspices of the Unesco, Eolss
Publishers, Oxford, UK (2007).
[5] G. Temir, D. Bilge, Thermoeconomic analysis of a
trigeneration system in Applied Thermal
Engineering, Vol. 24 (2004), pp. 2689-2699.
[6] Clean Energy Project Analysis RETScreen
Engineering & Cases Textbook, 3
rd
edition,
RETScreen International, Natural Resources
Canada, Varennes (2005).
September 5
th
to September 7
th

53

SLIMNET: AN INNOVATIVE INTEGRAL APPROACH FOR IMPROVING
EFFICIENSIES OF DISTRICT HEATING NETWORKS
M. W. P. van Lier
Stadsverwarming Purmerend B.V., the Netherlands
m.v.lier@svpbv.nl

ABSTRACT
This paper describes the innovative integral approach
improving district heating network efficiency, SlimNet.
SlimNet consists of five phases which lead to annual
energy savings of about 227.000 GJ and almost 37.000
ton CO
2
savings for the city of Purmerend in 2015.
INTRODUCTION
Company situation
In 2007 the new company Stadsverwarming
Purmerend B.V. (SVP) took over the responsibilities of
the district heating network from the municipality in
Purmerend, the Netherlands. With 25.000 customers
the grid is the fourth largest grid of the Netherlands.
District heating Purmerend started in 1980. The
network expanded organically following the city
expansions. While daily operations were outsourced to
external and changing partners, the final responsibility
stayed with the municipality.
A comprehensive business analysis performed by the
new management in 2008 showed severe problems. In
the present state the company would remain
structurally loss giving, (future) heat delivery was not
ensured, and sustainability and customer satisfaction
were below benchmark standards. Fall 2009 a new
business plan was presented that sets course for a
future proof company, based on sustainable, cost-
effective and 80% renewable heat. On the technical
side this is achieved by two major project programs, a.
improving network efficiency, SlimNet, and b.
incorporation of sustainable energy sources, the
Energy transition. The company mission is to become
the most sustainable district heating company of the
Netherlands.
Network description
The 520 km district heating network is fed by a CHP
(CCGT) plant of 65 MWth and seven natural gas fired
auxiliary boilers with a total power of 131 MWth. During
the last 6 years 64% of the total heat production came
from the CHP plant. The heat sources are operated by
a third party.
The production units feed the heat to the network via
buffering tanks to the primary network. The heat is then
directly transported through substations and a
secondary network to the 25.000 customer
installations. Specific to the secondary network are the
post-insulated steel distribution pipes and connections
to customer installations hanging in narrow crawl
spaces under blocks of buildings.
In the distribution process no heat exchangers are
used except from the production of hot tapping water in
the houses.
Hydraulics are controlled by decentralized pressurizing
valves, differential pressure valves and pumps
compensating for hydraulic deficiencies.
The supply temperature from production is directly
related to the ambient temperature (i.g. 95 C at
T
a
=-10 C and 75 C when T
a
=15 C). The maximum
supply pressure to the primary network is 6,8 bars and
to the secondary network 4,5 bars.
NETWORK CONDITION
Part of the business analysis was an extensive
technical research program covering all technical
aspects of the grid and finally entire district heating
chain. The main conclusions were:
1. The network characteristic had become
uncontrollable: Network builds out has occurred
without a master plan. Effectively SVP had no
control on the characteristics of customer
installations. Furthermore, hydraulic problems in
the grid had been masked with decentralized
pumps and control systems.
2. Heat production capacity was critical, reaching a
critical limit under the conditions of the winter of
2008. There was certainly no spare capacity to
facilitate the planned expansion of the grid and
thus the heat demand as shown in Fig 1.

Fig. 1 Required heat production
September 5
th
to September 7
th

54

3. In 2008 the network showed a heat loss factor of
33,6% (with a Dutch benchmark of 25%). requiring
32.683 m3 of water replenishment in the same
year.
4. Parts of the network showed excessive heat loss
and repairs, mainly due to high ground water table,
exposing the pipes in crawl spaces directly to
water for most of the year. Repairs with standard
material proofed insufficient and innovation on
material and building techniques was needed.
SLIMNET
SlimNet is part of a large restructuring program initiated
in 2008. SlimNet does contribute to stopping the
negative spiral glide of the above mentioned problems
SlimNet consists of the following phases:
A. Knowing where the heat flows
B. Defining key performance indicators (KPI)
C. Developing analyzing tools
D. Developing and defining measures
E. Quantifying KPI results from SlimNet
In the following those phases will be discussed.
KNOWING WHERE THE HEAT FLOWS
For SVP the heat losses are defined as:
sold produced loss
Q Q Q = (1)
The heat losses in the network, Q
loss
, were 427.158 GJ
(33,6%) in 2008. Causes for those losses
5
are:
1. Losses in buffering tanks
2. Losses in primary network
3. Losses in secondary network
4. Undefined losses
None of the above can be determined exactly within
the boundary conditions of the network but the
following describes the results of the research
performed on this matter and the localization of
hotspots, parts of the grid with excessive losses.
Buffering tanks
In [1] an estimated calculation was made for the heat
losses due to the buffering tanks, 5.562 GJ annually.
There are four buffering tanks with a 4.000 m3 capacity
in the network which are used for peak shaving. A
check upon this calculation [2], based upon an IR-scan
of one of the buffering tanks resulted in an estimate of
14.032 GJ annually which is considered to be a
maximum value.

5
Losses from heat plants are not taken into account.
Primary network
Most substations in the network are provided with a
SCADA6 system. This data in combination with a newly
developed network model made it possible to
calculating annual heat loss at 100.706 GJ.
According to [3] about 14% of this heat loss is caused
by cross-linked polyethylene (PEX) piping material
used in the early 90s.

Fig. 2 IR scan of a PEX pipe constructed in 1990
Considering that those PEX pipes are applied in only
3,5% of the primary network, these may be referred to
as hotspots.
Secondary network
With four public housing companies, SVP conducted
research on failures in the district heating related
systems in Purmerend [3]. It became clear that during
the period 2006-2008 74% of the unplanned repairs
were caused by the high ground water level in the
crawl spaces where post-insulated steel pipes with
Armaflex insulation are installed. In total research
identified areas of 4000 houses, where heat loss was
extreme, i.e. hotspots.
This research confirmed the conclusion of an earlier
research [4] that the thermal conductivity k for the wet
insulation in the crawl spaces will be close to 0,1 W/mK
and 0,2 W/mK instead of the 0,02 or 0,03 W/mK for the
current pre-insulated pipes. The total of heat losses in
the secondary network are estimated at 304.041 GJ.
Conclusion addressing heat losses
Table 1 gives the overall results of the heat loss
analysis.
Table 1: Overall results of heat loss analysis
Main network part Loss(GJ) % of total
Buffering tanks 14.032 3,3 %
Primary network 100.706 23,6 %
Secondary network 304.041 71,2 %
Undefined losses 8.170 1,9 %
Total 427.158 100%

6
Supervisory Control And Data Acquisition
September 5
th
to September 7
th

55

It was concluded that replacing the PEX pipes in the
primary network and post-insulated pipes in the crawl
spaces of houses in the areas identified as hotspots
was the most effective strategy for heat loss reduction.
DEFINING KEY PERFORMANCE INDICATORS
The main goal of SlimNet is improving network
efficiency as part of the new business plan that sets
course for a future proof company which provides
sustainable, cost-effective and 80% renewable heat.
The Key Performance Indicators (KPIs) can be divided
in four main criterions:
1. Economics
2. Sustainability
3. Reliability
4. Customer Satisfaction
Economics
Every GJ of heat lost in the network cannot be sold and
has therefore a negative effect on the balance sheet.
Consequently the heat loss in the DH-network is an
obvious and important KPI.
Another parameter that has a negative effect on
profitability is the amount of water that is replenished.
Sustainability
The avoided CO2-emissions are and should be an
important driver for DH grids. According to subsequent
directives in the Netherlands for assessing energy
performance of buildings NEN 7120, the avoided
CO2-emissions has to be determined on the required
primary energy sources and by referring to common
state-of-the-art technologies. The HR-107 type (107%
LHV efficiency) is the required and accepted common
state-of-the-art reference technology.
Reliability
The condition of the network in terms of reliability
presents itself in the amount of times that mechanics
have to deal with unplanned repairs. It was apparent
that SVP was facing an increasing trend curve. The
actual deprecation of the replaced piping provided a
another criterion for assessing system degradation.
Customer satisfaction
Reducing off time, during replacement was an
important element of the SlimNet approach.
KPI summary
1. Heat loss
2. Water replenishment
3. Avoided CO
2
emissions
4. Unplanned repairs
5. Network degradation
6. Off-time during replacement
DEVELOPING ANALYZING TOOLS
Research had located the hotspots of unplanned
repairs and heat loss in an area of 4000 houses. These
hotspots were responsible for 50% of the unplanned
repairs. In order to define and implement a suitable and
cost-effective replacement strategy a set of tools was
developed.
Upgraded network diagram
Analyzing networks requires reliable and
comprehensive network diagrams. All required
information such as dimensions, age, depth etc. should
be available in the diagram. Many network diagrams
are drawn using CAD-software. Analyzing from those
drawings is costly. It therefore was chosen to revise the
diagram completely and apply the possibility to add
element attributes to the drawing connected to an
integral database system. The upgraded network
diagram had a catalytic effect on two other models, the
network model and the grid valuation model.
Network model
In 2009 SVP replaced the outdated and inadequate
network with a validated dynamic model (TERMIS),
developed by 7-Technologies with COWI as system
integrator. With the upgraded network diagram SVP
had the first and validated model of the primary
network within five months.
In combination with a new CRM system, operational
since 2010, SVP will soon be able to tap into the
information on customer behavior and consumption.
This will allow SVP to dynamically calculate the current
state of flow, pressure and temperature throughout the
network at a configurable cycle time. Additionally, every
real-time model calculation cycle will include a forecast
simulation for a given period. This allows SVP to be
abreast of demands, enabling optimization of
operations and planning of the future.
Valuation model
The upgraded network diagram supplied database
information on lengths, dimensions, age and type. With
the following equations added to the database it was
possible to develop a valuation model, that could help
to prioritize and direct renovation efforts.
x
network
x
x
L R X

=
=
1
(2)
x x
network
x
x
L R
D
A D
Y |
.
|
\
|
=

=1
(3)
September 5
th
to September 7
th

56

( )
x x
network
x
x
L R
D
A D
Y Z |
.
|
\
| +
=

=1
1
(4)
X = value network in new state ()
Y = current network value ()
Z = required annual maintenance costs ()
x = pipe
Rx = construction costs per meter pipe dimension x ()
Lx = length of pipe x (m)
D = lifetime expectancy (year)
Ax = age of pipe x (year)
The network degradation is defined as factor |:
X
Y
= | (5)
From consultation with amongst others COWI, it was
concluded that networks with a |< 0.5 are in a critical
stage.
For the entire grid the | was above the threshold.
Discriminating the | for separate grid sections helped
to identify the hotspots and monitoring | will help to
determine the effect of SlimNet.
Sustainability assessment model
To assess the current sustainability results of the
network, SVP developed a sustainability assessment
model in accordance with Dutch law and guidelines,
resulting in Fig. 3 [5]. This model can also predict the
effects of optimization in the chain from production,
distribution and delivery to customer installations

Fig. 3 CO2 reduction DH-network Purmerend in past

It appeared that the ratio of CHP operation to the total
of heat produced and the heat loss factor have the
biggest impact on the sustainability results.
Strategic metering
It was concluded that actual data on heat loss on
smaller scale (houses and clusters of houses) would
facilitate decision making on future renovation projects
and grid management. To get hold of this information
SVP installed heat meters with radio transmission
modules on strategic positions in the network. Together
with the metering data from heat meters in customer
installations this firstly gives accurate data on the heat
loss in the corresponding part of the grid. This setup
will also provide us with empirical data on the long term
results of network improvement measures.
In order to make the data comparable, two areas where
chosen. One with the new SlimNet approach (Using
polybutene pipes and new construction techniques)
and one with conventional material and construction
techniques. First comparative results will be available
by the end of 2010.
Leak detection
Most producers of pre-insulated pipe systems offer the
possibility of leak detection wiring. Using a master plan
with proper zero and recurrent measurements this
would be a reliable method of leak detection.
Unfortunately this is not applicable to the situation in
Purmerend.
With one of its partners SVP developed a method using
tracer gas to detect leakages. The detection devices
proofed to be very sensitive and with this method
almost 2.500 houses have been inspected this year
and last year. Leakages were detected in 3% of those
cases, mostly in an early stage, that otherwise would
only have been detected through visual sighting of
damp.
DEVELOPING AND DEFINING MEASURES
It became clear soon that the only way to improve
network performance was to rigorously renovate the
hotspots and to start implementing a structural
maintenance program in accordance to Z, Eq. 4.
In sum the challenge was: a.) cost effectively renewing
the steel pipes with wet insulation in narrow crawl
spaces while b.) improving network efficiency.
To meet (a), SVP started the first two pilots in 2008
with pre-insulated steel flex piping material, using two
different construction methods. Both pilots met the
technical requirements but were too time consuming,
costly and, because access to the crawl spaces had to
be gained by digging in the gardens, meant huge
inconvenience for customers.
Parallel to this SVP had challenged pipe manufactures
to come up with innovative material construction
methods, suitable for the Dutch situation (groundwater
September 5
th
to September 7
th

57

and retrofit in narrow crawl spaces). The only viable
solution came from Flexalen of Thermaflex, using
flexible polybutene (PB) carrier pipes. The producer of
the PB material offers a 50 years plus life guarantee [6]
for the pressures and temperature profiles of the SVP
network.
A pilot with Flexalen was conducted in September
2009. The pilot used prefabricated joints of Flexalen,
called Flexalinks, which were under research and
development at that time. The pilot did meet all the
requirements. Costs were reduced by 30% compared
to the steel flex pilots, 16 houses were overhauled
within a week and access could be gained by the crawl
space hatches.
On the basis of this pilot decision has been made to
retrofit 4000 houses within four years. Works has
currently started at the first 309 houses, at a speed of
30 houses a week.
The second part of the challenge (b): improving
network efficiency, is furthered by SlimNet through
optimizing pipe dimensions and lengths (smart grid
redesign)
Heat losses can be reduced by optimizing:
1. Thermal conductivity
2. Pipe lengths
3. Radial dimensions
4. Fluid temperature
These elements are captured in the following equation
for heat loss in a pipe [7]:

in
out
out in
pipe loss
r
r
T T
L k Q
ln
) (
2
_
= t (6)
k = thermal conductivity (W/mK)
L = length of pipe (m)
Tin = temperature of inside layer pipe (K)
Tout = temperature of outside layer pipe (K)
rin = inner radius (mm)
rout= outer radius (mm)
The first three of the above heat loss parameters can
only be changed by renewing pipes. The last can only
be changed by chain modification, i.e. production and
customer installations. SlimNet addresses both.
SlimNet part I: Renovation and smart redesign
Applying Flexalen means an improvement of k from 0,1
of the wet post-insulated steel pipes to theoretically
0,031 W/mK (manufacturer information, at 50 C).
Key to the SlimNet approach was smart redesign.
Calculations in TERMIS showed that many parts of the
DH-grid in Purmerend are generally oversized, and that
the common circular grid can easily be changed into a
star shaped grid, whilst reducing pipe lengths. Using
TERMIS redesign focused on reducing radial
dimensions and pipe lengths by deleting obsolete
pipes.
The results for the part of the grid that is replaced this
year, Fig. 4 and Fig. 5, gave, Table 2 [8]:
Table 2: Results from redesign 2010 area
Heat demand Heat loss
Current situation 100,0 % 100,0 %
New dimensions 93,0 % 76,3 %
Finger system 91,0 % 69,5 %

Fig. 4 Existing network part to be renewed

Fig. 5 Redesigned and renewed network
September 5
th
to September 7
th

58

The actual effect of SlimNet on heat losses will be
closely monitored in the grid, through the strategic
metering project.
SlimNet part II: Smart chain management
The last heat loss parameter, fluid temperature (Eq. 6),
can only be changed by modification of the complete
chain.
To start at the production side, the current supply
temperature is dependent on the ambient temperature,
95 C at Ta=-10 C and 75 C at Ta=15 C, Fig. 6.
Lowering this curve, while still meeting the
requirements of customer installations, would reduce
the average network temperature hence the heat
losses. It was calculated through the network model
that the alternative temperature curve in Fig. 6 solely
would reduce the heat losses with 4%. Further
research will focus on matching the most effective
temperature curve with production characteristics.

Fig. 6 Existing and alternative temperature curve

This research will also look upon the possibilities of
implementing demand-driven heat production. This is
achieved by using a real time network model
connected to the substations and production SCADA.
The model uses the weather forecast with customer
information to adjust the temperatures and pressures
just to meet the requirements of customer installations.
It is expected that this will reduce the average fluid
temperature even more.
Further research is done to implement cascading
heating services, i.e. using the latent heat in the return
pipes of the network with temperatures between 45 C
and 60 C to the customer installations. This is
however only possible to implement in new houses with
low temperature heating installations. This research will
focus on further reducing the heat losses.
At the other end of the chain are the customer
installations. Since 1996 the district heating company in
Purmerend has only accepted installations in new
houses that have a 9050 C characteristic during
design conditions (-10 C). In most areas before that
time SVP found return temperatures that are
structurally higher than the required 50 C. Hence the
flows in those areas are also much higher than
necessary.
The high return temperatures and corresponding high
flows are caused by absence of pressurizing valves in
the customer installations and defective control valves
in the hot tapping water installations. By the end of
2010 SVP starts a campaign to encourage house
owners to improve or renew their installations, also for
their own benefit. This campaign will make use of local
approved installers of customer installations. Research
indicated that in certain areas the peak flow can be
reduced with 60% [9].
QUANTIFYING KPI RESULTS FROM SLIMNET
Summarized, the measures that SVP takes before
2014 to improve network efficiency:
1. Renewing the distribution pipes and house
connections in the crawl spaces of 4000
houses, while optimized to dimensions and
lengths.
2. Replacing 4,0 km PEX-pipes in the primary
network, while optimized to dimensions and
lengths.
3. Doing this with a minimum of off-time for
customers
4. Implementing demand-driven heat production
5. Implementing cascaded heating installations
6. Encourage house owners to improve or renew
their installations in accordance with SVP
guidelines.
7. Eliminating arrears of maintenance an
implementing a structural preventative
maintenance program.
Heat losses will reduce from 33,6% in 2008 to 22,1% in
2015. While heat consumption prognoses stays the
same, the corresponding required heat production falls,
Fig. 8. This results in a energy saving of 227.000 GJ
that year. In Fig. 9 the results of the sustainability
assessment model are shown regarding CO2 savings.

September 5
th
to September 7
th

59

Fig. 7 Required heat production with SlimNet

Fig. 8 CO2 savings with SlimNet

Replacing the post-insulated steel and PEX pipes
together with a maintenance program including leak
detection will have a positive effect on the water
replenishment. The leak detection actions have already
resulted in a 30.285 m replenishment in 2009, which is
a 7% reduction compared to 2008.
It is expected that al measures will result in a 50%
reduction in 2015. Unplanned repairs will also reduce
50% and consequently | is expected to improve
significantly.
REFERENCES
[1] A. D. Heidweiller, B. C. Van Leeuwen and C. L.
Paarmann, Systeemstudie Stadsverwarming
Purmerend, Tebodin B.V., Den Haag, the
Netherlands (2006)
[2] A. E. Klop, B. P. Mensink and C. F. Dervis,
Transitiestudie Stadsverwarming Purmerend,
DWA Installatie- en Energieadvies, Bodegraven,
the Netherlands (2009)
[3] A. L.J.A.M. Hendriksen and B. R.A. Brand,
Onderzoek naar storingen in het
stadsverwarmingnet van Purmerend (report 034-
APD-2009-0021), TNO Bouw en Ondergrond,
Apeldoorn, the Netherlands (2009)
[4] A. M. den Burger and B. D. Heidweiller,
Deelrapport 5: Warmteverliezen en
meetverschillen, Tebodin B.V., Den Haag, the
Netherlands (2005)
[5] F. Dervis, Nulmeting duurzaamheid SVP, DWA
Installatie- en Energieadvies, Bodegraven, the
Netherlands (2009)
[6] J.J. Ribberink, Lifetime prediction of PB pipes
used in a district heating network, KIWA N.V.
Certification and inspection, Rijswijk, the
Netherlands (2009)
[7] A. D. A. Kaminski and B. M. K. Jensen,
Introduction to thermal and fluid engineering,
John Wiley& Sons, Hoboken, USA (2005), pp 103
[8] T.A. stergaard, New dimensions for O16,
COWI A/S, Aarhus, Denmark (2010)
[9] A. B. Zitoony and B. E. Roukema, Rapport
inregelstatus onderstations Stadsverwarming
Purmerend 10.001.V2, Roukema B.V., Groningen,
the Netherlands (2010)
September 5
th
to September 7
th

60

A DIRECT HEAT EXCHANGER UNIT USED FOR DOMESTIC HOT WATER SUPPLY IN
A SINGLE-FAMILY HOUSE SUPPLIED BY LOW ENERGY DISTRICT HEATING
Marek Brand
1
, Jan Eric Thorsen
2
, Svend Svendsen
3
and Christian Holm Christiansen
4

1
Ph.D. student, Technical University of Denmark
2
Senior project manager, Danfoss District Energy, Nordborg, Denmark
3
Professor, Ph.D., Technical University of Denmark
4
Danish Technological Institute, Denmark

ABSTRACT
The increasing number of new and renovated buildings
with reduced heating requirements will soon make
traditional District Heating (DH) systems uneconomic.
To keep DH competitive in the future, the heat loss in
DH networks needs to be reduced. One option is to
reduce the supply temperature of DH as much as
possible. This requires a review of the behaviour of the
whole domestic hot water (DHW) supply system with
focus on the user comfort and overall costs. This paper
describes some practical approaches to the
implementation of this Low Energy District Heating
(LEDH) concept. It reports on the testing of the dynamic
behaviour of an Instantaneous Heat Exchanger Unit
(IHEU) designed for DHW heating and space heating in
detached family houses supplied by LEDH ensuring an
entry-to-substation temperature of 51 C. We measured
the time it takes for the IHEU to produce DHW with a
temperature of 42 C and 47 C when the tap is
opened. Measurements were made for control
strategies using internal and external by-pass and no
by-pass. Our results show the importance of keeping
the branch pipe warm if comfort requirements are to be
fulfilled, but this involves higher user costs for heating.
To increase user comfort without increasing costs, we
propose the whole-year operation of floor heating in
bathrooms, partly supplied by by-pass flow.
INTRODUCTION
District Heating (DH) is a well known concept of
providing buildings with heat for space heating (SH) and
Domestic Hot Water (DHW) heating in economical and
environmentally friendly way. Nowadays, building
regulations have been introduced worldwide and are
pushing to reduce energy consumption in buildings,
because 40% of all energy consumption takes place in
buildings. The energy policy of European Union is
recently focused on energy savings, reducing
production of CO
2
and increasing the ratio of renewable
energy [1]. DH is one of the most suitable solutions to
achieve these goals for building sector and it gives high
priority for further development of DH. But recenlty used
traditional high and medium temperature DH systems
are not optimal solution for the future. Sooner or later,
energy consumption of all buildings will be in
accordance with low energy building regulations and it
will form areas with lower heat demand than nowadays.
Currently used DH networks will not be able supply
these areas in economical way, because the ratio
between network heat losses and heat consumption in
buildings would be unacceptable and thus cost of heat
for end users will increase and DH systems will loose
concurrency with other solutions, e.g. heat pumps.
Recently, research in DH is focused to find the way how
to use DH in areas with low energy buildings and how to
increase ratio of heat produced by renewable sources of
energy as solar heat plants or heat pumps driven by
electricity from renewable sources.
One of interesting application of renewable energy in
DH is use of decentralised heat sources as e.g. solar
collectors installed on roofs of individual buildings,
supplying heat to DH network, but it still needs more
time and work to develop new substations and new
concept of DH networks to be able to handle these new
features. The solution for future development of DH is to
reduce heat losses of DH networks by means of pipes
with better insulation properties e.g. twin pipes, use
better concepts of network design (circular network
configuration, possibility of using circulation line for main
pipes) and to reduce the supply temperature of district
heating water to lowest level as possible.
The District Heating Systems designed due to this
philosophy are called Low Energy District Heating
Systems (LEDH). The main focus in LEDH system is to
reduce heat losses from network as much as possible,
exploit more sources of renewable energy for heat
supply and still maintain or improve level of comfort for
users, because without high level of comfort this
concept cant be successful. LEDH concept was
reported e.g. in project Development and
Demonstration of Low Energy District Heating for Low
Energy Buildings [2], where theoretical case study
documented, that LEDH concept is a good solution for
future and even in sparse housing areas is fully
competitive to heat pumps. This article is focused on
application of LEDH for DHW heating. Considerations
related to use of LEDH for space heating will be
reported in future in another article.
September 5
th
to September 7
th

61

LOW TEMPERATURE DISTRICT HEATING
CONCEPT
Reduced risk of Legionella by use of system with
minimal volume of DHW
Since LEDH is mainly developed for low energy
buildings already designed with low temperature space
heating, the lowest acceptable forward temperature of
LEDH system is defined by requirement for DHW supply
temperature. The hygienic requirement for heating of
DHW is due to recent standards 50 C for single-family
houses and 55 C for multi-storey buildings [3] where
DHW circulation is used. In case of using circulation,
temperature of recirculated water should never fall
below 50 C. These requirements are based on need to
avoid Legionella growth in DHW pipes and storage
tanks. It is widely believed, that Legionella grow in
temperature range between 46 C 20 C, in systems
with high volume of water. Mentioned temperature
levels are made in order to assure comfort and hygienic
requirements in furthest tap away from a heat source. It
is important to say, that there is high level of
discrepancy among different results and national
standards focused on Legionella.
Due to German Standard W551 [4], temperature of
DHW can be below 50 C and not cause Legionella
promotion, if total volume of DHW system connected to
one heat source is lower than 3 L. From literature
studied, it can be concluded that requirements to
produce DHW with temperature higher than 50 C are
defined for an old fashion DHW building installations,
which can be characterized as systems with vertical
riser, branched pipes with bigger diameter (increasing
water volume of the system), using DHW circulation.
For new and renovated buildings, DHW installations are
designed in much better manner, with individual
connection of DHW pipes between each tap and source
of DHW and with maximally reduced pipe diameter,
defined by requirements for noise propagation and
pressure drop.
Due literature, danger of Legionella growth in DHW
system is influenced by temperature of DHW, nutrients
in DHW, laminar or turbulent flow in the DHW pipes and
water stagnation [5]. Several on site measurements
were performed in buildings using DH for DHW heating.
From results of Martinelli [6] and Mathys [7] can be
concluded, that Instantaneous Heat Exchanger Unit
(IHEU) tend to have much less problems with Legionella
than traditional units with DHW storage tank. Both
studies concluded, that these findings are caused by the
fact that in IHEU, DHW is produced with temperature
60 C, while in case of storage units only with
temperature 50 C. But is necessary to mention, that in
case of traditional DHW storage tanks, overall volume of
DHW in a system is much higher than in case of IHEU
system. Due to our knowledge, there is not reported
investigation of Legionella in DHW system using IHEU,
producing DHW with temperature below 50 C and
reduced volume of the system below 3L.
For single family houses with appropriate close location
of tapping points, volume of DHW in IHEU and pipes will
be lower than 3 L and thus temperature of 50 C on
primary side will not cause Legionella problems. For
multi-storey buildings, district heating substations for
each flat is a state of the art solution [8]. In this case,
each flat has own completely separated DHW system
(with volume of water below 3 L) and thus has
increased users comfort and no huge DHW systems
with circulation, where Legionella is forming and
spreading [9]. The other advantage of using flat station
in multi-storey buildings is individual metering of each
flat and complete control over space heating and DHW
preparation, which is positively affecting energy savings.
With properly designed DHW building installations,
supply temperature of LEDH will be defined by
requirements for users comfort. These requirements are
discussed in following text.
Users comfort in DHW supplied by LEDH
Another important question, when concerning DHW
systems is level of user comfort. From comfort point of
view, requirements for temperature and waiting time for
DHW can be specified. Due to Danish Standard DS439
Code of Practice for domestic water supply
installations, [10] temperature of DHW should be 45 C
in kitchen and 40 C in other taps, provided with
nominal flowrate and desired temperature reached
within reasonable long time, without significant
temperature fluctuations. It is a question, if requirement
of 45 C degrees for kitchen tap is not too high, but
argument of problems with fat dissolving from dishes
can be objected and should be investigated. Based on
mentioned standard, desired temperature of DHW
flowing from fixture is 45 C. But in order to define
desired forward temperature of LEDH system, we
should be aware of temperature drop in DH network, in
users substation and in DHW installations in building.
The temperature drop in DH network is not in focus of
this paper, so our goal is to find needed temperature
level at the entrance of substation to produce 45 C
from tap in building. Desired temperature will be found
by experimental measurement of LEDH substation later
in article.
Beside temperature requirements, users comfort is
influenced by time needed for DHW to reach a fixture
after tapping was started. This waiting time is in
following text called tap delay. Due to DS439,
suggested value for tap delay is 10 sec and it is defined
in order to avoid wasting of water and to protect users
against too long waiting times for DHW. In large multi-
storey buildings with centralised preparation of DHW,
short tap delay and measures avoiding Legionella
September 5
th
to September 7
th

62

growth are assured by circulation line of DHW, but not
properly designed or maintained DHW circulation is
quite often responsible for increased risk of Legionella
[11]. Another disadvantage of DHW circulation is big
heat losses, sometimes even bigger than net energy
needed for DHW heating [8]. The 10 sec waiting time is
not rule and for some people it is a long time, for some
people short, but this value is used to evaluate tested
concepts if they are fulfilling requirements for high level
of users comfort or not. An overall tap delay can be
studied from different angles. From dynamic point of
view, tap delay consists of transportation time needed
for new volume of water travel to tap and dynamic
thermal behaviour of passed components, i.e. pipes and
substation. From point of view related to location, it
consists of three parts, tap delay in branch pipe (pipe
from DH pipe in street to users substation), in DH
substation and in DHW system in building. A tap delay
in branch pipe and substation are related to DH network
and substations control system strategy, while tap
delay in DHW pipes in buildings without DHW
circulation are defined only by thermal capacity of pipes,
volume of water in individual pipes, nominal flow and to
some extend also by their insulation.
Tap delay in DHW system in building
For DHW systems with individual feeding pipes and
overall volume of pipes lower than 3 L, DHW circulation
is not needed, because waiting time for DHW with
desired temperature is not critical. In Table 1, transport
delays for individual fixtures in typical house built in pilot
LEDH project in Larch Garden - Lystrup, Denmark [11]
are presented. It should be mentioned, that data are
only transport delay, without dynamic behaviour of
cooled pipe. From Table 1 can be seen, that reasonably
designed close locations of fixtures, not so far away
from substation, lead to maximal transport delay around
6 sec, for basin. The total volume of DHW system
consists of 0.99 L in pipes and 1.1 L in HEX (type
XB37H-40). It means, that it is possible to install longer
pipes or more fixtures and still fulfil requirement of DHW
system with volume lower than 3 L. The velocity of
flowing water is below 2 m/s and thus problems with
noise propagation during tapping are avoided.
Table 1 Transport delay for nominal flows for individual
fixtures due to DS439, in DHW system in typical house in
Lystrup, for pipes with inner diameter 10 mm

Tap delay on primary side
A transport delay on primary side consists of delay in
branch pipe and delay in DH substation. While tap delay
in DHW installations in building is for DHW system
without circulation uniquely determined, tap delay on
primary side varying as control strategies for substation
control varies. From energy consumption point of view,
the best solution is a control strategy without by-pass
(see Fig. 1). In this case, DH water staying in the branch
pipes is cooled down to temperature of ambient ground
(if tapping wasnt performed for long time) and DH water
in substation to room temperature. In general, waiting
time for DHW is influenced by controller used in
substation. Basic principles of controllers are
proportional flow controller and thermostatic controller.
Each controller has own advantages and
disadvantages, thus best solution is to combine both
controllers [12]. In case of proportional flow controller,
ratio between primary and secondary flow is fixed to
provide DHW with desired temperature and it means in
case of using LEDH primary and secondary flow will be
very similar. If proportional flow controller is used for
setup without by-pass, user will face long waiting time
for DHW. Waiting time for this case can be seen from
Table 2. For branch pipe with inner diameter 15 mm (as
is designed in Lystrup for IHEU), even transport delay to
reach substation for nominal flow for basin, kitchen sink
and shower will be 31.6, 17.7 and 12.6 sec,
respectively. This solution is from comfort point of view
and water savings completely unacceptable. If we
decrease inner diameter of branch pipe to 10 mm,
transport delay is decreased roughly to one half of value
for pipe with inner diameter 15 mm, but it is still long
time. In case of combined proportional flow controller
and thermostatic controller, from beginning of tapping
thermostatic part assures opening of valve on
approximately full capacity until desired temperature of
DHW is reached.
Table 2 Transport delay for nominal flows for individual
fixtures due to DS439, in branch pipe, 10 m long, for typical
house in Lystrup, data simulate using proportional flow
controller without by-pass

Full opening from beginning of tapping leads to much
higher flow rate on primary side than on secondary and
time delay is decreased substantially. This solution can
be used for short branch pipes with reduced diameters.
But it should be mentioned, that transport time in branch
pipe will be always limited by maximal allowed flow on
fixture
nominal
flow
(L/min)
length
to
fixture
(m)
volume
in
pipes
(L)
velocity
(m/s)
transp.
delay
(s)
shower 8.4 2.2 0.17 1.8 1.2
basin 3.4 4.1 0.32 0.7 5.8
kitchen 6 6.3 0.49 1.3 4.9
fixture
nom.
.flow
(L/min)
inner
pipe
d
(mm)
volume
in pipes
(L)
velocity
(m/s)
transp.
delay (s)
basin 3.4 15 1.77 0.3 31.6
kitchen 6 15 1.77 0.6 17.7
shower 8.4 10 0.79 1.8 5.6
shower 8.4 15 1.77 0.8 12.6
bath 12.6 15 1.77 1.2 8.4
September 5
th
to September 7
th

63

primary side defined by DH provider by means of flow
restrictor or by available differential pressure in DH
network. To reduce tap delay on primary side, control
concepts with by-pass, avoiding cooling of DH water in
branch pipes and substations, and thus reducing
substantially waiting time for DHW are available. There
are two concepts of by-pass in relation to the heat
exchanger: external and internal by-pass (see Fig. 1). In
case of external by-pass, DH water enters substation,
but not enters heat exchanger and is sent back to DH
return pipe and thus branch pipe is kept on desired
temperature. Desired temperature is controlled by
thermostatic valve situated in by-pass loop. Increased
level of comfort expressed by reduced tap delay can be
adjusted independently on temperature of DHW on
secondary side.

Fig. 1 Different by-pass strategies for IHEU: left - no-by pass; middle - external by-pass (cold HEX); right - internal by-pass
(hot HEX)
The set-point temperature of external by-pass is
always compromise between insufficient cooling of DH
water and additional heat consumed by customer and
reduced waiting time for DHW. In case of operation of
space heating system, the function of by-pass is to
some extend overtaken by space heating loop and
thus heat for by-pass operation is not wasted and
temperature of DH water returning to DH network is
cooled sufficiently..In case of internal by-pass, by-pass
flow is passing through heat exchanger and keep it
warm (see Fig. 1). The benefit of this solution is even
more reduced tap delay than in case of external
by-pass, but on the other hand, since heat exchanger
is kept warm, internal by-pass solution has additional
heat losses. If substation is installed in room with need
of space heating, heat losses are considered only
outside of heating season.
Contrary to external by-pass solution, where it is not so
important if space heating loop is installed in series or
in parallel to DHW heat exchanger, in case of internal
by-pass it is in importance. If space heating loop is
connected in parallel to DHW heat exchanger in
traditional way, by-pass water just pass through DHW
heat exchanger and is sent back to DH network with
still high return temperature, without any other use. If
space heating loop is connected in series to DHW heat
exchanger or in parallel but with possibility to sent by-
pass water flown through internal by-pass to space
heating loop (see Fig. 2), this solution provides high
level of comfort for users as well as proper use of heat
needed for by-pass operation.

Fig. 2 Combined by-pass concept, with possibility of use
by-pass flow in space heating loop
In order to run by-pass without drawback of insufficient
cooling of DH water and wasted heat also outside of
heating season, it is proposed to use by-pass flow for
floor heating, installed in bathroom and operate it all
year. From preliminary calculations it looks, that flow
needed to keep bathroom floor surface temperature on
24C will be enough as by-pass flow. Considering the
use of renewable sources of heat, the problem of
insufficiently cooled DH water is related to reduced
efficiency of these sources and whole year using of
floor heating for comfort in bathroom is reasonable.
Supply supply recirculation
As an alternative solution for customers who dont
want to use whole year bathroom floor heating,
solution called supply-supply recirculation is a
possibility how to use benefits of by-pass without
whole year heating of bathroom. In this case, district
September 5
th
to September 7
th

64

heating water is supplied by pipe 1 to substation,
circulated through HEX or external by-pass (see
Fig. 3) and then sent back to district heating network
(DHN) supply by pipe no.3. This concept is in early
stage of investigation but it looks promising. The main
question will be related to flow of DH water in branch
pipe in order not cool it down too much before will be
sent back to DH supply pipe in the street.

Fig. 3 Supply supply recirculation with external by-pass
This solution is expected to be favourable mainly for
circular shapes of DH networks, but it should be
mentioned, that re-heating stations will be probably
needed in point of DH network, where temperature of
DH water decrease bellow defined value.
Full scale demonstration of LEDH
Full scale demonstration of LEDH is recently running in
Larch Garden in Lystrup, Denmark [11], where 40 low
energy houses class 1 and 2 are connected to LEDH
system, with designed forward temperature from heat
plant 52 C. For primary side of substation, forward
temperature of 50 C and return temperature of 25 C
are designed. The DH network is built from highly
insulated single pipes (for main pipes) and main pipes
with smaller diameter, distribution and branch pipes
are built from twin pipes. Two types of district heating
substations providing houses with DHW and space
heating are tested by customers in real conditions. The
first concept is 29 Instantaneous Heat Exchanger Units
(IHEU), second is 11 District Heating Water Units
(DHWU). IHEU is classical concept of substation with
instantaneous heat exchanger, only with enlarged
number of plates. IHEU units have external by-pas,
with set point temperature of 35 C for customers
situated not at the end of street and 40 C for
customers situated at the end of the street. DHWU is
new concept of DH substation, reported e.g. by
Paulsen [13]. DHWU consist of buffer tank for district
heating water and when DHW is needed, DHW is
heated in instantaneous heat exchanger as in previous
case. Advantage of concept with buffer tank is peak-
shaved demand of DH water during charging and use
of branch pipes with lower diameter, connected with
lower heat loss. On site measurements were started in
Lystrup to evaluate performance of both types of DH
substations, but no detailed measurements requiring
short time steps are performed to evaluate level of
users comfort. The measurements more focused to
users comfort are planed to be performed this year in
Danish Technological Institute and Technical
University of Denmark (DTU) on DH systems
simulating the conditions in Lystrup. The DH systems
will consist of branch pipes, substation and DHW
building installations and different control approaches
(external or internal by-pass, different set up by-pass
temperatures, possibility of supply-supply recirculation,
etc.) will be studied for DH substations supplied by
LEDH. Measured data will be used for evaluation of
performance of different control concepts, level of
users comfort and lately also for validation of
numerical model which is aimed to be developed for
optimization LEDH systems.
TEST OF TEMPERATURE PERFORMANCE
As a first part of measurements planed to be
performed at DTU, the time needed for IHEU to
produce DHW with temperature of 42 C and 47 C
was measured, after tapping of DHW was started. The
tap delay was investigated for two control strategies,
one using internal and second using external by-pass.
The measurements were performed for different initial
conditions before tapping was started to simulate in
realistic way users behaviour. Finally, the period
between two by-pass flow operations was measured.
Experimental setup and instruments
Tested DH substation was prototype of Instantaneous
Heat Exchanger Unit (IHEU) developed specially for
LEDH pilot project in Larch Garden Lystrup,
Denmark. The IHEU is a type of district heating
substation consists of a heat exchanger (HEX) without
storage tank. DHW is heated instantaneously in HEX
only when tapping is performed and then supplied
directly to DHW taps by individual feeding pipes, while
space heating is using direct connection without heat
exchanger, i.e. concept typical for Denmark.
Substation is same concept as regular IHEU for
traditional DH. The difference is in increased number
of plates in heat exchanger assuring better heat
transfer. Water volume of primary and secondary side
is 1.1 L each and the heat exchanger is not insulated.
The experiments were focused only on dynamic
behaviour of substation related to DHW heating and
thus space heating loop wasnt connected and space
heating valves in substation closed. Desired
temperatures of DHW were chosen in accordance with
requirements in DS439 for temperature of DHW for
kitchen sink and other fixtures. Required temperatures
mentioned in DS 439 are 45 C and 40 C. In order to
September 5
th
to September 7
th

65

cover additional temperature drop in building DHW
installations, 2 C were added. This addition is based
on experience from previous measurements. During
the experiments, temperatures of four different flows
passing through the DH substation were measured.
On primary side it was temperature of DH water
supplied to substation (T11) and temperature of DH
water returning back to DH network (T12) and on
secondary side it was temperature of cold potable
water entering substation (T21) and temperature of
heated DHW (T22). All temperatures were measured
by thermocouples type T, installed directly in pipes, in
flowing water, so they do not have any practical time
delay for the measurements. The time constant to
reach 90% of step change was less than 1 second.
The distance of thermocouples from substation flanges
was 5 cm and thermocouples were previously
calibrated. We also measured surface temperature of
HEX in upper (HEX-UP) and bottom part (HEX-
DOWN) and temperature of air in the testing room.
Temperatures were measured and collected by
multifunction acquisition unit every second. For
authentic simulation of DH network, DH water with
constant temperature of 51 C was necessary. It was
solved by connecting of IHEU to source of DHW in
laboratory of DTU, where DHW is supplied by DH
system. DHW system of DTU is big enough, to assure
stable temperature 51 C without any fluctuations. In
order to prevent cooling down of pipes supplying DHW
to laboratory in periods when there was not flow
through substation (stopped by by-pass controller),
small guard flow, just before entrance to substations
was kept to maintain DHW always on 51 C and
drained to sink.
Experimental procedure
As a first step, both controllers were adjusted to
provide 47 C on DHW side with supply temperature of
DH water 51 C. Then we measured time delay in the
substation, i.e. time needed for substation to produce
DHW with temperature 42 C and 47 C on secondary
side outlet from the moment when DHW tap is opened.
The measurements were performed for different initial
conditions and secondary flowrate was always
8.4 L/min, which is nominal flow for shower.
1. For measurements of concept with external by-pass,
substation was controlled by PTC2+P controller with
by-pass set point temperature adjusted to 35 C. This
setup is exactly the same as is installed in Lystrup pilot
project. The testing procedure was made in following
steps. Substation was left idle for long time in the
testing room, so all components and water in HEX
were on room temperature. Than we opened the valve
on DH supply in substation and DH water with
temperature of 51 C started to flow in the substation
and flew through external by-pass, until closing
temperature was reached and by-pass flow stopped.
Then we wait until by-pass was opened again. Time
between two by-pass openings as well as volume and
temperature of DH water passed through by-pass was
written down and after by-pass was closed again, we
waited a little bit shorter time than was needed to open
by-pass flow again and we start tapping on secondary
side with flow rate 8.4 L/min. In this way, most
unfavourable condition for substation with by-pass, i.e.
highest recovery time, was measured. After tapping of
DHW was finished, we wait 5 minutes and we
performed one more tapping to simulate short time
step between two subsequent tapping of DHW.
2. For measurement of internal bypass concept, IHPT
controller was used. In case of IHPT, by-pass set point
temperature cant be adjusted independently and is
defined by desired temperature of DHW, i.e. 47 C for
our measurements. IHPT controller was developed for
traditional DH networks operating with forward
temperatures around 70 C. For traditional DH, by-
pass opens when temperature in HEX falls 57 C
below set point of DHW, but in case of LEDH with
forward temperature 51 C, by-pass opens 1 C below
DHW set point temperature, i.e. 46 C in our case.
The testing procedure was similar to measurements
with external by-pass. After supply valve on primary
side of substation was opened, DH water with
temperature of 51 C started to flow in the substation
and temper HEX, until by-pass closing temperature
was reached. Then we wait until by-pass was opened
again and we performed tapping of DHW just before
next by-pass opening was expected. In following steps
was procedure same as in case of external by-pass.
Moreover, we also performed measurements of time
delay in IHEU for control concept without by-pass.
RESULTS
Time delay for IHEU with PTC2+P controller and
external by-pass adjusted to 35 C to start supply
DHW water with temperature 42 C and 47 C after
long idling period just before opening of external by-
pass was expected, can be seen from Fig. 4 and is 11
and 22 seconds, respectively. This measurement
represents condition with the longest time delay for
PTC2+P controller. Temperature of room, where IHEU
was installed was 22.2 C. For this case, temperatures
of produced DHW in first 10 sec after tapping was
started are listed in Table 3.
September 5
th
to September 7
th

66

Fig. 4 Time delay for external bypass (PTC2+P), when tapping is performed just before expected start of by-pass flow, set
on 35 C.
In case, when tapping of DHW was performed after
long idling just after by-pass flow was stopped, time
delay decreased to 8,5 and 16,5 seconds. In this
measurement, temperature of substation and thus
water standing in the HEX was little higher than
ambient air temperature. It is expected that time delay
will be slightly longer, if substation will have real
ambient temperature but still shorter than in case 2.
We also performed measurement of tap delay five
minutes after previous DHW tapping was finished.
In this case, tap delay in substation to produce DHW
with temperature 42 C and 47 C was shorter, 7 and
14 seconds.
For room temperature around 22 C, external by-pass
was opened roughly every 30 minutes. The by-pass
was in average opened 2.5 minute and volume of DH
water needed to close the by-pass was in average 3 L,
i.e. when substation is idle, by-pass uses 6 L of DH
water per hour.
Table 3 Temperatures measured for PTC2+P controller in first 10 sec after tapping was started for situation after long
idling, just before by-pass was expected to run again
t(sec) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
T22 (C) 21.6 22.3 26.0 29.7 32.6 35.0 36.9 38.7 39.9 41.2 42.2 42.8 43.5 44.2 44.7 45.1 45.5

Time delay in IHEU equipped with IHPT controller with
internal by-pass adjusted by requirement of DHW to
47 C was 6 and 14 seconds to reach 42 C and 47 C
on outlet for situation when tapping was performed just
before by-pass was expected to open. The internal by-
pass opens 3 minutes after previous tapping is finished
and when is once opened never closes, only when
another tapping is performed, but again only on
3 minutes.
The average flow of internal by-pass was 24 L/hour
and average return temperature to DH network was
45 C. When internal by-pass is once opened, the time
delay in substation decrease substantially to 1.5 and
7 seconds to produce DHW with temperature 42 C
and 47 C. The condition with expected longest time
delay was solution without by pass. In this case time
delay to produce DHW with temperature 42 C and
47 C was 12 and 25 sec. All measured results are
summarized in Table 4.
Table 4 Overview of time delays for all measured cases

case number and description
T11
(C)
42
(sec)
45
(sec)
47
(sec)
T12
(C)
T12AVG
(C)
THEX-UP
(C)
THEX-DOWN
(C)
NO BY
PASS 1 after long idling, no by-pass (BYP)
50.1 12 18 25 16.2 19.5 20.4 21
EXTERNAL
BY-PASS
2 after long idling, just before BYP was
expected to open again
49.6 11 16 22 30.1 19.3 21.5 21.4
3 after long idling, just after BYP closed 50.6 8.5 12 16.5 42.6 19 29 26
4 5 minutes after previous tapping finished 50.8 7 10 14 25 19.1 22.3 37.4
INTERNAL
BY-PASS
5 just before BYP was expected to open (3
min after prev. tapp. finished))
50.5 6 10 14 19.5 19.1 22.6 38
6 anytime, when BYP was already in
operation
49.3 1.5 3.5 7 47.3 18.4 44 45.5
September 5
th
to September 7
th

67

DISCUSSION
Focused on level of users comfort and proper cooling
of DH water during idling, time delay of LEDH
substation to supply DHW with temperature 42 C and
47 C was measured. Three different control strategies
related to tap delay were investigated. Obtained
results represent case of IHEU used in single-family
house in period when space heating is not in
operation. Explored concepts can be evaluated from
two different points of view, due to highest advantages
for customer and for DHN.
The solution without by-pass is from energy savings
point of view very interesting because doesnt need
any DH water for idling, but from users comfort point of
view is very poor because of reduced comfort and
problems with wasting of water during waiting for DHW
with desired temperature. Solution without by-pass can
be probably used for substations equipped with
combined thermostatic and proportional flow controller,
for customers with short branch pipes or for customers
with low requirements for level of users comfort. If
solution without by-pass will be used for substation
controlled only with proportional flow controller, even
transport delay in 10 m long branch pipe for nominal
flow for basin will be 32 sec. For period when space
heating is operated, branch pipe will be kept warm
from flow needed for space heating and time delay for
solution without by-pass will be very similar to solution
with external by-pass. Anyway, in non-circularly
shaped DH networks, by-pass should be installed at
least at the end of a street, so it is better to find
solution how to use by-pass flow in useful way than
sent it directly back to DH return. Considering this, it is
suggested to use by-pass flow for whole year
operation of floor heating in bathrooms to increase
comfort for customers and at the same time solve
problem with by-pass flow which otherwise increasing
return temperature to DH network.
From user comfort point of view, better solution than
solution without by-pass, but consuming more energy,
is substation equipped with external by-pass. By
comparison of results of concepts without by-pass
(case 1) and solution with external by-pass, for case
when tapping is performed after long period of idling
just before by-pass opens again (case 2), we can see
that time delays are almost the same (see Table 4).
Difference is only that for external by-pass are pipes in
DH substation kept on higher temperature and it made
slightly faster reaction. In the case 3, time delay is
even more reduced since pipes in substation were
warmer by just finished by-pass flow. For control
concept with external by-pass and tapping repeated
5 minutes after previous one, time delay is again
reduced, since HEX is still hot from previous tapping.
The time delay for case 4 and 5 are almost the same,
only difference is that in case 5 (internal by-pass), tap
delay is again reduced because tapping was
performed 3 minutes after previous (to prevent
influence of by-pass) and thus HEX was warmer.
If the requirement is to fulfil 10 sec tap delay for less
favourable fixture, i.e. in our case basin (see Table 1),
DHW should leave DH substation with temperature
42 C in 4 sec after tapping was started, because it will
take 6 second to reach the tap. This requirement was
reached only by concept with internal by-pass and only
when by-pass was already opened. On the other hand
from Table 3can be seen, that even for concept with
external by-pass and tapping after long idling and just
before expected bypass opening, DHW at a
temperature 26 C leaving substation in 3 sec. DHW
with this temperature is not sufficient for taking a
comfortable shower for which temperature 371 C is
preferred, but for washing hands this temperature
should be enough. The values in Table 3 are for flow
rate used for shower, but it can be used to explain that
it is time to rethink the suggested value of tap delay
from 10 sec to another value and consider also
nominal flows and use of tapped water. The different
standards for the different use of DHW based on new
solutions in DHW supply systems and results from test
panels are needed, because it may have some
influence on design of optimized DHW systems.
Nevertheless, for customers requiring DHW in very
short time e.g. continuously or discontinuously (only
during rush hours) operated trace heating elements
can assure almost no tap delay by keeping DHW
staying in pipes on desired temperature.
CONCLUSION
Based on literature study it can be concluded that
hygienic requirement of DHW with 50 C on outlet of
DHW heater is not needed for systems with a total
volume of the DHW lower than 3 L.
From results of our measurements and evaluation of
IHEU supplied by LEDH, only substation with external
by-pass with set point 46 C is able to produce 47 C
DHW in time bellow 10 sec. The easiest step how to
decrease waiting time also for other concepts is to
insulate HEX. This measure will reduce time delay for
DHW tapping and also will decrease heat losses from
DH substation. The lower waiting times for DHW can
be also achieved by further optimisation of HEX in way
of decreased number of plates reducing volume of
water in HEX and thus transport delay, and by
increased thermal efficiency of HEX (followed on the
other hand by higher pressure loss). These
modifications can lead for higher temperature of DH
water returning to DH network, but during all our
experiments, average return temperature was below
20 C, what is 5 C less than is designed for LEDH.
September 5
th
to September 7
th

68

Traditional control concepts of DH substations are
always trade-off between users comfort and reduced
cooling of DH water during idling and thus customer
should have to some extent possibility to choose which
solution prefers. In case of traditional concepts,
decision is between longer waiting time for DHW and
energy savings or vice versa, if by-pass in substation
is used. In non-circularly shaped networks, by-pass
should be used anyway at least at the end of a street
line. The one of possible solutions how use by-pass
flow in better way can be proposed innovative concept
of whole year operated floor heating in bathrooms or
supply-supply recirculation. Both solutions will increase
level of user comfort and at the same time also energy
efficiency of DH system.
LEDH is a promising solution for providing buildings
with DHW and space heating regarding fulfilling
requirements of modern society with reduced CO
2

emissions and energy consumption. More detailed
investigations by testing of different parameters and
numerical simulations are needed in order to optimize
LEDH concept.
Future work
It will be very interesting to compare time delay of
substation for traditional DH with time delay for DHW
produced by LEDH substation. It is expected that
timed delay for LEDH will be higher because dynamic
response is slowed down by lower temperature
difference between DH water and desired temperature
of DHW, but on the other hand, lower temperature
difference is in some extend compensated by bigger
HEX. It is also suggested to rethink 10 sec tap delay
suggestion for different taping flows and purposes of
DHW use.
REFERENCES
[1] S. Froning, Low energy communities with district
heating and cooling, PLEA 2008 25th
Conference on Passive and Low Energy
Architecture, Dublin,
[2] Hovedrapport, Udvikling og Demonstration af
Lavenergifjernvarme til Lavenergibyggeri 2009,
(in Danish)
[3] EUROHEAT & POWER, Guidelines for District
Heating Substations, 2008, downloaded from
www.euroheat.org in October 2009, pp 8
[4] DVGW, W551 - Trinkwassererwrmungs- und
Trinkwasserleitungsanlagen ,1993, Bonn, (in
German)
[5] Z. Liu, Effect of flow regimes on the presence of
Legionella within the biofilm of a model plumbing
system, 2006, Journal of Applied Microbiology,
Vol. 101, pp 437-442
[6] F. Martinelli, A Comparison of Legionella
pneumophila Occurrence in Hot Water Tanks and
Instantaneous Devices in Domestic, Nosocomial,
and Community Environments, 2000, Current
Microbiology, Vol. 41, pp. 374-376
[7] W. Mathys, J. Stanke, et. al., Occurrence of
Legionella in hot water systems of single-family
residences in suburbs of two German cities with
special reference to solar and district heating,
2008, Int. J. Hyg. Environ. Health, Vol. 211, pp.
179-185
[8] H. Kristjansson, Distribution Systems in
Apartment Buildings, Published at the 11th
Cooling, August 31 to September 2, 2008,
Reykjavik, ICELAND
[9] T. Persson, District Heating for Residential Areas
with Single-Family Housing, paper IV, 2005,
Doctoral Thesis, Lund Institute of Technology,
Lund
[10] Dansk Standard, DS 439 Code of Practice for
domestic water supply installations, 2009
[11] P.K. Olsen, Low-Temperature District Heating
System for Low-Energy Buildings, 2009,
http://www.fbbb.dk/Files/Filer/Peter_Kaarup_Olsen
_-_COWI_29-10_2009.pdf
[12] H., Boysen, J.E. Thorsen, Control Concepts for
DH Compact Stations, Published in Euroheat and
Power IIII 2004
[13] O. Paulsen, Consumer Unit for Low Energy
District Heating Net, Published at the 11th
Cooling, August 31 to September 2, 2008,
Reykjavik, ICELAND

September 5
th
to September 7
th

69

CHALLENGES ON LOW HEAT DENSITY DISTRICT HEATING NETWORK DESIGN
M. Rm
1
and K. Sipil
1

1
VTT Technical Research Centre of Finland PB 1000, FI-02044 VTT, Finland

ABSTRACT
While district heating is an energy efficient solution to
provide heating to areas with high heat consumption,
mature systems extending out to more demanding
operational environment face challenges maintaining
competitiveness over alternative heating systems. As
the heat density falls below a certain level, district
heating is no longer economically feasible. Studying the
possibilities of extending this threshold by means of
district heating system design and pointing out the
operational challenges while approaching it are the
main topic of this paper.
The problem is investigated in a representative case of
a low heat density area bordering a more extensive
district heating network. A node-and-branch type
network simulation model is used study the operation of
the network and a simulation period of one year is used
to get a realistic view of the system in a normal
operational cycle.
Not taking into account the characteristics of a low heat
density area in network design can result in inefficient
distribution system. Operational problems, especially
maintaining the temperature level in summertime, must
be solved. Only concentrating on minimizing the heat
losses will not result in best possible design.
The temperature level issue can be solved with a by-
pass valve, auxiliary heating or accumulators, but in
overall more efficient system requires steps to be taken
in the houses. Floor heating and a heat pump coupled
with an accumulator enables the use of low temperature
design where the heat losses can be cut significantly.
INTRODUCTION
District heating remains to be one of the most efficient
alternatives to provide heating mostly due to its high
total efficiency especially when utilizing combined heat
and power production or waste heat from industrial
facilities or other sources. A wide choice of production
technologies, based on fossil or renewable fuels or
other sources of heat, provide flexibility to district
heating systems and enable the benefits from the
economy of scale unlike most consumer specific
heating systems. From the consumer point of view,
district heating is considered as a reliable and carefree
source of heating energy and is also often an
economically sound choice.
Areas with high heat consumption i.e. economically the
most attractive areas will be connected first to district
heating. The expansion of mature and large scale
systems take place in areas with lower heat
consumption. This transition to more demanding
operational environment both technically and financially
represents challenges to district heating network design.
This is also true in small scale systems of limited
consumption separated from a larger system.
A careless network design in these circumstances can
lead to deterioration of the advantages of district
heating; efficiency and reliability. An annual heat loss of
5% in district heating distribution is considered a good
result, but the case in question the heat losses can
easily reach 10% or even tens of percents if the
characteristics of low heat density areas are not taken
into account in design.
LOW HEAT DENSITY AREA
A detached house area consisting of 56 identical 150
m
2
houses with energy consumptions in compliance of
todays building standards is studied. Dedicated heat
exchangers between the network and the consumer
exist for both heating and domestic hot water. Total
energy consumption for the houses is 18.75 MWh/year
of which domestic hot water has a share of 20 percent.
The district heating network studied is presented in
Figure 1. The detached house connections are marked
as green dots and the connection to the main district
heating network as a red rectangle. The connections
have 1, 2 or 6 detached houses as consumers,
indicated by the size of the dot.
50 m 50 m

Figure 1. District heating network studied.
The total trench length in the area is 2 390 m of which
the service pipes (DN 15-25) account for 1 300 m. The
pipe size distribution is illustrated in Figure 2. The dark
blue coloured bar (DN 65) represents the pipe
connecting the area to the main district heating network.
September 5
th
to September 7
th

70

As the pipe diameters are quite small, twin pipes with
insulation class IV are used in the area as
recommended by Energy Industry [1], [2] in Finland.
The pressure drop design principle used here is roughly
~1.5 bar/km.
0
100
200
300
400
500
600
700
15 20 25 32 40 50 65
Pipe size (DN)
P
i
p
e

l
e
n
g
h
t

(
m
)

Figure 2. Pipe size distribution

The linear heat density is 0.44 MWh/m which makes the
area a low heat density area by definition [3].
The heat demand around the year is presented in
Figure 3. The peak demand for the area is 507 kW. As
expected, in the summertime the load consists almost
solely of domestic hot water consumption.
0
100
200
300
400
500
600
0 50 100 150 200 250 300 350
Days
T
o
t
a
l

h
e
a
t

d
e
m
a
n
d

(
k
W
)
Figure 3. Heat demand of the simulated area.
SIMULATION MODEL
A node-and-branch type simulation model [4] was used
to study the case in hand. The model calculates
temperatures and pressures for the nodes and flows
and heat losses for the pipes, i.e. the branches. From
these results pumping power can also be calculated,
although a constant efficiency of 0.5 is used for the
pump. The pressures are calculated separately from
temperatures. The temperature calculation is dynamic
while the flow and pressure calculation is not. A
minimum 0.6 bar pressure difference over a consumer
is assumed.
When defining the network, each pipe is given a start
and an end node, a pipe type (twin, single), an
insulation standard (class I to IV) and length.
The consumptions for both heating and domestic hot
water use were given as hourly time series as well as
the radiator supply and return temperatures on the
secondary side.
The heat exchangers were modelled with logarithmic
temperature principle in a design point (described in
Table 1) after which the conductance in W/K is
assumed to be constant. When heat demand, both
supply and return temperatures on secondary side and
supply temperature on primary side are given as input,
the primary return temperature and district heating mass
flow can be calculated.

Table 1. Design point for heat exchangers.
Description Value
Primary side temperatures 115/45 C
Radiator heating 70/40 C
Domestic hot water 55/10 C
Design heating load 8 830 W
Design DHW load 2 060 W

The design loads for domestic hot water are low
compared to a real life design load of a heat exchanger
in normal detached house in Finland, 50 kW is a
common choice. This is due to the simulation model
taking hourly data originally calculated for a multifamily
house as input so the domestic hot water demand is
also flatter than it really is. However, from the network
design point of view hourly data is considered accurate
enough.
Other input data used were the undisturbed ground
temperature of 5 C, assumed to be constant, and the
supply temperature from the main district heating
network as a function of outdoor temperature. The
outdoor temperature time series used described a
typical year in Southern Finland. The supply
temperature reaches its maximum value of 115 C in an
outdoor temperature of -26 C and its lowest value of
75 C in 5 C. Between these two points, the relation is
linear.
SIMULATION RESULTS
The most interesting results concern the heat losses
and the temperature variations within the network. The
pumping needed (less than 1 MWh) in a network of this
size is quite low and thus negligible.
In the initial simulation runs it was noted that the system
was struggling to maintain high enough temperature
level in the summertime when the load consist solely of
domestic hot water demand. This problem was met by
defining a flow through valve at the consumer, opening
September 5
th
to September 7
th

71

when the supply temperature on primary side dropped
too low (< 65 C). The valve allowed a constant mass
flow (0.015 kg/s) to go past the heat exchanger on the
primary side. This solution helped the situation
significantly although not without ill effects as can be
seen from the heat losses presented below.
The use of a by-pass valve to ensure the appropriate
temperature level for domestic hot water also mean
higher heat losses and pumping power and effectively
lower cooling; all of which are undesirable outcomes.
One possibility to solve the problem is just to accept the
flaw and to use additional electrical heating element to
raise the temperature of domestic hot water to the
required level. As the temperature boost needed is for
most of the time quite small and is only needed in
summertime, the increase in electricity consumption is
reasonable.
Because of the high capital costs of district heating, the
pipes should basically be sized as tight as possible
while keeping in mind the future demand for the pipeline
in question. As the pipes are small, the volume of water
contained is also low. This leads to water cooling more
rapidly than in larger pipes. The Figure 4 illustrates this
with a simplified example by showing the temperature
on supply side service pipes if there is no flow for three
different pipe sizes. The temperature drop of 15 C, for
example, takes 5 times longer with a pipe size DN 50
than with a small DN 15 pipe. The calculations assume
a constant return side temperature of 30 C and a
ground temperature of 5 C.
0
10
20
30
40
50
60
70
0 2 4 6 8
Time (h)
T
e
m
p
e
r
a
t
u
r
e

(
C
)
DN 15 DN 25 DN 50

Figure 4. Temperature drop in three pipe sizes when no
flow is introduced.

The use of smaller pipes reduces the heat losses in
W/m and this is accentuated if the temperature level
drops as described above. As a result, looking solely on
heat losses when designing a low heat density area
network on common design principles can lead to
reliability issues as the system cannot supply the heat
required by the consumers.
The relative heat losses (that is, heat losses per needed
production) for the simulated case are 13.8 % in a year.
The monthly values can be seen in Figure 5. While the
relative heat losses in the heating season are
acceptable, they reached 47 % in the summertime. The
high heat losses are partly because of the by-pass valve
letting hot water past the heat exchangers. The by-pass
valve is also responsible for small cooling, i.e. the
difference between supply and return temperatures,
within the system in summertime (Figure 6).
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
I II III IV V VI VII VIII IX X XI XII
Month
R
e
l
a
t
i
v
e

h
e
a
t

l
o
s
s
e
s

(
-
)
Figure 5. Monthly relative heat losses.

0
10
20
30
40
50
60
70
80
90
0 50 100 150 200 250 300 350
Days
C
o
o
l
i
n
g

(
C
)
Figure 6. Difference between supply and return
temperatures at the border of the area.

The most obvious way to cut heat losses in already
reasonable insulated network is to lower the supply
temperature. In the simulated system, this would cause
problems because aforementioned issues concerning
domestic hot water demand in summertime, and during
the heating season because of the traditional radiator
heating design temperatures of 70/40 C. However, if
more significant changes would be possible, a floor
heating system and a heat pump coupled with an
accumulator handling the higher temperature level
required domestic hot water would enhance the
efficiency of the distribution system at a price of a very
modest increase in electricity consumption and higher
investment costs for the consumer because of the
accumulator, heat pump and floor heating. If the
domestic hot water demand takes 3.75 MWh/year,
20 percent of the total consumption of 18.75 MWh/year,
the electricity consumption would be a very reasonable
September 5
th
to September 7
th

72

1.25 MWh with an average COP of 3. With this setup,
supply temperature would need to be just 40 C.
CONCLUSIONS
The use of traditional district heating network design
principles can lead to an inefficient area heating system
in areas with low heat density. Special attention must be
paid on operation of the system to ensure reliability, one
of the advantages of district heating.
When aiming for an efficient system, one goal is to
minimize the heat losses. However, concentrating solely
on this can make another problem, maintaining high
enough temperature level for domestic hot water in
summertime, even worse. The problem can be solved
using a by-pass valve, but this causes unwanted
effects; worse cooling and an increase in heat losses
and pumping power. Other solutions are auxiliary
heating (electrical heating or a heat pump) or the use of
an accumulator and with it, aiming for a steady domestic
hot water load.
Another approach is lower the supply temperature
significantly and to use floor heating and heat pump with
an accumulator for domestic hot water demand. This is
not suitable for existing areas with a heating system
already designed, but for new areas it is a reasonable
and, compared to the traditional district heating design,
an efficient way to provide heating.
REFERENCES
[1] Lappeenranta University of Technology,
Kaukolmpjohtojen optimaalisen eristyspaksuu-
den tarkastelu / Investigation of the optimal
insulation thickness on district heating pipes,
Energy Industry, 2009, 36 p.
[2] Preinsulated district heating pipes,
Recommendation L1/2010, Energy Industry, 2010,
44 p.
[3] Zinko, H., Bhm, B., Kristjansson, H., Ottoson, U.,
Rm, M., Sipil, K., District heating distribution in
areas with low heat demand density, IEA DHC
Annex VIII, 2008, 117 p.
[4] Ikheimo, J., Sderman, J., Petterson, F., Ahtila, P.,
Keppo, I., Nuorkivi, A., Sipil, K. 2005. DO2DES
Design of Optimal Distributed Energy Systems,
Design of district heating network. bo Akademi.
Report 2005-1.
September 5
th
to September 7
th

73

DESIGN OF LOW TEMPERATURE DISTRICT HEATING NETWORK WITH SUPPLY
WATER RECIRCULATION
Hongwei Li
1
, Alessandro Dalla Rosa
1
, Svend Svendsen
1

1
Civil Engineering Department, Technical University of Denmark

ABSTRACT
The focus on continuing improving building energy
efficiency and reducing building energy consumption
brings the key impetus for the development of the new
generation district heating (DH) system. In the new
generation DH network, the supply and return
temperature are designed low in order to significantly
reduce the network heat loss. Meanwhile, the low
network operational temperature can make a better
utilization of renewable energy and further improve the
CHP plant efficiency.
Though the designed return temperature is low, it may
increase considerably when the heating load becomes
low and the by-pass system starts to function. The aim
of this paper is to investigate the influence of by-pass
water on the network return temperature and introduce
the concept of supply water recirculation into the
network design so that the traditional by-pass system
can be avoided. Instead of mixing the by-pass water
with return water, the by-pass water is directed to a
separated circulation line and returns back to the plant
directly. Different pipe design concepts were tested and
the annual thermal performances for a selected
residential area were evaluated with the commercial
program TERMIS. The simulation program calculates
the heat loss in the twin pipe as that in the single pipe.
The influence of this simplification on the supply/return
water temperature prediction was analyzed by solving
the coupled differential energy equations.
INTRODUCTION
In European Union, one of the major energy
development targets is to reduce the building energy
consumption and increase the supply of renewable
energy. The introduction of European Energy
Performance of Building Directive (EPBD) poses
stringent requirement for the member countries to
effectively reduce their building energy consumption.
According to the national energy policy, the building
energy consumption in Denmark will drop to 25% of
current level by the year 2060, while the renewable
energy share will increase from 20% to 100% at the
meantime [1].
District heating (DH) benefits from economic of scale
with mass production of heat from central heating
plants. The significant reduction of building energy
consumption and wide exploitation of waste heat and
renewable energy, however, makes the current DH
technologies become barriers to further increase the
market share [2]. In order to sustain the economic
competiveness and realize the long term sustainable
development, the concept of design and operation of
DH system needs to be re-examined under the new
energy regulation and development trends. This is the
main impetus for the development of the new
generation DH system. Based on previous studies, in a
properly designed in-house substation system, the
network supply temperature at 55oC and return
temperature at 20oC can meet the consumer space
heating and domestic hot water demand [3].
The low return temperature has the advantages to
reduce the network heat loss, increase CHP plant
power generation capability, and utilize direct flue gas
condensation for waste heat recovery. However, the
return temperature can become much higher than the
designed value when the heating load becomes low
and the by-pass system at the critical user starts to
function. In this paper, the influence of by-pass water
on network return temperature was examined for a
reference residential area. The concept of supply water
recirculation was introduced to avoid the mixing of by-
pass water and the return water. Three network design
methods were tested. The annual thermal performance
was evaluated with the commercial district heating
network hydraulic and thermal simulation software
TERMIS [4]. The simulation program calculates the
heat loss in the twin pipe as that in the single pipe. The
influence of this simplification on the supply/return
water temperature prediction was analyzed by solving
the coupled differential energy equations.
SUPPLY WATER RECIRCILUATION
The solution to overcome the excessive temperature
drop along the supply pipe due to reduced flow rate is
to install by-pass system at the critical user in the
network. Figure 1 shows the principle of supply water
by-pass. Extra flow is called based on the temperature
measurement at the critical user until the minimum
supply temperature requirement is met. This extra flow
is then by-passed and sends back to the return pipe.
As the by-pass flow rate may be considerable and its
temperature is high, the mixing with return water will
significantly increase the return water temperature
which causes both increased heat loss in the return
pipeline and decreased power generation capability in
the CHP plant.
September 5
th
to September 7
th

74

A desirable design approach is to maintain the by-pass
system as the flow rate adjuster, while avoids the
mixing of the by-pass water and the return water. This
design concept is schematically shown in Fig. 2, which
is realized through adding a third pipeline for supply
water re-circulation. When the by-pass water is called,
the circulation line will transfer the extra supply water
back to the plant where it is re-heated up to the supply
temperature again. On the other hand, the addition of
the 3rd pipeline provides the possibility to supply water
in two supply lines when the heat demand is high. The
network, therefore, can be designed as two supply
lines with reduced diameter together with one return
line.

Fig. 1 Schematic for hot water by-pass system

Fig. 2 Schematic for by-pass water recirculation

Fig. 3 Annual heating load (blue columns) and duration
hours (red curve) at different ground temperature

NETWORK SIMULATION
Heating Load
The simulation was performed for a reference area with
81 low energy demand houses. The house was
designed based on the building standard Class 1,
following the Danish Building Regulation. The domestic
hot water draw-off profile was designed similar to the
Danish standard DS439 [5]. Detailed space heating
and domestic hot water heating load simulation can be
found from [6, 7]. Figure 4 shows the averaged heating
load and the corresponding duration hours. The annual
heating load is divided into 8 intervals, varying as a
function of undisturbed ground temperatures which
ranges from 0 to 15 C. The summer season lasts 3281
hours and the heating load comes only from the
domestic hot water demand. The space heating is
required for the rest of the year.
House Installations
Two house installations were considered in this study.
Figure 4 shows the instantaneous heat exchanger (HE)
in the DH system. Without a buffer tank, the branch
pipe which connects directly to the HE installation must
have the capability to supply the instantaneous hot
water demand without causing significant pressure
drop, which otherwise can be compromised by
installing a booster pump. The HE design load is 32kW
per houses at the network supply temperature 55oC
and return temperature 22 C. On the other hand,
simultaneous factors which are the probabilities for
multiple users concurrent use of hot water are
considered for the design of street pipes and main
pipes, as shown in Table 1 [3]. Fig. 5 shows the
domestic hot water storage tank (DHWS) in the DH
system. The DHWS design load is 8 kW per house. To
avoid the legionella problem, the design temperature
for DHWS is higher than HE, at 65 C /30 C for supply
and return respectively.
September 5
th
to September 7
th

75

Fig. 4 In-house heat exchanger (HE) in DH system

Table 1 Simultaneous Factors

Fig. 5 Domestic hot water storage (DHWS) in DH system

District Heating Network
The DH network and the connection to the end users
are shown in Fig. 6. The house is designed to connect
to the plant directly through different diameter pipes
which were optimized with the simulation program. The
direct connection allows the primary DH network to
circulate water directly into the end user installation. It
is suitable for a moderate pressure level network and
the differential pressure of DH network is sufficient to
circulate water to the house installation. The networks
and house installations are assumed to withstand
maximum pressure 10 bar. The consumer differential
pressure is set as 0.5 bar. It is controlled at the end
user along the network critical route which is shown in
green color.
Three network design scenarios were investigated for
each house installation:
- Case 1: It is the reference case. The total
network length is 3080 m and the network line
heat density is 177 kWh/year. Network was
designed in the traditional way for two pipes
with one supply and one return, respectively.
The differential pressure is controlled at user
A. Twin pipes were selected for the
DH network. They are called reference pipe
in this paper.

- Case 2: By-pass water recirculation. A third
pipeline (Fig. 6 grey color line) was introduced
to separate the by-pass water with return
water and re-circulate the by-pass water back
to the plant. The third pipeline was sized
based on the summer by-pass water flow rate.
The differential pressure is controlled at point
B.

- Case 3: Double pipeline supply. The main pipe
(from plant to the junction point at each street)
in the third pipeline which was sized in case 2
functions all year round. It acts as supply pipe
during winter season and functions as supply
water recirculation pipe when there has by-
pass water demand. In this case, the main
pipe in the reference case was resized as a
portion of supply water is shared by the
recirculation pipe. The connection of
recirculation pipe to the reference pipe is
shown with red color.
The thermal by-pass temperature was set as 50 C for
HE and 60 C for DHWS with dead band 2 C. The by-
pass is placed on the end user at each street in case 1,
while at the virtual point adjacent to the end user in
case 2.
September 5
th
to September 7
th

76

Network Heat Loss Calculation
The reference network was designed with twin pipes by
placing the supply and return pipe in the same casing.
Two types of twin pipes were considered in the
simulation: AluFlex multilayer flexible pipe and straight
steel pipe. The pipes were selected with continuous
dimension ranging from Alx14 to 32 for AluFlex pipe
and DN 32 to DN40 for steel pipe, based on the market
available products [8]. Single AluFlex pipe is selected
for the 3rd recirculation line. This 3rd pipeline can be
assumed being placed in the same trench along the
twin pipes. The thermal interaction between the twin
and the single pipe is assumed negligible.
The heat loss in the twin pipe was calculated according
to the reference [7,9]
[1]

[2]

Fig. 6 District heating network
The supply and return pipe are assumed identical and
placed horizontally in the same depth from the ground.
The linear thermal transmittance U
ij
reduces to
U
11
=U
22
=U
1
and U
12
=U
21
=U
2
. In addition, the thermal
conductivity of insulation foam was assumed constant.
U
1
and U
2
were then calculated with the analytical
solution developed from the multi-pole method [10].
The simulation program cannot handle two heat
transfer coefficients in the same pipe, U
s
and U
r
were
derived to represent the overall heat transfer
coefficients corresponding to the temperature
difference between the flow and the ground.
The temperature variation along the pipeline was
calculated as internal flow with isothermal boundary
condition. The downstream temperature in the pipe is
expressed as [4]:
[3]

T
d
, T
u
and T
a
represent the downstream fluid
temperature, upstream fluid temperature, and ambient
temperature respectively. M and K are parameters
include the overall heat transfer coefficient. As the
overall heat transfer coefficients have to be calculated
beforehand, the influence of flow temperature variation
on U
s
and U
r
along the pipeline is neglected. It is a
reasonable assumption when the thermal by-pass
temperature is set close to the plant temperature,
however, may cause appreciable errors if the
temperature drop along the network is high.
It is worth to be noted that though the design return
temperature (22
o
C) is higher than ground temperature,
the net heat transfer in the return pipe may absorb heat
from surrounding which makes U
r
negative. However,
negative U
r
has to be set to zero as the simulation
program cannot handle negative heat transfer
coefficient.
RESULTS AND DISCUSSION
Heat Exchanger
Network simulation starts from proper selection of pipe
dimension, based on the design condition and the
design criteria introduced in the previous section.
Table 2 shows the selected pipe types and
corresponding length for three different cases. Case 1
is the reference case. Flexible twin pipe Alx 20 to 32
and steel twin pipe DN32 and DN 40 were selected.
The third recirculation pipe was designed in case 2
based on the summer by-pass flow rate. Pressure
gradient 1500 pa/m for street pipes and 500 pa/m for
main pipes were set as the dimension criteria. Though
smaller pipe was suggested by the program, the Alx16
single pipe was selected as the minimum diameter pipe
available on the market. It was assumed that the
recirculation pipe can be used as water supply in winter
in case 3. Therefore, the main pipes in the reference
line were re-designed with considering that a portion of
supply water goes through the recirculation line. It can
be seen that the supply pipe has smaller diameter than
return pipe in some sections in the twin pipe line.
September 5
th
to September 7
th

77

Table 2 Selected pipe types and length in Case 13

Figure 7 shows the pressure profile along the critical
route. The network is designed for a 10 bar system.
The minimum network static pressure is 2 bar and the
minimum differential pressure at consumer is 50 kPa.
The plant static supply pressure is 853 kPa in case 1 at
design condition. In case 3, the designed plant supply
pressure head rise to 917 kPa, which is due to the
increased flow rate indicated in Table 4. The pressure
drop along the reference line during summer is quite
low due to the reduced flow rate. However, extra
pressure head has to be applied to overcome the
pressure loss along the recirculation line in Case 2.
The required static supply pressure is 800 kPa during
summer as a result of small dimension recirculation
line.
Table 3 shows the simulation results for case 1.
By-pass is required when the heating load is smaller
than 1.53 kW. The return water temperature increases
along with the increase of by-pass water flow rate. In
summer, the amount of by-pass water flow rate
exceeds the actual flow rate passing through the
consumer, and the return temperature at the plant
increases up to 35.5 C. The heat loss in the return
pipe is accounted when the plant return temperature is
raised to higher than 30 C.
Simulation results for case 2 and case 3 are shown in
Table 4. They were put in the same table as case 2
operates when there has by-pass requirement, while
case 3 operates in the rest seasons. Italic is used for
case 3 to distinguish the two cases. Thanks to the
recirculation line, the return temperature at the plant in
the reference line remains low at 22 C, while the return
temperature in the recirculation line can reach 44 C in
the summer, after deducting the single pipe heat loss.
The low plant return temperature can help extract more
power in the CHP plant or be used in other
circumstance like direct flue gas condensation. On the
other hand, high temperature return water in the
recirculation pipe can be re-heated by an additional
heat exchanger or boiler with minimum energy input.
Case 2 has higher return pipe heat loss comparing with
case 1 due to the introduction of recirculation line. At
constant supply temperature 55 C, the heat transfer
coefficient Us decreases with increase the return water
temperature. As shown in Table 3, the return water
temperature in case 2 (at 22 C) is lower than that of in
case 1. This leads to a higher heat loss in the supply
pipe in case 2. As a consequence, more by-pass flow
is required to compensate the extra supply pipe heat
loss, therefore, the by-pass flow rate in case 2 is higher
than in case 1 in the summer season.
Supply water in the recirculation pipeline in winter
increases the supply pipe heat loss in case 3. The
concept of double pipe supply may not economical
feasible, according to the simulation results. However,
it may be used as an alternative solution to supply
water in the 3
rd
pipeline under extreme whether
condition, which otherwise has to raise the plant supply
temperature to meet the increased heating demand.
Furthermore, results in table 4 were limited to fixed
recirculation pipe diameters. The double pipe supply
concept may be economical feasible by free selection
both reference pipe and recirculation pipe diameter
with the objective to minimize the annual network
operational cost or exergy consumption. This study is
out of the scope of current paper due to the limitation of
the simulation program.

Fig. 7 Pressure profile on the critical route in Case 13
September 5
th
to September 7
th

78

Table 3 Simulation results in Case 1

Table 4 Simulation results in Case 2 (First 5 rows) and
Case 3 (Last 3 rows with italic)

Domestic Hot Water Storage Tank
Table 5 shows the pipe types and corresponding length
in the DHWS installation. Alx 14 was selected as
branch pipe due to the smaller design heating load.
Similar to the HE, the by-pass flow rate exceed the
actual flow rate through the consumer in summer
season. The plant mixed return water temperature in
case 1 is 46
o
C. The introduction of the recirculation
line can keep the plant return temperature in reference
line as low as 30
o
C, while increases the return
temperature in the recirculation pipe to 54
o
C at the
plant. Extra heat loss has to be tolerated due to the
recirculation pipe in both case 2 and case 3.
Table 5 Selected pipe types and length in Case 13

Table 6 Simulation results in Case 1

Table 7 Simulation results in Case 2 (First 5 rows) and
Case 3 (Last 3 rows with italic)

Further Discussion on Heat Transfer
As shown in Eq. 13, the simulation program simplifies
the calculation of the heat loss in the twin pipe as that
in the single pipe. The influence of the adjacent pipe
was accounted through converting the linear thermal
transmittance U
ij
to the overall heat transfer coefficients
U
s
and U
r
, with pre-assumed constant network
supply/return temperatures. To assess the influence of
this simplification on the temperature predication, the
thermal interaction between the supply and return
pipes was calculated by solving the coupled pipe heat
transfer differential equations. The governing equations
for supply and return pipes can be expressed as:
[4]
[5]
The boundary conditions can be expressed as:
[6]
The dimensionless temperature is introduced with:
[7]

September 5
th
to September 7
th

79

The governing equations then change to:
[8]
Where
The boundary conditions change to :
[9]
The system linear ordinary differential equations can be
solved with Eigen value method or with Laplace
transformation. The Laplace transformation was
applied in this study. Eq. 8 is transformed to:
[10]
The final solutions are given as:
[11]
[12]
Where :
, [13]
, [14]
Tws- DN32, which is the longest main pipe in HE of
case 1, is selected for the assessment with U
1
=0.141
and U
2
=0.0523. The pipe length is assumed 500 m.
Ground temperature ranges from 0 to 15
o
C. The inlet
of supply and return temperatures are known as 55
o
C
and 22
o
C respectively. The outlet temperature of
supply pipe is controlled as 50
o
C and 45
o
C,
respectively.
Table 8 shows the temperature prediction based on
single pipe simplification and the coupled pipe
equations. T_Difference represents the coupled
solution minus the single pipe solution. When the
temperature drop along the supply pipe is controlled at
5
o
C, the prediction between the single pipe and the
coupled pipe is very close. The prediction errors
increase with increase the ground temperature. The
single pipe approach predicts lower supply water
temperature and higher return temperature than those
of coupled pipe solutions. It was also observed that
when the ground temperature is higher than 4
o
C, the
net heat transfer effect in the return pipe is to absorb
heat to the surrounding.
Table 8 Pipe temperature predication comparison (supply
outlet temperature is controlled at 50
o
C)

The by-pass water temperature in this study was set in
a conservative way. In many practices, the by-pass
water can be set 10 C lower than the supply water
temperature. Even lower by-pass temperature is
proposed for the low temperature district heating
network [3]. Table 9 shows the simulation results based
on a10 C temperature drop along the supply pipe. It
shows the prediction errors increase in both supply and
return pipes. The heat transfer was predicted in a
reverse trend in the return pipe at 4 C. Considerable
prediction error was found in the return pipe at high
ground temperature.
It is worth to be noted that the increase of supply
temperature drop has more influence on the return pipe
temperature prediction than that of supply pipe. The
reason can be explained from the expression of U
s
and
U
r
in Eq. 12. As the magnitude of T
s
-T
g
is higher than
T
r
-T
g
, the same amount of return water temperature
variation will have more influence on U
r
than U
s
,
therefore causes a larger prediction error in the return
pipe than in the supply pipe.
Table 9 Pipe temperature predication comparison (supply
outlet temperature is controlled at 45
o
C)

CONCLUSION
In this paper, a preliminary study was conducted on the
influence of by-pass flow on the network return water
temperature in a designed low temperature DH
network. The concept of supply water recirculation was
September 5
th
to September 7
th

80

introduced to avoid the mixing of by-pass water to the
return water. Double pipe water supply concept was
tested to use the recirculation pipe supply water during
winter season. Two different house installation modes
were considered in the analysis.
The by-pass water significantly increases the return
water temperature in the traditional design. The mixed
return temperature can reach 35.5
o
C for HE and
45.6
o
C for DHWS. With applying the by-pass water
recirculation, this return temperature can be maintained
at 22
o
C, while the re-circulated by-pass water can be
kept as high as 44
o
C and 53.5
o
C for HE and DHWS at
the plant, respectively. It was found that the double
pipe supply leads to the highest network heat loss.
However, the conclusion that whether the concept of
double pipe supply is inferior to other network design
methods can only be drawn after further network
thermal-economic optimization.
The simulation program simplifies the twin pipe heat
transfer prediction as a single pipe, and neglects the
return pipe heat loss when the return pipe absorbs heat
from the surroundings. The temperature prediction
errors due to the single pipe assumption were analyzed
through solving the coupled supply/return pipe
differential energy equations. The prediction errors
increase with increase the allowable temperature drop
in the network. Considerable error was found for the
return pipe at high ground temperature.
NOMENCLATURE
c
p
= specific heat capacity [ J/kg.K]
q = Heat transfer rate [kW / m]
s = Laplace transform variable
T = Temperature [ K]
U = Overall heat transfer coefficient [ kW /m.K]
U
ij
= Linear thermal transmittance [kW/m.K]
= mass flow rate [ kg/s]
Greek Letter
u = Dimensionless temperature
Subscripts
g = Undistributed ground
r = Return
s = Supply
u = Upstream
d = Downstream
Abbreviation
DH = District heating
HE = Heat exchanger
DHWS = Domestic hot water storage tank
REFERENCE
[1] H. Lund, B. Moller, B. V. Mathiesen, A. Dyrelund,
The role of district heating in future renewable
energy systems, Energy, 35, pp. 1381-1390,
2010.
[2] Charlotte Reidhav, Sven Werner, Profitability of
sparse district heating, Appliced Energy, 85, pp.
867-877.
[3] Udvikling og Demonstration af Lavenergifjern-
varme til Lavenergibyggeri, EFP 2007.
[4] TERMIS Help Manual, Version 2.093,
7-Technologies A/S.
[5] Dansk Standard DS 439, 2000. Norm for
vandinstallationer, Code of Practice for domestic
water supply installations, 3. udgave, www.ds.dk.
[6] Otto Paulsen, Jianhua Fan, Simon Furbo, Jan Eric
Thorsen, Consumer Unit for Low Energy District
Heating Net, The 11th International Symposium
on District Heating and Cooling, 2008, Iceland.
[7] P. K. Olsen, et.al, A new low-temperature district
heating system for low energy buildings, the 11th
Cooling, Iceland, 2008.
[8] Logstor. http://www.logstor.com/
[9] Benny Bohm, Halldor Kristjansson, Single, twin
and triple buried heating pipes: on potential
savings in heat losses and costs, International
Journal of Energy Research, 29, pp. 1301-1312,
2005.
[10] P.Walleten, Steady-state heat loss from insulated
pipes, Thesis, Lund Institute of Technology,
Sweden, 1991.

September 5
th
to September 7
th

81

STEADY STATE HEAT LOSSES IN PRE-INSULATED PIPES
FOR LOW-ENERGY DISTRICT HEATING
A. Dalla Rosa
1
, H. Li
1
, S. Svendsen
1
1
Technical University of Denmark

ABSTRACT
The synergy between highly energy efficient buildings
and low-energy district heating (DH) systems is a
promising concept for the optimal integration of energy
saving policies and energy supply systems based on
renewable energy (RE). Distribution heat losses
represent a key factor in the design of low-energy DH
systems. Various design concepts are considered in
this paper: flexible pre-insulated twin pipes with
symmetrical or asymmetrical insulation, double pipes,
triple pipes. These technologies are potentially energy-
efficient and cost-effective solutions for DH networks in
low-heat density areas. We start with a review of
theories and methods for steady-state heat loss
calculation. Next, the article shows how detailed
calculations with 2D-modeling of pipes can be carried
out by means of computer software based on the finite
element method (FEM). The model was validated by
comparison with analytical results and data from the
literature. We took into account the influence of the
temperature-dependent conductivity coefficient of
polyurethane (PUR) insulation foam, which enabled to
achieve a high degree of detail. We also illustrated the
influence of the soil temperature throughout the year.
Finally, the article describes proposals for the optimal
design of pipes for low-energy applications and
presents methods for decreasing heat losses.
INTRODUCTION
The energy policy on energy conservation poses
stringent requirements in the building energy sector, so
that the entire DH industry must re-think the way district
energy is produced and distributed to end-users [1, 2].
This is a requirement to be cost-effective in low heat
density areas. Low-energy DH networks applied to low-
energy buildings represent a key technology to match
the benefit of an environmentally friendly energy supply
sector and the advantages of energy savings policy at
the end-users side. Future buildings with a high
performance envelope will lead to reduced space
heating load and therefore to a lower required
distribution temperature for heating. The introduction of
low-energy DH networks is an appropriate and natural
solution to enhance energy and exergy efficiencies.
Distribution heat losses represent a key-point for
designing low-energy DH systems, due to the critical
role they have in the economy of the system. The
industry could meet the requirements of higher
insulation series to reduce heat losses and thus saving
operational costs; however, this option would increase
investment and installation costs. The design principles
for DH networks could instead be changed towards the
use of media pipes with small nominal diameters, with
a higher permissible specific pressure drop. All-year
around lower supply temperature and return
temperature constitute an effective option to reduce
heat losses [3]. These principles have a big potential
for heat supply to low-energy buildings, as explained in
[4] and they are investigated in this paper.
The total length of branch pipes can be significant in
proportion to the total length of the network, above all
in areas with a low-energy demand density. Moreover
the temperatures in the critical service lines affect the
temperature level in the whole network, so that the heat
losses and the temperature decay in building
connection pipes are decisive for the overall
performance of the system. In this paper particular
focus was given to branch pipes.
State-of the art of district heating pipes
At present time DH distribution and service lines are
based either on the single pipe system, where the
supply/return water flows in media pipes with their own
insulation, or on the twin pipe system, where both pipes
are placed in the same insulated casing, or in a mixture
of them. All plastic pipe systems are characterized by
having the water medium pipe made of plastic (cross-
linked polyethylene (PEX) or polybutylene (PB)). They
are covered by insulation, usually polyurethane foam,
but in some cases of PEX foam or mineral wool; the
outer cover is formed by a plastic jacket. Durability of
plastic pipes is not a real issue, since it has been
proved that the expected life of PB pipes and PEX
pipes is, respectively, more than 40 years and approx.
100 years [5]. As consequence of even lower average
operational temperature, longer lifetime can be
predicted according to Annex A in [6]. Studies have
indicated that cross-linked polyethylene (PEX) pipes
have a cost advantage over steel pipes at pipe
dimensions less than DN60, due to their greater
flexibility since the joints do not require welding [7].
Alternative design concepts must be considered in
branch pipes from street lines to consumers
substations: a pair of single pipes, twin pipes or triple
pipes. Traditionally most DH branch connections have
been built with two single steel pipes: one supply pipe
and one return pipe. Twin pipes can be made of steel,
September 5
th
to September 7
th

82

copper or PEX, with the supply and return pipe in the
same casing. The heat losses from twin pipes are
lower than from single pipes, considering same
dimensions and temperatures.
Furthermore commercially available twin pipes, with
dimensions up to DN200 for traditional steel media pipe
or up to DN50 for PEX media-pipes are usually less
expensive to install than single pipes [7]. This
technology has been introduced in Nordic countries
(and it is used in daily operation in many DH networks.
Triple pipes might be considered in the near future, due
to flexibility in the way the system can operate and
lower heat losses in case of optimal configuration. The
choice of house connections depends mainly on the
length of the branch pipe, on supply and return
temperatures, building heating load and type of
substation. The latter is decisive with regard to energy
performance and thermal comfort. The types of
substations are typically divided into three concepts:
unit with domestic hot water (DHW) storage tank,
where the tank is the secondary-loop and consumer
unit with DH water tank, where the tank is placed in the
primary loop. In this paper branch pipe solutions are
considered for the concept of a consumer unit with heat
exchanger and no storage tank. Two possible
configurations of user connection to the distribution line
are shown in Figure 1.

Figure 1: Sketch of a user connection with heat
exchangers: twin pipe connection with/ without booster
pump (12) and triple pipe connection (1-2-3).
1: supply
2: return
3: supply/re-circulation

A simple and cost-effective configuration is composed
of the control system and two heat exchangers for,
respectively, space heating (SH) and domestic hot
water (DHW). The main disadvantage of such type of
substation unit is that only rather short lengths of
service pipes can usually be applied; otherwise it would
not be possible to assure the required DHW
temperature at tapping points in the required time, due
to the unsatisfactory transportation time. A modified
unit is therefore proposed and it is equipped with a
booster pump which assures quicker response to DHW
demand, although a non perfect cooling of DH water
occurs when tapping of DHW starts. The concept
based on twin pipes and a substation with
instantaneous production of DHW in a heat exchanger
is an optimal solution, if certain conditions are
respected. The first requirement is that the control
method gives priority to DHW preparation over space
heating; the second condition is that the space heating
load during summer, to keep a high level of comfort in
bathrooms for example, has to guarantee a sufficient
cooling of the return water. As a result media pipes with
inner diameters as small as 10 mm can be applied in
the primary loop and the water return temperature can
be kept sufficiently low, even in summer conditions.
The triple pipe system is applicable in three different
operational modes. The first one (mode I) occurs in
case of DHW demand, when pipe 1 and pipe 3 both act
as water supply pipes; the second operational mode
(mode II) is activated when an idle water flow is
supplied by pipe 1 and pipe 3 acts as re-circulation line
to the supply distribution line, while the return line (pipe
2) is not active: this is often the case when there is no
demand for space heating, but a small amount of water
circulates in the DHW heat exchanger, keeping the
loop warm to satisfy the instantaneous preparation of
DHW in the required time. This system avoids an
undesirable heating of the water in the return
distribution line. The third operational mode (mode III)
occurs during the heating season when there is only
demand for space heating and no tapping of DHW:
pipe 1 and pipe 2 operate as a traditional supply-return
system, while there is no water flow in pipe 3. The
different modes are summarized as follows:
- Operational mode I: DHW tapping, pipe 1, 2, 3
active.
- Operational mode II: supply-to-supply
re-circulation, pipe 1, 3 active; pipe 2 not active.
- Operational mode III: space heating demand, pipe
1, 2 active; pipe 3 not active.
METHODS
Theory of steady state heat loss in buried pipes
In order to calculate steady-state heat losses in DH
buried pipes there are analytical methods [8] and
explicit solutions for the most common cases [9]. A
complete review of the available literature about
steady-state heat losses in district heating pipes has
been carried out in [10]. Here the methods are
presented with reference to the present status of the
technology in the district heating sector. Furthermore
key-points and critical aspects are discussed; finally,
improvements in the methodology of how to calculate
steady-state heat losses are proposed, with particular
focus on low-temperature and medium-temperature
September 5
th
to September 7
th

83

applications. Low-temperature district heating systems
are defined as networks where fluids at a temperature
below 50 C are used, while a medium-temperature
district heating system is defined as using fluids at
temperatures not higher than 70 C [11, 12].
Steady-state heat losses from pre-insulated buried
pipes are generally treated by use of the following
equation [10], which is valid for each pipe-i:
(1)
where Uij is the heat transfer coefficient between pipe-i
and pipe-j, Tj is the temperature of the water in pipe-j
and T0 is the temperature of the ground. In case of two
buried pipes, which is the most common application in
the DH sector, the heat losses can be calculated as
follows, respectively for the supply pipe and the return
pipe, where T1 is the supply temperature and T2 is the
return temperature.

Supply pipe: (2)

Return pipe: (3)

Equations (2) and (3) show how the heat transfer from
each pipe can be seen as linear superimposition of two
heat fluxes, the first one describing the heat transfer
between the pipe and the ground, the second one
representing the heat transfer between the supply pipe
and the return pipe. The equations can also be
re-arranged in the following way:
Supply pipe: (4)

Return pipe: (5)

Equations (4) and (5) show how the heat transfer from
each pipe can be calculated by use of only one linear
thermal coefficient, which is function of the temperature
in this case. U-values are dependent both on
temperature and time. If the time-dependency due to
the ageing of the foam can be restrained by introducing
effective diffusion barriers, that is not true for the
intrinsic dependency on temperature. It is practice to
evaluate the steady state heat loss applying a thermal
conductivity value that corresponds to a hypothesized
mean temperature of the insulation. Nevertheless we
need models based, for example, on the finite element
method (FEM) when complex geometries or a high
degree of detail are requested.
Temperature dependant thermal conductivity of
PUR insulation foam
In this paragraph the authors want to explain and
demonstrate the importance of taking into account the
temperature-dependency of the thermal conductivity of
the insulation (lambda-value). The temperature
gradient in the insulation foam in the radial direction is
often higher then 10 C/cm, meaning that the thermal
conductivity of the material locally varies remarkably. In
the example shown Figure 2, it varies more than 10%
of the prescribed mean value. This affects the
magnitude of the heat transfer. Considering a life cycle
assessment of a DH system, the main impact to the
environment is represented by heat losses [13]. The
thermal conductivity of the insulation material in
pre-insulated DH pipes is usually stated at a
temperature of 50 C. The lambda-coefficients were
chosen according to the available data at the end of
2009; the lambda-value at 50 C for straight pipes,
axial continuous production was set to 0.024 W/(mK)
and for flexible pipes to 0.023 W/(mK). Since April 2010
new results are available [14]. It is preferable to have a
model that takes into account the temperature-
dependency of the thermal conductivity of the
insulation foam. The calculations in this paper use the
following expression, if not differently stated. It derives
from experimental data [15]:
(T) = 0.0196734 + 8.0747308
.
10-5.T [W/(mK)] (1)

Figure 2: Thermal conductivity in the insulation, horizontal
cross-section of the pipe. Pipe: Aluflex 16-16/110,
temperatures supply/return/ground 55/25/8 C.
September 5
th
to September 7
th

84

Temperature field in the soil around the pipe
In this paper we address the question of how to create
a simple yet detailed FEM model for steady state heat
loss calculations. The overall heat transfer resistance
between the DH water and the environment is mainly
composed of the thermal resistance of the insulation
and the thermal resistance of the soil; compared to
these two factors, the thermal resistances of the pipe
wall and the convective resistance at the surface water-
pipe are in practice negligible. The insulation foam
always offers the greatest share in the overall
insulation effect. The contribution of the soil is smaller
on small-sized pipes than on large-sized pipes. The
share is smaller in Insulation Series 2 and 3 [3]. The
heat conductivity coefficient of the soil is the main
parameter affecting the thermal resistance of the soil
itself, and its value is often unknown in practice.In the
calculations we chose a value of 1.6 W/(m.K). The soil
temperature influences heat losses from DH pipes. The
soil layer around the heating pipes slightly warms up
around the pipes. The evaluation of the temperature
field in the soil is a prerequisite to create a realistic
model for calculations of heat losses. Finite Element
Method (FEM) simulations were carried out and
temperature conditions in the soil around a typical DH
service pipe, suitable for low-temperature applications
were evaluated over a 10-year period.
Table 1: Thermal properties of materials.
[W/(mK)] [kg/m] Cp [J/(kgK)]
soil 1.6 soil 1600 Cp_soil 2000
PE 0.43 PE 940 Cp_PE 1800
PUR 0.023
PU
R
60 Cp_PUR 1500
PEX 0.38
PE
X
938 Cp_PEX 550
Steel 76
Ste
el
8930
Cp_Stee
l
480
Cu 400 Cu 8930 Cp_Cu 385

The simulation calculated the soil temperature at
various x-coordinates from a commercial branch pipe.
The selected pipe was the Aluflex twin pipe 16-16/110.
Temperatures were set at 55 C and 25 C,
respectively for the supply pipe and the return pipe.
The heat transfer coefficient at the ground surface was
assumed to be 14.6 W/(m2K), including convection and
radiation [16]; we set the outdoor air temperature
during the year according to the harmonic function valid
for the Danish climate [17]:
|
.
|
\
|
+ =
12
2 sin 5 . 8 0 . 8
M
T
air
t (6)
Combined heat and moisture transfer is disregarded.
The material properties are homogeneous and phase
changes, i.e. freezing and thawing were not
considered. Table 1 lists the material properties, used
as input values also for the following models; a sketch
of the slab-model, where the boundary conditions are
described, can be seen in Figure 3.

Figure 3: Sketch of the model. Dimensions are in [mm].
FEM model
A rectangle representing a semi-infinite soil domain
(width: 1020 m, height: 2040 m) is the most used
geometry to model the ground in heat loss calculations
[18, 19]. In this paper a finite, circular soil domain was
applied, instead. Its diameter is 0.5 m and it is equal to
the distance between the surface and the centre of the
casing pipe. Calculations show that the introduced
simplification hardly affects the accuracy of the results.
The mesh model and an example of the temperature
field in a small size twin pipe are shown in Figure 4.

Figure 4: Mesh model of a pre-insulated twin pipe
embedded in the ground (top and left). Temperature field
in Aluflex twin pipe 16-16/110 (bottom-right); temperature
supply/return/ground: 55/25/8 C.

In [3], where FEM simulations were performed, it is
stated that for media pipes size from DN 50 to DN 400,
the deviation of the lineal thermal coefficient between
the piggy-back laying (arranging the supply pipe below
the return pipe) and the traditional system (horizontal
laying) is less than 1%. The same conclusion can be
stated for twin pipes; this is confirmed by calculations
September 5
th
to September 7
th

85

with the multipole method in [20] for two examples of
twin pipe (DN 20 and DN 80) and by [10]. For twin
pipes of even smaller size, such as in branch
connections, the heat losses occurring in case of
vertical layout are only slightly more favorable than the
losses occurring on horizontally arranged pipes; this
result is shown with an example in the results section.
RESULTS AND DISCUSSION
In this section we discuss the influence of the soil
temperature on heat losses; next, we present the
validation of the FEM models; finally we apply the
method to show the potential for energy saving in the
case of asymmetrical insulation of twin pipes, in the
case of double pipes and triple pipes.
Temperature field in the soil
Temperature conditions in the soil around a typical twin
pipe, type Aluflex 1616/110, were evaluated over a
10-year period. Figure 5 shows the all-year
temperature profiles of the outdoor air and of the
ground at depth equal to 0.5 m, at three horizontal
distances from the centre of the casing, during the first
year of operation. No notable differences in the yearly
profile were noticed in longer periods of time.
We found that in state-of-the-art well insulated twin
pipes (series 2 or 3) a certain amount of soil is slightly
heated up by the warm twin pipe; nevertheless the
level of such heating can be neglected because its
effect is not noticeable in comparison to the fact that
the uncertainties about the thermal properties of the
soil usually have a bigger impact. Considering yearly
average temperatures, the magnitude of the soil
heating is about 1 C for distances of around 0.2-0.3 m
from the centre of the casing, and less than 0.5 C by
0.5 m. The temperature raise is considered in
comparison to the undisturbed temperature of the
ground at a distance of 10 m.
Figure 5: All-year temperature profiles of the outdoor air
and of the ground at depth equal to 0.5 m and 3 horizontal
distances from the centre of the casing.
FEM model: geometry of the ground and of the
pipes
We considered the geometric model of the pre-
insulated Aluflex twin pipe type 16-16/110; the
temperatures of supply/return/ground are 55/25/8 C.
We calculated the heat losses for vertical or horizontal
placement of the media pipes inside the casing, which
was embedded in a rectangular or a circular model of
the ground. The same calculations were repeated for
other twin pipe size, up to DN 32 and other medium
pipe materials, i.e. steel and copper. The results
confirm that the vertical placement of the media pipes
inside the insulation barely affect the heat transfer,
being the difference between the two configuration less
than 2% for the considered cases.
Table 2: Heat loss for various placements of the media
pipes and various model of the ground.
Ground
model
Media
pipes
layout
Heat loss
supply
[W/m]
Heat loss
return
[W/m]
Heat loss
total
[W/m]
A Vert. 3.79 -0.17 3.62
A Horiz. 3.80 -0.18 3.62
B Vert. 3.84 -0.18 3.66
A: Semi-infinite, rectangular (width x depth: 40 m x 20 m)
B: Finite, circular (diameter: 0.5 m)
Steady-state heat loss in commercial pipes
The model was validated by comparing the results from
FEM simulation to the analytical calculation for pre-
insulated pipes embedded in the ground [14].
Calculations were carried out for four different sizes of
Aluflex twin pipes (size 1414, 1616, 2020, 2626)
and for chosen sets of supply (50, 55, 60 C), return
(20, 25, 30 C) and ground (8 C) temperatures. The
selected pipes are suitable to be used as branch pipes
in low-energy demand areas. There is a good
accordance between the two methods, the deviation
being lower than 1%. Figure 6 gathers the values of
total heat loss for the Aluflex twin pipe category; four
different approaches are reported. The term standard
is used when the effect of the temperature on the
thermal properties of the insulation is neglected and the
thermal conductivity of the PUR foam is thus constant.
This is in accordance with [21]. The term advanced is
used when the calculation method takes into account
that the thermal conductivity of the insulation depends
on the temperature. Based on the temperatures
calculated for a number of points in the insulation the
program calculates an average temperature for the
material; the lambda-value of the insulation is then
calculated as a function of such temperature. An
average temperature of the ground is similarly
calculated. The calculation is repeated until the mean
temperature difference for the insulation material, pipe
September 5
th
to September 7
th

86

shell and surrounding soil is less than 0.005 C for two
consecutive calculations. Thestandard and
advanced model are available online [14]. In the
FEM advanced model we directly implemented
equation (1) in the insulation domain, instead. The
results indicate that in case of low-temperature
operation, lower total heat losses are calculated if the
temperature-dependency of the insulation lambda-
value is taken into account. Moreover the heat transfer
between the pipes in twin or triple pipes can be
properly evaluated.
DN 14 DN 16 DN 20 DN 26 DN 32 DN 40
Standard 3.61 4.24 4.62 5.71 6.45
FEM Standard 3.34 3.68 4.33 4.80 6.02 6.76
Advanced 2.86 3.36 3.69 4.55 5.10
FEM Advanced 3.19 3.51 4.14 4.59 5.75 6.47
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
T
o
t
a
l

H
e
a
t

L
o
s
s

[
W
/
m
]

Figure 6: Comparison of 4 different approaches for steady-
state heat loss calculation. Aluflex twin pipe series,
supply/return/ ground temperatures: 55/25/8 C.

Asymmetrical insulation in twin pipes
The results show that improvements are possible,
thanks to asymmetrical insulation (see Table 3). We
proved that a better design leads to lower heat losses
from the supply pipe (leading to a lower temperature
drop); next, the heat loss from the return pipe can be
close to zero, maintaining isothermal conditions in the
return line. If commercial available casing sizes are
kept, we suggest two design strategies, depending on
the size of the pipes. For small pipe sizes
(Aluflex: DN 26, steel: DN 50) the best design is to
put the supply pipe in the centre of the casing, assuring
the best possible insulation for the supply pipe. This
strategy guarantees also the lowest temperature drop
in the supply side, which is a critical figure in
low-temperature applications.
For bigger sizes (Aluflex: DN 26, steel: DN 50) the
best design is achieved by moving up the media pipe
layout and at the same time by keeping the same
distance between the media pipes as in the
symmetrical case.
Double pipes
A double pipe consists of a pair of media pipes of
dissimilar size, co-insulated in the same casing. It is a
further development of the twin pipe concept. A sketch
of a possible application of the double pipe concept is
shown in Figure 7. Though these measures, network
heat loss reduction is possible, in case of operation
during low heating load periods.

Figure 7: Sketch of the possible application of the double
pipe concept in a simple district heating network.

The space heating demand in summer is diminished,
except for the energy requirement in bath room
heating. According to the energy balance, the reduced
heating load requires less
network flow rate as far as the
designed building temperature
drop is sustained. However, the
reduction of network flow rate
will increase the supply water
temperature drop along the
pipeline due to heat loss. As a
consequence, the supply
temperature at the end user
may lower down below the
minimum requirement. This
problem is relevant to tow-
energy DH systems with an
already low supply temperature.
This design is based on the fact
that the supply line acts also as
re-circulation line during low
Table 3: Comparison between asymmetrical and symmetrical insulation in twin pipes.
The centre of the casing is the origin of the Cartesian system.

Coordinates
(x; y) [mm]
Heat loss
[W/m]
asymm.-
symm. [%]
Size
(DN)
Mat. Sup. Ret. Sup. Ret. Tot. Sup. Tot.
14
Alx.
(0; 0) (0; 27) 3.24 0.01 3.25 -7.6 2.0
16 (0; 0) (0; 28) 3.56 -0.01 3.55 -5.1 1.1
20 (0; 0) (0; 30) 4.16 -0.04 4.12 -4.2 -0.3
26 (0; 0) (0; 36) 4.67 0.00 4.67 -5.1 1.9
32 (0; -16) (0; 28) 5.54 0.00 5.54 -5.8 -2.5
50
Steel
(0; -25) (0; 55) 5.69 -0.03 5.66 -7.7 -2.4
65 (0; -36) (0; 60) 6.70 -0.02 6.68 -7.8 -3.2
September 5
th
to September 7
th

87

heating load periods; hence by-pass at the critical
consumers are not necessary and the exergy loss due
to the mixing of warm water into the return line is
avoided. Furthermore the water flow in the return line
has the same direction as in the supply line (clockwise
in the example), so that the smallest size for the return
pipes are expected in correspondence to the biggest
size for the supply size, and vice versa. This results in
lower local pressure differences between supply and
return lines and savings in operational costs, thanks to
lower heat losses. This is shown in Table 4 and Table
5, by means of two examples: the first one refers to a
small to medium-size distribution network, the second
one to a bigger one, being capable to supply four times
more energy than the previous one.
Table 4: Comparison between a distribution network
based on twin pipe (DN40-40 and DN80-80) with a
distribution network based on double pipe (DN40-80 and
DN80-40). Supply/return/ground temperature: 55/25/8 C.
Heat loss [W/m]
Size
(DN)
Sup. Ret. Tot.
Total
(system)

[%]
40-40 -6.24 0.04 -6.20
Twin:
-13.79
6.1
80-80 -7.66 0.07 -7.59
40-80 -5.55 0.05 -5.58
Double:
-12.94
80-40 -7.41 0.05 -7.36

Table 5: Comparison between a distribution network
based on twin pipe (DN100-100 and DN200-200) with a
distribution network based on double pipe (DN100-200
and DN200-100). Supply/return/ground: 55/25/8 C.

Heat loss [W/m]
Size
(DN)
Sup. Ret. Tot.
Total
(system)
[%]
100-100 -7.83 -0.55 -8.39
Twin:
-17.06
11.8
200-200 -8.92 0.24 -8.68
100-200 -6.4 0.08 -6.36
Double:
-15.05
200-100 -8.07 -0.03 -8.69

We considered an optimal placement of the media
pipes in case of double pipes, thus asymmetrical
insulation is applied. The total amount of insulation is
used both in the twin pipe-based distribution network
and in the double pipe-based one, so that the
investment costs are equal in both cases. Results show
that the heat loss can be reduced by 6% by means of
double pipes instead of twin pipes for the low to
medium-size distribution network. Even higher energy
savings (around 12%) are possible in the case of the
large-size distribution network.
Triple branch pipes
The development of an optimized triple pipe solution for
low-energy applications is reported to show the
potentiality of utilizing detailed models for steady-state
heat loss calculation. In this survey focus was given on
the choice of media pipes diameters as small as
possible. The triple pipe geometry is based on
modifications of the 14-14/110 (outer diameters in [mm]
of respectively supply pipe, return pipe, casing) twin
pipe design which has been reported in [18]. Four
geometrical variations have been considered (see
Figure 8) and the Cartesian coordinates describing the
placement of media pipes inside the casing are listed in
Table 6.

Figure 8: four different geometries for a triple service pipe
type Aluflex 14-14/110.

Table 6: placement of media pipes inside the casing for
four triple pipe geometries, type Aluflex 14-14-20/110.

Variation
Coordinates (x, y) [mm]
Pipe 1
(Sup.)
Pipe 2
(Ret.)
Pipe 3
(Sup. or
re-circ.)
A (14;-14) (0;20.5) (-14;-14)
B (10;-14) (0;20.5) (-21;-7)
C (3;-14) (0;20.5) (-21;-7)
D (0; 0) (0;25) (0;-28)

The results of FEM simulations are listed in Table 7 for
the four geometries (A, B, C, D) and the three
operational modes (I, II, II), previously described. Since
mode II occurs in case of no demand of space heating
and then outside of the heating season, simulations
were additionally performed with a more realistic
temperature of the ground during that period
(14 C),considering Danish weather. This gives also an
insight in the effect of ground temperature throughout
the year.

September 5
th
to September 7
th

88

Table 7: Steady state heat losses of triple pipes type
Aluflex 14/14/110 for 4 geometries and 3 operational
modes. Temperature supply/recirculation/return/ground:
55/55/25/8 C.

Table 8: Steady state heat losses of triple pipes type
Aluflex 14/14/110 for 4 geometries and operational mode
II. Temperature supply/recirculation/ return/ ground:
55/55/25/14 C.

Geom.
Heat loss [W/m]

Pipe
1
Pipe
2
Pipe
3
Tot.
II
(supply-to-
supply re-
circulation)
A 2.35 / 2.35 4.70
B 2.37 / 2.51 4.88
C 2.39 / 2.63 5.02
D 2.20 / 2.42 4.62

We conclude that an absolute best design for the
service triple pipe does not exist, but it depends on the
operational mode that is chosen as critical. In fact the
results reported in Table 7 and Table 8 show that
geometry C gives the lowest total heat loss for
operational modes I and II, while geometry D has the
best thermal performance for operational mode III and
for operational mode II, if a temperature of the soil of
14 C is considered. It has to be underlined that,
considering the operational mode III, geometry D
shows no heating of return water; this is a situation
always desirable, although it has a slightly higher heat
loss from the supply pipe than the other geometries. It
is proved that usually operational mode I occurs for
less than 1 h/day [20]. Moreover the temperature drop
in the supply pipe to the DHW heat exchanger is critical
in low-temperature applications, so that it is strongly
recommended to minimize the heat loss from this
media pipe. Considering all this and the fact that mode
III is the most likely during the heating season and
mode II is the most likely outside heating season, the
conclusion is that geometry D is preferable.
CONCLUSIONS
The soil temperature at 0.5 m below the surface varies
between 2 C in January-February and 14 C in
JulyAugust, for Danish conditions. This knowledge
can be used to better predict the winter peak load and
the temperature drop in the distribution line during
summer.
The slab-model for steady state heat loss calculations
can be replaced, in case of small size
distribution/service pipes, by a model where the effect
of the soil is represented by a circular soil layer around
the district heating pipe.
The results confirm that the vertical placement of twin
media pipes inside the insulation barely affects the heat
transfer, in comparison to the horizontal placement; the
difference between the two configurations is less than
2% for the considered cases.
We proposed a FEM model that takes into account the
temperature-dependency of the thermal conductivity of
the insulation foam; in this way we enhanced the
accuracy of the heat transfer calculation among pipes
embedded in the same insulation.
We applied the model to propose optimized design of
twin pipes with asymmetrical insulation, double pipes
and triple pipes. We proved that the asymmetrical
insulation of twin pipes leads to lower heat loss from
the supply pipe (from -4% to -8%), leading to a lower
temperature drop; next the heat loss from the return
pipe can be close to zero.
It is possible to cut the heat losses by 612% if an
optimal design of double pipes is used instead of
traditional twin pipes, without increasing the investment
costs.
The development of an optimized triple pipe solution
was also reported. It is suitable for low-energy
applications with substations equipped with heat
exchanger for instantaneous production of DHW.
REFERENCES
[1] S. Froning, Low energy communities with district
heating and cooling, 25
th
Conference on Passive
and Low Energy Architecture, Dublin (2008).
[2] S. F. Nilsson et al., Sparse district heating in
Sweden, Applied Energy 85 (2008), pp. 555564.
[3] F. Schmitt, H.W. Hoffman, T. Gohler, Strategies to
manage heat losses technique and economy,
IEA-DHC ANNEX VII, (2005).
[4] P.K. Olsen, B. Bhm, S. Svendsen et al., A
new-low-temperature district heating system for

Mode
Geo
m.
Heat loss [W/m]
Pipe
1
Pipe
2
Pipe
3
Tot.
I
(DHW
tapping)
A 2.67 -0.08 2.67 5.30
B 2.91 -0.29 2.75 5.38
C 2.52 -0.22 2.74 5.06
D 2.46 0.05 2.74 5.24
II
(supply-to-
supply re-
circulation)
A 2.67 / 2.67 5.34
B 2.69 / 2.85 5.55
C 2.48 / 2.70 5.18
D 2.49 / 2.75 5.25
III
(space
heating)
A 3.46 0.48 / 3.95
B 3.39 0.43 / 3.83
C 3.41 0.35 / 3.76
D 3.53 -0.01 / 3.53
September 5
th
to September 7
th

89

low-energy buildings, 11
th
international symposium
on district heating and cooling, Reykjavik (2008).
[5] M. Klompsch, H. Zinko, Plastic pipe systems for
DH, handbook for safe and economic application,
IEA-DHC ANNEX V (1999).
[6] DS-EN 253:2009, District heating pipes Pre-
insulated bonded pipe systems for directly buried
hot water networks - Pipe assembly of steel service
pipe, polyurethane thermal insulation and outer
casing of polyethylene.
[7] H. Zinko, GRUDIS-tekniken fr vrmegles
fjrrvrme (The GRUDIS technology for low heat
density district heating), Swedish District Heating
Association, Stockholm (2004).
[8] J. Claesson, J. Bennet, Multipole method to
compute the conductive heat flows to and between
pipes in a cylinder. Department of Building
Technology and Mathematical Physics, Lund
(1987).
[9] P. Wallenten, Steady-state heat loss from insulated
pipes, Lund (1991).
[10] B. Bhm, On transient heat losses from buried
district heating pipes, International Journal of
Energy Research, 2000, Vol. 24, pp. 1311-1334.
[11] Terminology of HVAC, ASHRAE, Atlanta (1991).
[12] I.B. Kilkis, Technical issues in low to medium-
temperature district heating, International Journal
of Global Energy Issues, 2002, Vol. 17,
pp. 113-129.
[13] J. Korsman, S. de Boer and I. Smits, Cost benefits
and long term behavior of a new all plastic piping
system, DHC ANNEX VIII (2008).
[14] www.logstor.com (March 2010).
[15] Udvikling og demonstration af lavenergifjernvarme
til lavenergibyggeri (development and
demonstration of low energy district heating for low
energy buildings), 2007.
[16] B. Kvisgaard, S. Hadvig, Varmetab fra
fjernvarmeledninger (Heat loss from pipelines in
district heating systems), Copenhagen (1980).
[17] DS418:2002, Calculation of heat loss from
buildings.
[18] H. Kristjansson, F. Bruus, B. Bhm et al.,
Fjernvarmeforsyning af lavenergiomrder (District
heating supply of low heat density areas), 2004.
[19] T. Persson, J. Wollerstrand, Calculation of heat
flow from buried pipes using a time dependent
finite element model, 45th International
Conference of Scandinavian Simulation Society,
Copenhagen (2004).
[20] B. Bhm, H. Kristjansson, Single, twin and triple
buried heating pipes. On potential savings in heat
losses and costs, International Journal of Energy
Research (2005), Vol. 29, pp.1301-1312.
[21] EN 13941:2003, Design and installation of pre-
insulated bonded pipe systems for district heating.

September 5
th
to September 7
th

90

TRANSIENT THERMAL CONDUCTIVITY OF FLEXIBLE DISTRICT HEATING
TWIN PIPES
C. Reidhav and J. Claesson
Department of Civil and Environmental Engineering, Division of Building Technology,
Chalmers University of Technology, Gteborg, Sweden.

ABSTRACT
The standardized methods to measure the thermal
conductivity of straight district heating pipes are not
applicable on flexible district heating pipes. This paper
presents a transient method determining the
temperature dependent thermal conductivity of flexible
twin pipes.
A transient method to determine the temperature-
dependent thermal conductivity of flexible single district
heating pipes is presented in this paper. A flexible pipe
coil is immersed into cold water. Hot water is
distributed in the coil. The temperature decline of the
coil water is measured and calculated. Minimizing the
difference between the calculated and measured
temperatures gives (T) of the flexible polyurethane
foam. The method gives small errors.
INTRODUCTION
District heating is supplied to the customers in one pipe
and returned to the heat generation plant in another
pipe. The two pipes may be placed in separate casings
(single pipes) or in one casing (twin pipes), see Fig 1.
The temperature difference between the district heating
supply (~80-110C) and return temperature (~40-50C)
gives an internal heat flow from the supply pipe to the
return pipe in a twin pipe. The total distribution heat
loss from a twin pipe is lower than that of comparable
single pipes due to this internal heat flow. When
distributing district heat to areas with single-family
houses to heat sparse areas, the issue of distribution
heat losses is of special importance. The relative
distribution heat losses are considerably higher in
sparse areas than in more heat dense areas due to low
heat densities. Flexible district heating twin pipes are
widely used when single-family houses are connected
to district heating systems due to their light weights,
flexibility and long lengths. In the efforts of minimizing
distribution heat losses, the possibility of determining
the insulation capacity of flexible twin pipes is an
important issue.

Fig. 1 Cross-section of single (left) and twin (right) district
heating pipes
The standardised method used for determining the
steady-state thermal conductivity of district heating
pipes, the guarded hot pipe method, is only applicable
on straight single pipes. The method is based on [1],
described in [2] and [3]. A heater pipe is placed inside
the service pipe and the heat transferred through the
insulation is measured. The measurements must be
conducted with a constant distance between the heater
pipe and the service pipe along the test specimen
which can not be achieved with flexible pipes. An
alternative method was presented in [4] and applied in
[5] where the thermal conductivity of flexible district
heating single and twin pipes can be determined. The
temperature decline of hot water pumped in a flexible
pipe coil is measured. A long pipe coil is needed to
have a sufficient temperature decline along the pipe.
The Danish method is based on steady-state
measurements at different temperatures to get the
temperature dependence of the decline. The Finite
Element Method is used to determine the thermal
conductivity (T) of twin pipes.
A transient method to determine the temperature-
dependent thermal conductivity of flexible single district
heating pipes was presented in [6]. A pipe coil is
immersed into cold water and the temperature decline
of hot water inside the coil is measured. The measured
temperatures are compared to numerically calculated
values to characterize (T). A Kirchoff transform is
used to simplify the calculations. Finally, the mean
square difference of the measured and calculated
temperatures are minimized which gives (T). In this
paper, a similar experimental set-up is used for a
flexible twin pipe. The numerical and mathematical
model developed in [6] cannot be used for twin pipes
due to the complicated geometry of twin pipes. In [7]
and [8] a method was presented where heat losses
from district heating twin pipes were calculated with
conformal coordinates describing the twin pipe
geometry. In this paper, the conformal coordinate
model is used to calculate the temperature decline in a
flexible twin pipe. The calculated temperatures are
compared to experimentally measured temperatures.
This gives the temperature-dependent thermal
conductivity of semi-flexible polyurethane foam of the
studied flexible twin pipe.
September 5
th
to September 7
th

91

EXPERIMENTAL MEASUREMENTS
The experimental set-up is similar to that used when
determining the thermal conductivity of single district
heating pipes in [6]. A flexible twin pipe of about 18
meters coiled with a diameter of 1.8 meters is
immersed into a pool with circulating water. In this
experiment, the pool water was about 17 C. Previous
tests show that air is unsuitable as surrounding media
due to difficulties in keeping stable temperatures. The
supply and return service pipes are connected in a loop
circulating water at a temperature of about 80C. When
steady-state is established in the insulation, at time
t = 0, the circulation is stopped. Then, the temperature
decline of the stagnant loop water is measured at one
position in the coil. The thermocouples are placed at
three positions of each service pipe, see Fig. 2. One is
placed on top of the service pipe, one on the side and
underneath the service pipe. The insulation is peeled
off at the positions of the thermocouples and then put
back and sealed to be water proof. The reattachment of
the insulation was probably insufficient and it appears
as if pool water permeated after about 5.5 hours and
disturbed the measurements.

Fig. 2 Experimental set-up and positions of thermocouples
at the service pipes

The twin pipe studied in this paper has two copper
service pipes, semi-flexible polyurethane foam and a
slightly corrugated LDPE casing. The pipe is of
dimension DN 20 with the pipe dimensions described in
Table 1. The pipe producer declares a thermal
conductivity of the semi-flexible polyurethane foam of
50 = 0.0255 Wm-1K-1of a newly produced pipe of
this kind. The pipe has no diffusion barrier. The density
of the polyurethane foam was = 60 kg/m
3
.

Table 1. Dimensions of the twin pipe studied
Casing outer diameter (mm) 91
Casing thickness (mm) 2.2
Service pipe outer diameter (mm) 22
Service pipe thickness (mm) 1.0

The initial coil temperature was T
0
= 81.3C. The water
temperature at the service pipe (T
w,meas
(t)) decreases
during the 16 hours of measurements. The pool
temperature was initially T
1
= 17.4 C and increased
slightly to T
1
=17.9 C during the 16 hours. The
measured coil and pool temperatures are showed in
Fig. 3.
A sawtooth disturbance of about T = 0.07 C and a
small noice of about T = 0.015C can be seen in the
pool water measurements in Fig. 4. A detailed study of
the coil temperatures T
w,meas
(t) shows that the
temperature of some thermocouples decreases abrupt
occurred at about t = 5.5 hours. The marked chosen
interval in Fig. 3 and Fig.4 is chosen to minimize the
errors.
September 5
th
to September 7
th

92

0
10
20
30
40
50
60
70
80
90
0 2 4 6 8 10 12 14 16
Time (h)
T
e
m
p
e
r
a
t
u
r
e

(
C
)
Tw, meas
T1, meas
Chosen interval

Fig. 3 Measured coil Tw, meas and pool temperature T1, meas.

17,4
17,5
17,6
17,7
17,8
17,9
18,0
0 2 4 6 8 10 12 14 16
Time (h)
T
e
m
p
e
r
a
t
u
r
e

(
C
)
Chosen interval

Fig. 4 Measured pool temperature T1, meas

September 5
th
to September 7
th

93

MODELLING USING CONFORMAL COORDINATES
It is rather complicated to calculate the temperature
decline in twin pipes due to the pipe geometry. A so
called conformal mapping presented in [8] was used to
map the twin pipe geometry onto a rectangular
geometry. In the experimental measurements, the
supply and return service pipes were assumed to have
equal temperatures in the test-procedure. Then,
symmetry is assumed between the four quarters of a
pipe cross-section. A quarter of a twin pipe is studied,
see Fig. 5. In the x,y-plane, the temperature
development is described by the heat equation:

( ( ) ) ( ( ) )
T T T
c T T
t x x y y
p
c c c c c
= +
c c c c c
(1)
The (x,y)-coordinates ( ) z x i y = + are transformed to
suitable conformal coordinates ( ) w u i v = + with the
aid of line sources and so called multipoles.

Fig. 5 A quarter of a twin pipe in x-y-plane geometry

The heat equation in the conformal coordinates is:
( , ) ( ( ) ) ( ( ) )
T T T
c A u v T T
t u u v v
p
c c c c c
= +
c c c c c
(2)

Here, A(u,v) is the area factor in the conformal
transformation.
The considered region shown in Fig. 5 is transformed
to a rectangular region in the u, v-plane, see Fig. 6. In
the figure, the left-hand boundary lies against the coil
water and the right-hand boundary against the poll
water. The heat flux in the vertical v-direction is zero on
the horizontal boundaries due to symmetry.
Fig. 6 Initial temperature distribution in the cross-
section of a pipe quarter in the u, v-plane.

In the numerical solution, the region is divided into a
rectangular mesh. The area factor is now the area of
each of the cells shown in Fig.5. They are shown in
Fig. 7. The largest cell is the one in the lower left corner
in Fig.5 near the stagnation point (usp). The areas are
used to calculate the heat capacity of each cell in the u-
v-plane.

Fig. 7 Areas of the computational cells in the x, y plane
transferred to a u, v-plane. The stagnation point is denoted
usp.

The initial steady-state condition for a twin pipe with
coil water temperature T
w
= 81.3C immersed into pool
water at T0=19.7C is showed in u-v coordinates in
Figure 6. Then, the temperature decline of stagnant
water in the twin pipes are calculated
The density and the heat capacity c of the
polyurethane foam are assumed constant in the
temperature interval studied. The boundary
temperatures at the casing are given by the pool
temperature.
The thermal conductivity (T) of the polyurethane foam
is determined by the thermal conductivity at 50C
50

(W/mK) and a coefficient to account for a linear
temperature dependence.
| |
50 50
( ) 1 ' ( ) T T T = + (3)
September 5
th
to September 7
th

94

EVALUATION OF MEASUREMENTS
The temperature-dependent thermal conductivity of the
polyurethane foam is obtained by calculating the
temperature decline of the coil water. Certain values of
the thermal conductivity of the polyurethane are
chosen,
50
and . The actual (T) are obtained by
minimizing the difference between the measured and
calculated coil temperatures, (4).
The heat capacity c (Jkg
-1
K
-1
) of the polyurethane
foam is input to the calculations. Literature references
for the heat capacity of polyurethane foam varies, 1300
Jkg
-1
K
-1
at 50C in [9], 1400 Jkg
-1
K
-1
in [10], 1400-
1500 Jkg
-1
K
-1
for rigid polyurethane foam in [11].
The densities and heat capacities of water, service pipe
and insulation were assumed to be constant in the
temperature interval studied.
The optimal parameter values of
50
and c were
obtained by minimizing the difference D (C) between
the calculated T
w
(C) and measured coil temperatures
T
w,meas
(C).
, , 1 2
max ( ) ( )
w calc w meas
D T t T t for t t t = s s (4)
A certain time interval, 0.5<t<5.5 h, was chosen for the
optimization of the parameters
50
and c. Outside this
interval, the optimization was unstable. This is mainly
due to the disturbances in the measurements probably
caused by penetration of pool water at the position of
three thermocouples. The precision decreases also
with time due to the decreased difference between the
coil temperature and pool temperature. The problems
at the start can be an effect of not having steady-state
conditions before starting the temperature decline.
The maximal values of D for the interval studied are
compared for different combinations of the parameters.
A single optimum point is obtained with the lowest
value of D as illustrated in Fig. 8.

o
o
o
o
o
o
o
c
50
'
D = 0.23
D= 0.43
D= 0.28
D = 0.60
D = 0.43
D = 0.69
c = 900
c = 1500
D = 0.55
50 = 0.43
' = 0.0036
50 = 0.029
' = 0.0072
50 = 0.0285 W/mK
= 0.0054 W/mK
c = 1200 J/kgK
(J/kgK)
(W/mK)
(W/mK)

Fig. 8 The maximum difference D for different
combinations of 50, and c in the chosen interval.The
optimal parameter values giving the lowest D are showed
in the box.
The difference between the calculated and measured
temperatures for the optimal parameter values of
50
,
and c are showed in Figure 9. The saw tooth
disturbance from the measurements is seen. The
difference giving the best fit lies in the interval -0.20 to
0.25 (C). The error is small.

Fig. 9 The difference between the measured and
calculated temperatures for the optimal parameter
values

The optimal thermal conductivities
50
and give the
final result for (T), as shown in Fig.10.

0,025
0,026
0,027
0,028
0,029
0,03
0,031
20 30 40 50 60 70 80
Temperature (C)
(
T
)

(
W
/
m
K
)
Fig. 10 Estimated thermal conductivity of the studied
semi-flexible polyurethane foam
DISCUSSION AND CONCLUSIONS
In this paper, the twin pipe method is evaluated only
from the experiences of one experiment. However, this
experiment is part of a greater context with experiences
from previous experiments on single pipes.
The experimental procedure can be improved. In this
experiment, the reattachment of the polyurethane foam
at the positions of the thermocouples was probably
insufficient. Penetration of pool water lead to a certain
change of slope at about t = 5.5 h. The problem with
penetration could have been solved by measuring the
coil temperature inside the copper pipe instead.
September 5
th
to September 7
th

95

The sawtooth variation seen in the measurements in [6]
is seen in this measurement as well and should be
further investigated.
A large difference between the pool and coil
temperature is desirable to minimize the relative errors.
It is also important to assure that steady-state
conditions are established before starting the
temperature decline.
The final result, the obtained thermal conductivity:
5
( ) 0.0235 10 10 ( 0.0285
50
T T C),
= + = (5)
is in reasonable agreement with the declared
50
=
0.0255 Wm
-1
K
-1
for a newly manufactured pipe. This
pipe piece had been in store for some time and had no
diffusion barrier. The temperature-dependent part of
the thermal conductivity is in well agreement with [12].
REFERENCES
[1] U. Jarfelt, Test apparatus of pipe insulation.
Doctoral thesis. Chalmers University of
Technology, Gteborg (1994)
[2] European standard EN 253:2009, District heating
pipes - Preinsulated bonded systems for directly
buried hot water networks Pipe assembly of steel
service pipe, polyurethane thermal insulation and
outer casing of polyethylene, Brussels, Belgium.
(2009)
[3] European committee for standardization. European
standard EN ISO 8497:1996, Thermal insulation-
Determination of steady-state thermal transmission
properties of thermal insulation for circular pipes,
Brussels, Belgium. (1996)
[4] Danish District Heating Association. Development
of an experimental set-up for measuring the heat
conduction properties of flexible pipes, Project nr.
2006-05, rhus, Danmark. (2006), In Danish,
available at Dansk Fjernvarmes F&U-Konto,
www.danskfjernvarme.dk
[5] District Heating Association (2008), Heat plan
Denmark, Ramboll Danmark A/S and Aalborg
University, (2008), In Danish, available at Dansk
Fjernvarmes F&U-Konto, www.danskfjernvarme.dk
[6] C. Reidhav and J. Claesson, A transient method to
determine temperature-dependent thermal
conductivity of polyurethane foam in district heating
pipes, Building Physics 2008 - 8
th
Nordic
Symposium, Copenhagen, Denmark, (2008)
[7] C. Persson and J. Claesson, Prediction of heat
losses from district heating twin pipes, The 11
th

Cooling, August 31 to September 2, Reykjavik,
Iceland, (2008)
[8] C. Persson and J. Claesson, Numerical solution of
diffusion problems using conformal coordinates.
Application to district heating pipes, Report
Department of Civil and Environmental
Engineering, Chalmers University of Technology,
Gteborg, Sweden (2008)
[9] S. Peng, P. Jackson, V. Sendijarevic, K.C. Frisch,
G.A Prentice, A. Fuchs, Process Monitoring and
Kinetics of Rigid Poly(urethane-isocyanurate)
Foams, Journal of Applied Polymer Science,
(2000) Vol 77, 374-380
[10] R. Zevenhoven, Treatment and disposal of
polyurethane wastes: options for recovery and
recycling, Helsinki University of Technology,
Report TKK-ENY-19, Espoo, Finland, June (2004).
[11] BING, Federation of European Rigid Polyurethane
Foam Associations, Thermal insulation materials
made of rigid polyurethane foam (PUR/PIR),Report
No1 October (2006)
[12] U. Jarfelt and O. Ramns, Thermal conductivity of
polyurethane foam best performance,
10
th
International Symposium on District Heating
and Cooling, Sept 3-5, Hanover, Germany, (2006)
.
September 5
th
to September 7
th

96

DISTRICT HEATING PIPES 200 MM BELOW SURFACE
IN A STREET WITH HEAVY TRAFFIC
Anders Fransson
1
and Sven-Erik Sllberg
2

1
Gteborg Energi AB, Sweden
2
Building Technology and Mechanics, SP Technical Research Institute of Sweden

ABSTRACT
This article reports the results from a field experiment
initiated by Gteborg Energi AB with an extreme
shallow burial of district heating pipes 162/76,1 (DN 65)
casaflex under a street with heavy traffic designed for
an average of 20004000 passes of vehicles a day and
line. The pipes were laid only 200 mm below the
surface. The backfill was of 040 mm particle size.
Several consecutive measurements were done to study
the effects from instant and long term loads from the
traffic. The tests were done on a test pipe prepared
with displacement gauges and on operating pipes.
The aim is that the results will inspire and give input for
making district heating and cooling more cost effective.
The tests showed that both the instant and long turn
deformation of the pipes are small at the actual laying
depth and also that the acceleration in the ground as
heavy vehicles passes does not seem to be alarming.
The conclusion is that shallow burial is technically
possible if the road and backfill is done properly.
1. INTRODUCTION
1.1 New conditions for district heating
The branch of district heating is in need of a new
generation of district heating pipes.
The conditions for selling district heating are slowly
changing due to new legislation, harder competition,
new technique and climate changes. Since 2003
Gteborg Energi AB is connecting more and more
customers but is selling less and less energy. New
legislation from 2006 allowed new buildings in
Gothenburg to use a maximum of 110 kWh/m2
externally supplied energy for heating (,cooling) and
producing domestic hot water. Today the municipality
of Gothenburg wants new buildings to use 60 kWh/m2
at most.
These changes are not unique. New houses are using
less and less energy per square meter. There are
already households that are not using but producing
energy. The former energy suppliers in Europe are
finding themselves not as suppliers but distributors,
buying and selling energy. Climate changes are global
and have already measurable effects on district heating
sales (i.e. less district heating, d.h. to be sold). To
connect new district heating customers in the future,
with the competition of other heating suppliers, it is not
enough to use just smaller pipes because of the
smaller demands. Building the grid and maintaining the
grid needs to become more cost efficient.
The purpose of this article is to inspire and if possible
help whoever is interested in making district heating
and cooling in the world more cost effective using the
ideas or test results from this article.
1.2 Cost-cutting due to shallow burial in roads
When reducing costs, it is important to maintain the
qualities that are required. The road owner needs the
road to be functional and has its standards. The district
heating supplier is responsible for its pipes and
deliveries of heat and has its standards. Finally there
are workers (contractors and maintaining staff) who
need acceptable working conditions.

Fig. 1 Standard shaft section
In a standard shaft section the drainage may be taken
away in roads. A properly built road has a hard top and
is drained as it is. You do not need to drain it any
more.
It is also possible to make the shaft more narrow and
maintain acceptable working conditions if either long
pipes with no joints are used or if the joints are welded
on top of the shaft.
Less coverage is also an alternative. Earlier studies
[1][3] shows that the pipes are solid enough to be
placed with very little coverage (180 mm) and in rough
materials. It is also shown that there is less settling in
September 5
th
to September 7
th

97

the street the more shallow the shaft is [3]. On the
other hand if too little coverage over the pipe is chosen,
it may get hit by a rock curb or a supporting leg from a
truck with no cover.

Fig. 2 Left; Supporting leg with no cover, Right; Rock
curb.
If the existing fraction is used as backfill transports can
be reduced which lower the costs and the
environmental influence.
Normally the district heating pipes have no problems
coping with the traffic load. The extra pressure and the
movement in the soil are making extra loads that are
quite negligible compared to the thermal load, the inner
pressure load and the load from the outer pressure
from the soil.
To get an idea of if the graphite gaskets used in the
pipe joints (for casaflex) can stand the traffic load
200 mm below the surface in a street with heavy traffic
the traffic load was empirically measured in the test
area.
EXPERIMENTAL
The tests were done during 20092010.
The test site
An industrial street classified as a street with
20004000 passes of vehicles for every lane and day
as an average through the year was chosen as the test
site. The extension of the test area was 8 meters as the
pipes crossed a street.
The gauges were placed in one of the lanes close to
the centre of the street. The location of the gauges
were visualised with a cross on the asphalt (see Fig. 4).

Fig. 3 Drawing over the test area.

Fig. 4 Marking of the location of the gauges.
The test shaft
Three displacement gauges were placed in a test pipe
containing air (see section 2.1) beside two operating
pipes (see Fig. 35). Two accelerometers were placed
in separate boxes near the test pipe. The gauges were
monitored through wires at the bicycle path beside the
street.
The test was done with two single district heating pipes
162/76,1 (DN 65) casaflex buried in a fraction with
grain size 040 mm. (Normal standard is a fraction of
sand 0.216 mm.) The distance from the top of the
pipes to the top of the fraction was 60 mm. (Normal
standard is 460 mm.) The distance from the top of the
pipes to the top of the asphalt was 200 mm. (Normal
standard is 600 mm.)
As extra protection, the operating pipes were wrapped
in a grid of polyethylene, PE.

Fig. 5 Shaft section in the test area including the test pipe.
Before laying the asphalt, the fraction was compressed
with a 500 kg plate compactor.
The pipes are designed for 1.6 MPa but the local
hydraulic pressure is approximately 1.4 MPa. The
designed temperature for the district heating water is
110 C. The real temperature varies between 70 and
100 C in the supply pipe and between 40 and 60 C in
the return pipe.
About casaflex
The type of pipe, casaflex, was chosen to overcome
the thermal loads and the working conditions (i.e. the
September 5
th
to September 7
th

98

length 140 m of a single pipe means that there is no
need to work with the pipe in the shaft).
Casaflex is a type of pipe that differs from ordinary
district heating pipes in several ways. It is compared
with ordinary d.h. pipes in earlier studies [4]. This type
of pipe is not particularly common in the Nordic
countries. The pipe is supposed to be used as ordinary
pipes with sand as backfill.
The media pipe is made of corrugated stainless steel
and surrounded by CFC-free polyisocyanurate foam.
The foam is wrapped in a multi layer barrier foil at the
outside covered with a corrugated low density
polyethylene casing. Inside the insulation along the
pipe there are three surveillance wires. The casaflex
pipe can be delivered in very long lengths. The pipes
used in this test were 140 m.
To connect different casaflex pipes a system with
flanges, bolts and gaskets are used. The gaskets are
made of graphite.

Fig. 6 Left; Casaflex pipe, Right; Casaflex pipe with a
joint.
2.1 Deformation of the pipe over time
The test pipe was 1.66 m long and prepared with three
displacement gauges inside to measure the radial
deformation in three directions. The displacement
gauges were installed at the half length of the pipe with
a distance of 100 mm in between. One displacement
indicator measured on the upper side of the pipe
casing, the second on the underside of the pipe casing
and the third at the side of the pipe casing. The
displacement gauges were fixed to the media pipe to
measure the changes in the pipe casing.
Before the test pipe was installed reference
measurements were done at the laboratory to create
zero values for the displacement gauges.

Fig. 7 1.66 m long district heating test pipe of type
Casaflex 162/76.1 (DN 65).

100 100
Indicator 3
Indicator 2
Indicator 1

Fig. 8 Locations of displacement gauges inside test pipe.
During the test period that lasted for one year
indications from the displacement gauges were
measured twelve times. During the test period the
temperature varied between summer temperatures to
winter temperatures.
2.2 Instant deformation of the pipe and
accelerations from traffic load
Two accelerometers were placed 200 mm respectively
600 mm below the asphalt surface (see Fig. 5), close to
the test pipe, to measure the vibrations in the road
structure when heavy vehicles pass over the test pipe.
The measure equipment used were a signal analysator
01dB Harmonie, ser. nr 4227 and accelerometers of
the type ST Microelectronincs type LIS2L02AL with a
sampling rate of 3200 Hz and resonance frequency of
at least 2 kHz. The accelerometers were installed in
small boxes and calibrated within the frequency interval
46 Hz. The calibration is traceable to the Swedish
national centre for acceleration metering.
It was arranged so that a heavy lorry passed over the
test area several times at different speed (20 and 40
km/h) while the vibrations in the road structure were
registered with the two accelerometers. The weight of
the lorry was 26 400 kg.
To investigate the instant deformations in the test pipe
when heavy vehicles pass over the pipe the indications
from the displacement gauges (see section 2.1) were
measured at the same time as the vibrations in the
road structure were registered.
2.3 Radial and axial stiffness of pipe
In laboratory the physics of the test pipe were tested
concerning radial stiffness and axial stiffness. The tests
were done on a 165 mm long test specimen from the
same pipe as the test pipe. The arrow in Fig. 9 shows
the direction of the applied load during the test.

Fig. 9 Arrangement for test of radial stiffness.

September 5
th
to September 7
th

99

The axial stiffness in the pipe was tested in three ways.
Fig. 10 describes the three arrangements for applying
the load in the tests, (a) the applied load acts on the
whole cross section, (b) the applied load acts on only
the steel pipe and (c) the applied load acts on the outer
steel net including the pipe casing. The arrow in the
figures indicates the direction of the load.
90 140
165
(a) (c) (b)

Fig. 10 Three types of arrangement for applying the load
for test of axial stiffness.
2.4 Pipe prolonging while pressurizing
A casaflex pipe does not expand because of the
thermal load. It is self compensating. But because of
the geometry of the media pipe it expands when it gets
pressurized. On the other hand the multi layer barrier
foil in the pipe holds the expansion back. Because the
pipe is flexible, it will still be able to expand, but only
until the multi layer barrier foil stops the expansion.
To see how much the pipe expands because of the
pressurization, a distance indicator, Hilti PD4, was fixed
on the pipe before it was installed and pressurized
while it was still on the ground. The distance was
measured three times against an iron angle which also
was fixed on the pipe. After the pressurization the
distance was measured again three times.

Fig. 11 Left; Fixed distance indicator, Right; Fixed iron
angle.
To see with which force the casaflex pipe was
expanding, the following equation was used:
F
p
= PA (1)
Where Fp is the prolonging force [N], P is the internal
over pressure [Pa] and A is the maximum inner area of
the pipe [m2].
2.5 Leak test of the pipe casing
To discover moisture or even water in the insulation,
there are different indicators on the market.
The typical indicator system used in Gothenburg is the
so called Nordic System. The Nordic System is a
system which is using two naked cupper wires inside
the insulation along the pipe at 10 am and 2 pm.
The casaflex pipe uses the Hagenuk System. That
system uses three wires
a) Ni Cr,
b) Cu, insulated and
c) Cu, not insulated.
In this test different pipes and different systems were
connected. The Ni Cr wire in the Hagenuk System was
left disconnected.
The resistance was measured with an ordinary ohm
meter, BM 400.
The pipes were also three times tested with a, State
meter, Time Domain Reflectometer (TDR) from
Stateview.
2.6 Test of degree of compaction of the street
To get the permission from the road owner to do this
test in the street there were certain standards to follow
[5] and [6].
Before the asphalt could be put on the shaft there were
to be some tests of the degree of compaction of the
street with certain limits. It is a German test that is also
used in Sweden [7]. Basically the soil gets compressed
with a known load over a known area and one
measures the Youngs modulus Ev for the soil two
times. The demands were that;
a) E
v2
/ E
v1
< 2,8
b) E
v2
> 50 MPa
c) At least 4 out of 5 tests should be correct.

Fig. 12 Test of degree of compaction of the street
2.7 Visual control of the surface of the street
As an extra precautionary measure, the street was
optically inspected every month through a year. During
the first month, the street was inspected every week.
And there was an extra inspection in spring in order to
September 5
th
to September 7
th

100

find potential frost action damages. The inspections
were documented with photos.
3. RESULTS
3.1 Test results from deformation of the pipe over
time
The measured pipe deformations during the test period
turned out to be very small. The diagram in Fig. 13
describes the measured changes in the casing since
installation and average air temperatures during the
test period. All three displacement gauges were set to
zero before the installation. The diagram shows that
the casing of the test pipe during the installation was
squeezed out up to 0.5 mm at the three measurement
points. The deformations in the casing are most likely
caused by the packing of the backfill surrounding the
pipe.
After the installation during the test period the results
indicate that the upper side (violet curve in the
diagram) of the test pipe casing have been pressed in
0.2 mm. The side of the test pipe casing have
squeezed out approximately 0.1 mm. The under side
(red curve) was squeezed out approximately 0.1 mm
during the period between the first and second
measurement results. During the rest of the test period
the casing have been pressed back in 0.1 mm.
It is to be observed that these measured changes are
very small relative to the test pipe casing diameter.
Compared to the zero values in the laboratory the
measured changes are not more than 0.3 % relative to
the casing diameter.
-100
-80
-60
-40
-20
0
20
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
0 100 200 300 400
m
m
Days since installation
Averagetemperatures (C)
Under Side
Upper Side
Side

Fig. 13 Average air temperatures and changes in casing
at installation and during test period.
3.2 Test results from instant deformation of the
pipe and accelerations from traffic load
The diagrams in Fig. 14 and 15 describe the vibrations
process at 200 mm, the same depth as the test pipe,
and 600 mm below the road surface as acceleration
(m/s2) in the ground when a heavy lorry pass over the
test area at a speed of 40 km/h.

In Fig. 16 and 17 the diagrams show the vibration
velocity (m/s) in the ground when a heavy lorry pass
over the test area at a speed of 40 km/h. The vibration
velocity is calculated from the acceleration signal by
integration.
The diagrams in Fig. 18 and 19 show the maximum
amplitude of the acceleration in the ground as a
function of the speed of the lorry when it passes over
the test area in 20 km/h and 40 km/h, respectively the
maximum vibration velocity as a function of the speed
of the lorry.

200 mm below surface, vehicle speed 40 km/h
Time (s)

Fig. 14 Vertical acceleration 200 mm below the road
surface when a lorry passes at 40 km/h.

Time (s)

Fig. 15 Vertical acceleration 600 mm below the road
surface when a lorry passes at 40 km/h.

V
e
l
o
c
i
t
y
(
m
m
/
s
)
Time (s)

Fig. 16 Vibration velocity 200 mm below the road surface
when lorry passes at 40 km/h.

September 5
th
to September 7
th

101

V
e
l
o
c
i
t
y
(
m
m
/
s
)
Time (s)

Fig. 17 Vibration velocity 600 mm below the road surface
when lorry passes at 40 km/h.

Fig. 18 Maximum amplitude of acceleration as a function
of speed.

Fig. 19 Maximum amplitude of vibration velocity as a
function of speed

The acceleration of the ground increases with the
speed of the traffic. And the effect is more sensitive the
closer you are to the surface (see Fig. 18).
The vibration velocity also increase with the speed of
the traffic. The effect is not as sensitive as for the
acceleration when it comes to the coverage (see
Fig. 19).
In Fig. 20 the diagram describes the instant
deformations in the test pipe when the heavy lorry
passes over the test area at a speed of 40 km/h. It can
be seen from the red and the violet curves that the pipe
casing is squeezed together approximately 0.17 mm
from top to bottom at the instant when the lorry passes
over the test area. At next instant it is squeezed
together approximately 0.07 mm at the same time as
the blue curve indicates that the pipe goes eccentric
approximately 0.04 mm.
-0,06
-0,04
-0,02
0,00
0,02
0,04
0,06
0,08
0,10
0,12
0,00 1,00 2,00 3,00 4,00
m
m
time (s)
Side (mm)
Up (mm)
Down(mm)

Fig. 20 Instant deformation 200 mm below the road
surface when lorry passes at 40 km/h.
3.3 Test results from radial and axial stiffness of
the pipe
The test pipe was compressed 1.8 mm two times with a
feed speed of 1 mm/min. In Fig. 21 it can be seen that
the maximum force at 1.8 mm turned out to be 1.4 kN.
Using this result to look at what the corresponding
forces should be in the test when a heavy lorry passes
over the test area (see Fig. 20) it can be established
that the instant forces from passing vehicles is small.

0
200
400
600
800
1000
1200
1400
1600
0 0,5 1 1,5 2
L
o
a
d
,

N
Deformation, mm
Test 1
Test 2

Fig. 21 Diagram radial stiffness of a casaflex pipe.

Feed speed 3 mm/min
0
2
4
6
8
10
12
14
16
18
20
0 0,5 1 1,5 2
A
x
i
a
l

l
o
a
d
,

k
N
Axiell compression %
Load case (b)
Load case (c)
Load case (a)

Fig. 22 Axial stiffness of a casaflex pipe
September 5
th
to September 7
th

102

3.4 Test result of pipe prolonging while
pressurizing
The test results are as follows:
Table 1. Test result of pipe prolonging while pressurizing
Tests
1 2 3 Average
[mm] [mm] [mm] [mm]
Distance
before
pressurizing
4 360 4 360 4 360 4 360
Distance after
pressurizing
4 365 4 365 4 365 4 365

The pressure that was used was approximately
1400 kPa. The test results show that a single pipe
casaflex prolongs itself 100*5/4360 = 0,11%.
his could be compared to the more common steel pipe
for district heating. If that pipe would be loaded with a
thermal load of 100 C it would prolong itself 0,12%.
The force with which the casaflex pipe is expanding
because of the inner pressure would for 11400 kPa be
7,3 kN according to the supplier. That would mean that
the diameter would be 81,5mm. In real life the diameter
was measured to be 83,9 mm. The corresponding force
for the diameter 83,9 mm would be 7,7 kN.
If the pressure would have been 1 600 kPa and the
diameter would have been 81,5 mm then the
corresponding force would have been 8,3 kN.
This could again be compared to the steel pipe with the
thermal load of 100 C. This pipe would prolong itself
with the force of 164,9 kN.
So the casaflex pipe expands with a force that is
approximately 100*8,3/164,9 = 5,0% of the force from a
steel pipe when heated 100 C.
Looking at Fig. 22, case a), one sees that the inner
force (axial load) that the expanding force has to
overcome is negligible.
3.5 Test results from the leak test of the pipe
casing
Different TDR graphs have been made in May 2009, in
June 2009 and in April 2010.
Through metering the resistance and making TDR
graphs it is proven that:
a) It can be done to connect the two different
systems (The Nordic System and the Hagenuk
System).
b) There are no leaks in the test area, neither in
the supply pipe nor in the return pipe.

The diagram (Fig. 23) below contains two different TDR
measurements. It is one graph per wire and test. If
there are no changes in the impedance there are no
changes in the profile in the graph. And there are no
changes in the profiles.

Fig.23 TDR graph for the supply pipe.
3.6 The results from the test of degree of
compaction of the street
The different tests were plotted in diagrams and gave
the different Youngs modulus E
v1
and E
v2
for different
places. The places were documented in a photo. The
results can be read in the table below.
Table 2. - Results from test of degree of compaction

Spot Ev1
(MN/m
2
)
Ev2
(MN/m
2
)
Ev2/
Ev1

1 38,75 91,44 2,36
2 18,23 25,25 1,39
3 61,29 126,4 2,06
4 51,53 100,65 1,95
5 29,30 70,08 2,39

For every test the division E
v2
/E
v1
are approved. Spot 2
E
v2
is to low but the other four spots are approved so
overall the test is ok.
3.7 Results from the visual control of the surface of
the street
The strength of the road
There was no change in the surface of the street due to
the shallow laying of the district heating pipe what so
ever the first eleven months. In spring after an
unordinary cold winter one could see a small crack
(approximately 12 cm) in the street along the pipes
extension. As this article is getting written it is not
investigated why the crack has appeared nor of the
importance of it. The street has much worse injuries
from frost action damages outside the test area.
September 5
th
to September 7
th

103

Heat loss
The snow did not melt over the pipes in the test area. If
one would study other older district heating pipes they
would reveal themselves by melting the snow over
them. This effect never happened in the test area.

Fig. 24 Left; Test area in January, Right; Test area in
February.

Fig. 25 Left Test area in March, Right; D. H. chamber
revealing itself in February.
However the heat loss is of course bigger compared
with normal standard because of that the pipes are
placed closer to the air.
Frost action damages
In theory one could imagine that the street on both
sides of the district heating pipe would erect during the
winter if there were soil that could frost heave. This
could of course damage the asphalt. But streets are not
supposed to be built with soil that could frost heave. So
there should not be any problem.
There was no notable difference in the height of the
street over the district heating pipes compared to the
street beside the test area during the winter.
4. CONCLUSION
The article probably describes the first operating d.h.
pipes placed in backfill of 0-40 only 200 mm below the
surface in a street with heavy traffic. As expected, the
pipes are working nicely. The loads that have been
measured are acceptable or even low for the d.h. pipe.
As it seems also the street is satisfactory working even
though there are d. h. pipes close to the surface.
The pipe deformations are negligible with respect to the
pipes function both over time and under an instant
traffic load.

The acceleration and vibration velocity are also
negligible under the traffic load from e.g. a heavy lorry.
Probably there will be no problems using graphite
gaskets also with only 200 mm of coverage.
The casaflex pipe is prolonging itself if it may but the
force with which it is prolonging itself is but a fraction of
what comparable ordinary steel pipes uses. This effect
makes it more suitable for shallow shafts.
It is possible to combine different leak indicator
systems and still get the TDR-graphs. The graphs done
in this test indicates that there are no leaks in the
operating pipes after one year.
The demanded levels for the degree of compaction of
the street are possible to reach also with a d.h. pipe
60 mm below the surface as it gets compressed
without hurting the pipe.
There are still other issues that can be considered that
are not included in the tests presented in this paper,
aspects as e.g. heat losses.
5. ACKNOWLEDGEMENT
The authors would like to express their appreciation to
a couple of key persons. There had been no test of
this kind without their support and permitting. The
persons are:
Mr. Bo Andersson Planing Manager at
Trafikkontoret Gteborgs Stad,
Mr. Lars Ljunggren Manager at Gteborg
Energi AB and
Mr Gran Johnsson Technical Manager at
Powerpipe Systems AB.
6. REFERENCES
[1] Molin J., Bergstrm G. and Nilsson S. (1997).
Kulvertfrlggning med befintliga massor, Swedish
District Heating Association FOU 1997:17, (in
Swedish)
[2] Bergstrm G., Nilsson S. and Sllberg S-E. (2001),
Tthet hos skarvar vid terfyllning med befintliga
massor, Swedish District Heating Association FOU
2001:58, (in Swedish)
[3] Nilsson S, Sllberg S-E, Bergstrm G, (2006)
Grund frlggning av fjrrvrmeledningar, Swedish
District Heating Association FOU, FOU Vrmegles,
2006:25, (in Swedish)
[4] Gudmundson T. F-Processdesign AB, (2002),
Casaflex-rr i Malm 2001,. Swedish District
Heating Association, FVF 021241, (in Swedish)
September 5
th
to September 7
th

104

[5] Trafikverket, (2005) Allmn teknisk beskrivning fr
vgkonstruktion ATB Vg,
http://www.vv.se/Startsida-foretag/vagar/Tekniska-
dokument/ATB-Allmanna-tekniska-
beskrivningar/Vagteknik/Aldre-versioner/ATB-Vag-
2005/, visited 2010-04-27, (in swedish)
[6] Gteborg Stad Trafikkontoret, Bestmmelser fr
arbeten inom gatu- och spromrden i Gteborg,
http://www.goteborg.se/wps/portal/!ut/p/c0/04_SB8
K8xLLM9MSSzPy8xBz9CP0os3gjU-
9AJyMvYwMDSycXA6MQFxNDPwtTo2Anc_2CbE
dFABCTfUM!/?WCM_GLOBAL_CONTEXT=/wps/
wcm/connect/goteborg.se/goteborg_se/Foretagare/
Upphandling_staden%20som%20kund/Specifik%2
0upphandlingsinformation/art_N400_FOR_Up_SU
_Trafikkontoret, visited 2010-04-27, (in swedish)
[7] Trafikverket, (1993) Publikation 1993:19
Bestmning av brighetsegenskaper med statisk
plattbelastning Metodbeskrivning 606:1993, (in
swedish)
September 5
th
to September 7
th

105

STUDY ON THE HEAT LOSS REDUCTION METHOD FROM THE SECONDARY
PIPELINES IN THE APARTMENT COMPLEX
Byung-Sik Park
1
, Yong-Eun Kim
2
Sung-Hwan Park
1
, Yong-Hoon Im
1
, Hyouck-Ju Kim
1
, Dae-Hun Chung
1,
Mo Chung
3
1
Building Energy Research Center, Korea Institute of Energy Research,
102 Gajeong-ro, Yuseong-gu, Daejeon 305-343 KOREA, bspark@kier.re.kr
2
Energy System Engineering, University of Science and Technology,
113 Gwahangno, Yuseong-gu, Daejeon 305-333 KOREA, rainyday@ust.ac.kr
3
Dept. of Mechanical Engineering, Yeungnam University,
214-1 Dae-dong Gyeongsan-si Gyeongsangbuk-do 712-749 KOREA

ABSTRACT
This study aims to suggest better methods for reducing
heat losses from the pipelines installed as secondary
heating pipes in the apartment complex in which hot
water is being supplied for space heating and hot water
by a district energy supply company. Right now the
district heat supplier is responsible only for the primary
district heating pipelines just before the substations in
the apartment complex. That is why the heat loss
reduction becomes more important in the secondary
pipelines after the substation in the Korean apartment
complex.
Several methods to reduce the heat loss from the
secondary pipelines were set up and compared by a
simulation technique. One of the methods is to
combine the hot water heating pipes and space heating
pipes. Another method is to install a small heat
exchanger in each house to supply hot water from the
single space heating pipeline. In this case we can
easily change the means of heat supply and the right
choice of end users can be ensured for the means of
heat supply.
In this study the preferable method to reduce the heat
loss in the secondary pipelines has been suggested.
The simulation result has shown about 30% heat loss
reduction compared to the existing scheme for the
simple change of methods and much more reduction
for the optimization of pipe diameter and insulation
thickness or surface enhancement by low emissivity.
INTRODUCTION
Korea is characterized as having four distinct seasons.
Apartment complexes became a typical type of
residence in urban areas after the recent rapid
industrialization of last 30 years. At the moment over
half of the population chooses to live in apartments
rather than in individual houses and the trend will
increasingly continue in the future. There are three
typical heating methods for apartment complexes -
individual heating, central heating and district heating.
At the moment there is a lot of potential for district
heating and cooling. Korea has seen about a 10%
supply of DHC among total residential houses which is
very low compared to that of European countries which
supplies over 50% DHC.
If we are to increase green growth with low carbon, it is
crucial to supply DHC, which has higher energy
efficiency than any other method, in dense regions of
population. The recent Korean government has shown
effort in making a point of energy efficiency throughout
main energy consuming sectors including building
area. However, the supply policy of DHC is now being
crippled due to various reasons. Makers or consumers
of individual heating devices do not have positive
attitudes toward DHC. Therefore it is important to draw
attention to the multitude of benefits and merits of
DHC.
There is certainly some heat loss from the pipelines
installed under the ground to supply the district energy
from the power plant to the consumers. To reduce the
heat loss from these primary pipelines many
innovations and advancements have been made for a
long period since the district energy was supplied in the
northern European countries. The heat loss generally
differs according to the network type of pipelines. The
more compact that the network is, the less heat loss
occurs. But the type of network cannot be made
arbitrarily by the designer. The designer can simply
optimize the network in view of geological and
environmental conditions, such as population density,
not the type of network. Although heat loss exists with
the primary pipelines, it can be controlled and
maintained effectively by the district heat supplier. On
the other hand, the heat loss from the secondary
pipelines cannot be controlled properly by the building
owners who are responsible for.
The apartment complex is a unique housing system in
Korea. It contains many high rise buildings of over 10,
often over 20, stories high. In many cases it has over
September 5
th
to September 7
th

106

one thousand homes. However, the basic structure is
almost the same as that of western apartment buildings
except for the pipeline network between buildings and
substation. In the past there was one substation in one
apartment complex. The substation has a minimum of
two heat exchangers which are in general shell and
tube type or plate type. Nowadays the number of
substations grows bigger and bigger. That means that
the designer plans to install the heat exchanger
separately and respectively according to the buildings
which stand nearby each other. The secondary
pipelines have been said to have much heat loss in
Korea. There have been a few studies related to heat
loss from the secondary pipelines. It is very hard to
distinguish between positive heat gain and heat loss
from the pipelines installed within the buildings. If the
pipelines are installed in the center of the building, the
heat loss from the heating pipes or hot water supply
pipes can be regarded as positive heat to the
consumer. But if the pipelines are installed near the
building surface, the heat from the pipes can be
regarded as loss.
Measurement and analysis of the heat loss from
several apartment complexes in Korea has been tried.
The heat loss data from the several sites has been
stored and accumulated throughout the year. A
simulation method has been set up and the accuracy of
the simulation has been investigated. Some
alternatives to reduce the heat loss have been
prepared from the existing scheme. The simulation
method and results have been presented in this paper.
TYPE OF APARTMENT COMPLEX
The apartment complex was built and opened in
November 2007. It has 8 buildings which are
comprised of 518 homes. Each home has 112 m
2
of
heating area. Fig. 1 shows the location and overall
shape of the apartment complex which was chosen
CHINA
JAPAN
North
KOREA
South
KOERA
Fig. 1 The location and the shape of the apartment
complex
to be measured and evaluated on the heat loss from
the secondary pipelines. Many thermocouples and
flow-meters were installed in the region of the pipelines
to collect information on heat demand pattern,
temperatures and heat loss from certain regions to be
evaluated.
Supplementary
Water
DHWS
DHWR
204
206
208
203
205
207
202
201

City water
H
W
S
M
Calorimeter
H
W
R
H
W
R
H
W
R
H
W
R
H
W
S
H
W
R
H
W
R
H
W
R
H
W
R
H
W
S
H
W
R
H
W
R
H
W
R
H
W
R

Fig. 2 Secondary pipeline network from the substation

ANAYSIS OF HEAT CONSUMPTION PATTERN
1) Space heating water flowrate
Space heating amount is being measured daily. Toom
temperature does not differ much between the homes
in the apartment complex. Thus the temperature
difference (T) between inlet and outlet of the pipeline
of individual homes remains fairly constant except in
the summer season. Therefore the heating water flow
rate can be estimated from the following equation
Q=CmT. In other word, the flow rate could be
evaluated from the measured calorific amount. A good
example of this is shown in Fig. 3.
Fig. 3 Temperature difference between supply and return
In this study the flow rate was measured for the two
months of November and December 2009 using the
flow meter installed in the space heating water pipeline.
And the hourly heating water flow rate of individual
homes for the year 2009 was extracted from the
comparison of the total measured amount and the
individual house measurement. Fig. 4 represents the
annual heating water flow rate. Some differences exist
during the cold winter season.

September 5
th
to September 7
th

107

Fig. 4 Heating water flow rate

2) Hot water flow rate
Hot water consumption is measured by ton from the
general water flow meter. Hot water supply line is
designed to have a supplementary recirculation line in
order to supply instant hot water. By adding the water
which was used by individual homes to the heat
exchanger, the flow rate can be constantly maintained.

Fig. 5 Hot water flow rate

3) Various temperatures outside pipes
The temperatures outside the pipeline were
categorized into four cases taking into consideration
the atmosphere outside the pipeline. The first one is
the underground space which is fairly open to the
outside of the building. The second one is the
underground space which is not so open to the outside
of the building. The other two are the spaces of low and
high regions of the building which is not open to the
outside of the building. These temperatures, measured
according to the categories, were applied in the
simulation in view of the pipelines outside
characteristics.

1st floor
2nd floor
3rd floor
4th floor
5th floor
6th floor
7th floor
8th floor
9th floor
10th floor
11th floor
12th floor
Secondary pipe line Primary pipe line
DHWS
DHWR
T base T sb
T low
T high
T out

Fig. 6 Category of temperature charicterization.

Tout: Temperature outside the building
Tbase: Temperature of the underground space
frequently open to the outside surrounding.
Tsb: Temperature of the underground space closed
to the outside surrounding.
Tlow: Temperature of low-rise region in the building
Thigh: Temperature of high-rise region in the building

Fig. 7 Varous temperature of the surroundings of
pipelines.

SIMULATION METHOD
For the heat loss simulation commercial tool,
Flowmaster of 1 D system analysis has been used.
Flowmaster is a program which can analyze the
thermo-hydrodynamic characteristics of pipe systems if
the following items are given such as the physical
properties of pipes, flow rate and outside temperature
through the following equations. Annual heat loss can
be simulated by using the information such as
measured temperatures, flow rates and various
physical properties of pipes and insulation materials
according to the drawing of all the pipelines which are
September 5
th
to September 7
th

108

installed underground between buildings and
substation as well as in the buildings themselves in the
apartment complex. The simulation was performed
daily in view of calculation time.

(1)
(2)
(3)

fluid temperature, C

ambient temperature, C

convection heat transfer coefficient,

radiation heat transfer coefficient,

internal heat transfer coefficient,

insulation thermal conductivity,

pipe thermal conductivity,

external pipe diameter,

external insulation diameter,

internal pipe diameter,

For accurate simulation, individual flow rate was used
respectively and differently based on total measured
flow rate and read amount of individual flow meter of
518 homes throughout the year. By doing this, the flow
rate in the individual pipelines can be determined
according to the usage amount of heating water and
hot water. This similarly leads to actual flow rate in the
pipelines. From this complicated process, the number
of individual flow rate of hot water and heating water
comes to 378,140. Macro which was combined with
Excel and Flowmaster was used for a 365 day
analysis. This process requires 32 hours for 8 buildings
for only one case.
VARIOUS SCHEMES FOR SIMULATION
1) Present Scheme
1st floor
2nd floor
3rd floor
4th floor
5th floor
6th floor
7th floor
8th floor
9th floor
10th floor
11th floor
12th floor
DHWS
DHWR
HWR
HWS
HWR
HWR
Heating water supply
Heating water return
Hot water supply
Hot water return
HWR
HWS
HWR
HWR
H
W
R
H
W
S
H
W
R
H
W
R

Fig. 8 Pipeline configuration of present scheme

The present scheme is composed of 4 pipelines, two of
which are for heating water supply and return and the
other two of which are for hot water supply and return.
Heating water is supplied to each home and returned
from each home and resultantly the flowrate of supply
and return are equal. On the other hand hot water has
a certain amount of recirculation in order to keep
supply water hot. Hence the same amount of
consumed hot water in each individual home should be
supplied to the hot water heat exchanger to guarantee
instant hot water supply.
2) Alternative A
Alternative A is a scheme which removes the hot water
pipelines and combines with the heating water
pipelines. Thus there are only two pipelines of supply
and return from substation to each building. These
supply and return pipelines have two functions of
heating and hot water supply and return. By reducing
from 4 to 2 pipelines the heat transfer surface area can
be decreased. However, this scheme has
disadvantages in summer when the heating water
supply has been closed. If the pipelines should be used
for the supply of hot water, the resultant water speed in
the pipelines would be very small.
1st floor
2nd floor
3rd floor
4th floor
5th floor
6th floor
7th floor
8th floor
9th floor
10th floor
11th floor
12th floor
H
W
R
H
W
R
E-2
HWR
HWS
P-2
DHWR
H
W
S
HWR
HWS
HWR
HWR
Hot water supply
Hot water return
P-4
P-3
P-1
P-6
E-1
E-3
P-9
DHWS

Fig. 9 Pipeline configuration of Alternative A
September 5
th
to September 7
th

109

For hot water, this scheme is effective in heat
exchange because the heat exchanger acts for the
consumed hot water only and can avoid extra
recirculation pipelines. But it has the drawback of
supplying cold water or non heated water when
intermittently using hot water.
3) Alternative B
1st floor
2nd floor
3rd floor
4th floor
5th floor
6th floor
7th floor
8th floor
9th floor
10th floor
11th floor
12th floor
H
W
R
E-1
HWR
HWS
P-4
DHWR
H
W
S
P-1
E-4
E-5
E-6
E-8
E-7
E-9
E-13
E-12
E-14
E-11
E-10
E-15
DHWS
HWR
HWS Heating water supply
Fig. 10 Pipeline configuration of Alternative B

Alternative B is the same as Alternative A in the point
of unifying the heating water and hot water pipelines.
But it is different in the point of the individual hot water
heat exchanger being installed in each home among
the heating water pipelines. This scheme is said to
resemble the pipelines installed in the apartments of
European countries.

4) Alternative C
1st floor
2nd floor
3rd floor
4th floor
5th floor
6th floor
7th floor
8th floor
9th floor
10th floor
11th floor
12th floor
H
W
R
H
W
R
HWR
HWS
DHWR
H
W
S
P-7
P-8
H
W
R
P-9
P-14
E-5
DHWS
HWR
HWS
HWR
HWR
Hot water supply
Hot water return

Fig. 11 Pipeline configuration of Alternative C

Alternative C is a variation of Alternative A to make up
for the defect of cold water supply when intermittently
supplying hot water. In this alternative the recirculation
pipelines are equipped in the buildings.
5) Use of PEX PIPES
PEX is being used in western countries as district
heating pipes for low temperature service from
renewable resource application. For this reason PEX
can be used as secondary pipelines which normally are
under service of low temperature.
VALIDATION OF SIMULATION RESULT
To validate the simulation result, the measurement
value was compared with the prediction result by the
present simulation method. The measured value of
December 2009 was used for heat loss reference
value. As seen in Table.1 the simulation result fairly
agreed with the measured data. And the simulation
method can be used without much modification. For
more accurate prediction it needs slightly more
supplementation in the numerical modelling of heat
transfer phenomena of outer pipe surface and
environment.
Table1 Comparison of measured and predicted
values[unit: MWh]
Measurement heat Hot water
Heat loss rate
9.44%
loss 20.2 11.2
supply 264.7 67.8
Simulation heat Hot water
Heat loss rate
9.90%
loss 17.0 14.29
supply 252.6 63.53
SIMULATION RESULT
Fig. 12 shows typical heat supply and heat loss for the
24 hours of 11.11.2009. The simulation result of each
scheme is shown from Fig. 13 to Fig. 17. Each graph
shows similar patterns and the heat loss comparison of
each scheme is summarized in Table 3.

Fig. 12 Hourly heat supply and heat loss of 11.11.2009
September 5
th
to September 7
th

110

Fig. 13 Annual heat supply and heat loss (Present
scheme)

Fig. 14 Annual heat supply and heat loss (Alternative A)

Fig. 15 Annual heat supply and heat loss (Alternative B)

Alternative B should supply heating water in the
summer season when it is not required for the supply of
heating water in order to supply hot water to the
individual home. From the comparison of Fig. 14 and
15, Alternative B is more efficient than Alternative A in
the cold region.

Fig. 16 Annual heat supply and heat loss (Alternative C)

Fig. 17 Annual heat supply and heat loss (PEX Pipe)

The heat loss comparison of each scheme can be
summarized in Table 2.
Table 2 Heat loss comparison of each scheme[unit:
MWh]
Supply Loss Heat loss rate
Present scheme 4863.7 681.4 14.01%
Alternative A 4477.6 456.3 10.19%
Alternative B 4317.1 516.3 11.96%
Alternative C 4929 523 10.61%
PEX Pipe 4691.2 508.9 10.83%

HEAT LOSS COMPARISON DUE TO THE CHANGE
OF INSULATION THICKNESS,NOMINAL DIAMETER
AND INSULATION MATERIAL
Table 3 shows a comparison among parameters which
affect heat loss. There is frequent excessive design for
the nominal diameter of pipelines which are installed in
both the underground and the buildings. In this
comparison the increases of insulation thickness by
10 mm and 20 mm were considered. Also the decrease
of nominal diameter by 1 level was considered. It was
taken into consideration of insulation material change
and each combination of affecting parameters. In this
comparison, the present popular design method of
pipelines and insulation was regarded as a reference
for 100% of heat loss and other alternatives were
evaluated from the reference heat loss relatively.
Table 3 shows the result of relative comparison of
annual heat loss from the pipelines of the apartment
complex.

September 5
th
to September 7
th

111

Table 3 Heat loss comparison due to various
parameters

Heat loss comparison
Present pipe of insulation thickness 40 mm 100%
Downsizing nominal diameter by 1 level 88.7%
Insulation thickness 50 mm 86.1%
Insulation thickness 60 mm 75.5%
Closed-Cell Elastomeric thermal insulation 89.3%
Pipe diameter downsizing + insulation
thickness 50 mm
76.7%
Pipe diameter downsizing + insulation
thickness 60 mm
67.6%

CONCLUSION
Present scheme for the secondary pipelines is
evaluated to have 14% annual heat loss based on the
total heat supply to the apartment complex. This is a
very large amount when we consider that the primary
district heating pipeline has only about 4 to 5% annual
heat loss in dense population urban areas.
Heat loss by Alternative A can be reduced about 30%
compared to that of the present scheme which has
been widely adopted in Korea until now. However,
Alternative B has more heat loss compared to that of
Alternative A, which was not the common expectation.
The main reason was the increase of the pipe
insulation surface area which acts as a heat transfer
area. Alternative C and PEX system can be
replaceable when they have merits in the point of initial
cost.
More enhancements in heat loss can be extracted from
heat loss reduction by the selection of optimum pipe
diameter, good insulation material, increasing
insulation thickness and changing surface emissivity of
insulation material.
ACKNOWLEDGEMENT
The financial support from KDHC made this work
possible. This paper is based on the results of an
ongoing research project which will be completed at the
end of 2010.
REFERENCES
[1] W. F. STOEKER, DESIGN OF THERMAL
SYSTEMS 3rd edition, Mc Graw Hill, pp. 53-160
[2] Incropera, HEAT TRANSFER 5th,WILEY
[3] Byung-sik Park et al, Study on the Reduction
method of Heat Loss from the Secondary Pipelines
installed in the Apartment Complex, 2008
[4] Byung-sik Park et al, Study on the Reduction
method of Heat Loss from the Secondary Pipelines
installed in the Apartment Complex, 2009
[5] Flowmaster, Flowmaster heat transfer manual
[6] Manfred Kl psch, Plastic pipe system for DH,
Handbook for safe and economic application, IEA
R&D Programme on District Heating & Cooling

September 5
th
to September 7
th

112

HEAT LOSS OF FLEXIBLE PLASTIC PIPE SYSTEMS,
ANALYSIS AND OPTIMIZATION
EJ.H.M. van der Ven
1
, R.J. van Arendonk
2

1
Thermaflex International Holding B.V
2
Liandon B.V.

ABSTRACT
A newly developed, in-house, test rig for measuring
heat loss of pipe systems allows the user to analyse
various systems in a short timeframe. This allows quick
insights into heat loss variables and mathematical
analyses. The effect of alternative compositions of
insulation and other layers can be evaluated within a
short time span.
This already led to improvements of the production
process and of the insulating foam.
INTRODUCTION
Liandon developed a heat loss testing rig for
Thermaflex to test their produced flexible pipe systems.
Within a short time the pipe system, undergoing a heat
loss test, tends towards the controlled temperature in
the sections of the sample, the added power reaches
equilibrium and the test results can be collected.
Due to the short time required for testing, the results of
alternative production methods are easily available.
Due to the short response time the test is a great help
in the search for product and production improvements.
The objective of this paper is to present the results of
the research to the overall heat loss performance of a
flexible plastic pipe product, Flexalen 600.
The objective of the research is:
1 Find correlations between heat loss and other
parameters of the pipe system such as outer
diameter, inner diameter, foam surface and foam
structure. These correlations are determined by the
mathematical analysis of practical heat loss
measurements.
2 Find possibilities for the improvement of the pipe
parameters by analysing the heat loss correlations.
NOVELTY AND MAIN CONTRIBUTION
The actual heat loss of pre-insulated pipe products is
determined under controlled, similar conditions for an
entire diameter range. This range comprises various
outer diameters, various inner diameters and various
compositions in materials and pipe systems. The time
required for one single test run varies from half an hour
for a small-sized pipe to about eight hours for the
largest sized diameter.
The novelty of the test rig is described in the paper
Verification of heat loss measurements conducted on
(semi) flexible pipe systems [3].
The novelty for product improvement is that due to the
reduced time required for a test run the effect of
alternative systems can be mathematically analysed
and evaluated in a short time. In this way the analyses
of alternative production methods has a short
feedback. Optimization of the product can be effected
in a short time.
In the near future the test rig will be used for quality
control of the production process. This test will partly
replace other currently applied standard tests, such as
density and cell size measurements.
METHOD DESCRIPTION
In the Flexalen 600 pre-insulated pipe a PB medium
pipe is encapsulated in insulating foam, which is
protected against wear and tear in a corrugated hard
cover pipe. The pipe product has a solid bonding
between the insulation and cover and no bonding
between the insulation and the medium pipe.
According to EN 15632, the European Standard for
pre-insulated flexible pipe systems, this pipe system is
classified part 3: Non bonded system with plastic
service pipes. The Flexalen 600 plastic pipe system
differs in some areas significantly from most other
systems in this class:
1 Physical bonding between foam and outer casing,
2 One layer of foam, filling the complete space
between service pipe and cover,
3 Next to other connection methods the service pipes
can also be connected by welding.
Annex D of part 1: Classification, general requirements
and test methods give rules for calculation of the heat
flow to ambient (heat loss) from measured values,
making the heat flow of various parameters
comparable.
The heat loss calculations of annex D are based on the
thesis of Wallentn as published in Steady-state heat
September 5
th
to September 7
th

113

loss from insulated pipes; Lund Institute of Technology,
Sweden [2].
The described heat loss calculation is valid for the pre-
insulated pipes. Branches and connections are
excluded.

Figure 1, Thermaflex heat loss equipment

In the new developed test rig (figure 1) a test spool
(figure 2) is put in a slim fitting sleeve.
The test spool is heated internally in three sections.
The middle section of the spool is the test section.
Heating in this section is controlled to obtain the
required test parameters. The two ends are heated to
compensate for the heat loss from the ends of the
middle test section. In this way an endless pipe is
imitated.
The outer side of the sleeve is water-cooled to obtain
heat transport from the test spool.
During the start of the test, heat is lost into the heating
of the pipe system and into the surrounding cooling
water. When heat losses have reached equilibrium, the
steady state heat transfer can be measured.

Figure 2, Longitudinal section guarded end heating probe

The time span required for testing in the test rig is
rather short. The time to reach equilibrium lies in the
order of hours, depending on the diameter and
insulation thickness. Comparable tests often require
time spans in the order of days.
Containing various diameters, which are based on the
standard production outer diameters, the test rig
enables heat loss tests for various diameters of piping
systems. Test spools can be extracted directly from
production. In this way tests can be executed with fresh
product.
Also testing of cured piping and piping that is aged and
has been degassed during storage or in high
temperature aging is possible.
Knowing the relationship between heat loss and
various parameters, the most prominent parameters for
heat loss can be evaluated.
The most prominent parameters may lead to the
improvement of the flexible pipe system to ensure
optimal performance with minimal heat loss. The
influence of several prominent parameters is
determined and recommendations are given in order to
optimize the insulation performance.
Reliability and reproducibility of the test rig is discussed
in [3] Verification of heat loss measurements conducted
on (semi) flexible pipe systems (van der Ven et al).
MANUFACTURING PROCESS
The Flexalen 600 product has been developed by
Thermaflex, located in The Netherlands. The
development started in 2002 and resulted in a first
small-scale commercial production in 2005. During the
production of Flexalen 600 four different production
techniques are combined, partly simultaneous and
partly sequential:
1 Production of PB service pipes optionally covered
with an EVOH oxygen barrier layer.
2 Production of LDPE insulation foam to fill the area
between medium pipe and outer casing.
3 Production of outer casing of HDPE.
4 Assembly of the different elements (1, 2 and 3) with
a full bonding of the foam and the outer casing,
while corrugating the casing.
These techniques are based on extrusion technology.
The production line consists of purchased equipment
combined with technology developed in-house. The
complete production is a (semi)-in-line production. All
pipe systems are produced at Waalwijk in the
Netherlands. Unique for the process is the ability to
produce continuous lengths. For practical reasons the
lengths produced depend on the outer casing of the
product and the size of the reel. The maximum length
produced can reach up to 2000 meters.
The complete Flexalen 600 pipe system includes pre-
insulated pipes, couplings, sleeves, pre-insulated T-
connections, etc. The production range is described by
Engel and Baars. [5]
September 5
th
to September 7
th

114

Production of PB service pipes
All service pipes are made of Poly-butene. This is a
plastic with a special combination of properties. Poly-
butene has excellent heat and creep resistance,
flexibility and strength at a long lifespan and is fully
recyclable. In accordance with the temperature
duration profile mentioned in the BRL5609/EN15632,
PB is suitable up to a maximum temperature of 95 C.
All PB-pipes are weldable by socket fusion, electro
fusion and butt welding, which allows for an all plastic
distribution system without metal parts that are prone to
corrosion. The service pipes are produced via state-of-
the art extruders.
The production line consists roughly of an extruder,
calibration tools for adjusting pipe size, cooling baths,
and marking and cutting equipment.
Pipe dimensions are checked inline every second
during and after production with ultrasonic
measurements. PB-pipes can be produced up to an
outer diameter of 225 mm.
If desired an outer oxygen barrier layer may be applied
via co-extrusion up to an outer PB pipe diameter of 90
mm.
After production the PB-pipes are stored for a minimum
period of 5 days and cured to create the correct
crystalline polymer structure. After curing the pipe is
used for the Flexalen 600 production.
Every batch produced is verified by the in-house QC
department according to Dutch directives
- BRL-K5609, for PB pipes with oxygen barrier or
- BRL-K17401 for PB pipes without oxygen barrier.
Product and manufacturing processes are checked 6-8
times a year and certified by independent agencies
such as Bureau Veritas, KIWA and CSTB and le
Centre Scientifique et Technique du Btiment. The
quality of the QC department is validated by these
checks and by internal and external audits.
Production of LDPE insulation foam
Thermaflex has now approximately 35 years of
experience in the production of LDPE foam via
extrusion techniques.
Most of the raw materials that are used are tailor-made
mixtures according to Thermaflex specifications.
Through these specifications and proper production
quality control the companys philosophy related to
core business is also realized for raw materials.
During the heating, melting and mixing of the raw
materials the foaming agent is injected into the
extruder. This foaming agent is a hydrocarbon that
causes the expansion of the LDPE.
The quality of the foam, moreover the insulation
properties of the foam, depends on parameters such as
density, cell size and chemical composition. All
parameters are measured and adjusted within limited
tolerances to meet the specifications. The best
efficiency is further improved when the space between
the service pipe and the foam is filled better.
As there is no bonding between the service pipe and
foam, there is no risk of damaging the foam and
properties by expansion of the pipes due to thermal
fluctuations in the applications.
Production of corrugated outer casing (HDPE) and
assembly of complete product
The outer casing is applied by an extrusion process
and thermally welded to the foam after the pipe has
been inserted. The outer casing is also corrugated to
optimize the flexibility of the finished product. Now the
product is ready for coiling.
After production the final product must cure for 5 days.
During this curing period the degassing of the foaming
agent starts, while the insulation foam is still stabilizing
Production testing and controlling
During the production, parameters are checked and
controlled, such as:
1 Chemical composition of the foam.
2 Settings of all extruders involved (foam extruder and
extruder for corrugated outer casing).
3 Density and cell size of the foam.
4 Dimensions of the foam (outer diameter and inner
diameter).
5 Line-speed of all involved products (Foam / PB-pipe
/ End-product).
6 Thickness of the corrugated outer casing and the
connection of the corrugated outer casing to the
foam.
7 Since a few months: in-line production control of
Heat Loss of the pre-insulated pipe system.
The most prominent factors to influence heat loss are
parameters one to four. The foam lambda is directly
affected by these factors as represented in equation 1.

total

convection

conduction
+
radiation

blowingagent
+ +
...
|
\
|
|
.

(1)
September 5
th
to September 7
th

115

RESULTS AND CORRELATIONS
It is again the companys philosophy that paves the
way for innovations. One of these innovations is the
development of an in-house device for testing pipe
systems, directly from production, cured and degassed.
This proved to be a suitable device for production
control, but also for gaining more insights into product
parameters.
The objective of testing is to improve the knowledge of
the produced pipe systems in order to optimize:
Production methods:
Machine data can be adjusted based on test results.
Test results may lead to new production methods with
new equipment.
Chemical and physical composition of layer material:
Knowledge of various composition materials may lead
to an improvement of insulating values.
Cell structure and gap between insulating foam and
medium pipe:
Cell size influences values, test result based
improvements are possible. The gap is a bad insulator.
The quest for a minimal gap started with testing.
Improvement and minimization of this gap was an
achieved challenge in the testing period.
Thermaflex is a lean and mean organisation that
responds quickly to new insights. Therefore new
insights were applied even before the complete range
of production testing was performed.
For the company, improvements of product and
production have the highest priority. Although the
production range is wide, the insight into specific and
general parameters increased considerably.
The research provides the prominent variables to
improve insulation performance. Practical heat loss
determination, in combination with analytical studies,
results in a clear understanding of heat loss behaviour
in single and twin flexible pipe systems during their
entire lifetime.
As a result of the tests the manufacturing process is
improved in two steps.
The emphasis of the first step was to diminish the cell
size of the foam. This succeeded in a decrease of cell
size by some 20%.
The latest step is altering production such that the
content of anti-radiation agent increases. The initial
results are promising but are not yet conclusive as the
anti-radiation agent is also a good heat conductor.
DISCUSSION OF PARAMETERS
Table 1 summarises the results of measurements and
calculations of tests on Flexalen 600 pipes directly from
production. Various diameters are tested and
calculated according to EN 15632 for a surface
temperature of 10 C and a common medium
temperature of 70 C (Instead of the maximum medium
temperature of 95 C).
The table indicates the relationships between product,
cross sectional area of the foam, foam density, cell size
of the foam, remaining foaming agent, calculated
thermal conductivity and the calculated heat loss of a
buried piping system.
The products 50A25, 63A32, 75A40 and 90A50 are
newly developed. These products are not necessarily
District Heating products. However, they are produced
using the same process and have their application in
the connection between the district heating network
and the building or house. It is also applicable in case
of low temperature differences, cooling or in-house
heating or cooling.

Table 1: Test results of fresh, uncured piping systems
section density cell size agent
mm kg/m (mm) % mW/m.K W/m
50 A 25 1.473 50,0 0,47 52 39 15,3
63 A 32 2.313 34,0 0,50 52 38 15,2
75 A 40 3.044 38,0 0,40 46 44 17,8
90 A 40 5.105 42,0 0,80 64 51 17,1
90 A 50 4.398 39,0 0,80 62 55 23,0
125 A 63 9.155 39,0 0,88 70 56 22,0
160 A 75 15.688 40,3 1,20 81 54 21,0
160 A 90 13.745 35,0 1,30 85 61 25,2
200 A 110 21.913 45,0 1,60 81 68 27,4
50,calc
Heat Loss*
*) calculated heat loss of buried system at temperature difference of 60 K
Product
Foam

In Graph 1 foam density and cell size are related to the
cross sectional surface.
In Table 1 the foam density varies from about 35 kg/m
to about 50 kg/m. Graph 1 shows hardly any
relationship with the surface of the cross section.
Table 1 shows that cell size varies from 0.47 to 1.60
mm. Graph 1 shows that cell size is directly related to
the cross sectional surface, however less than 1 to 1.
This relationship is influenced by physical production
parameters.
September 5
th
to September 7
th

116

0
10
20
30
40
50
60
- 5.000 10.000 15.000 20.000 25.000
Cross section of foam [mm]
A
v
e
r
a
g
e

c
e
l
l

s
i
z
e

[
m
m
]
0,0
0,3
0,6
0,9
1,2
1,5
1,8
D
e
n
s
i
t
y

[
k
g
/
m
]
Graph 1: Density and cell size in relation to foam cross
section

15,0
20,0
25,0
30,0
35,0
40,0
50 100 150 200
Outer diameter pipesystem [mm]
H
e
a
t

l
o
s
s

o
f

b
u
r
i
e
d

p
i
p
i
n
g

[
W
/
m
]
30
40
50
60
70
80
C
o
n
d
u
c
t
i
v
i
t
y

[
m
W
/
m
.
K
]
Graph 2: 50,calculated and calculated heat loss of buried
pipe in relation to outer pipe size

Graph 2 shows the influence of outer pipe size to
calculated conductivity
50
and heat loss of buried pipe
systems.
It also shows that part of the increase of the heat loss
with the diameter is caused by increase of conductivity.
40
50
60
70
80
90
50 100 150 200
Outer diameter pipesystem [mm]
P
e
r
c
e
n
t
a
g
e

f
o
a
m
i
n
g

a
g
e
n
t
Graph 3: Foaming agent content in relation to outer
diameter size
Graph 3 shows the relationship between outer pipe
system diameter size and the percentage of foaming
agent directly from production. With increasing
diameter the foaming agent increases, possibly to an
asymptotic value.
An additional interesting factor is the ongoing process
in the foam during and after production. As described
before the final step in the production is a 5-day curing
stage.
During the curing stage the foam expands and part of
the foaming agent releases from the foam. As the foam
is locked by a hard outer shell, expansion is directed
inwards. By this the gap between the foam and the PB
medium pipe, typical for our production method, is
decreased.
Table 2 shows the effect of curing and degassing on
both the contents of foaming agent and the calculated
heat loss.
Even when forced, degassing takes time. The number
of degassed samples manufactured in the same way
as the fresh samples is therefore limited. Table 2 is
short due to a lack of adequate and comparable
samples.
Table 2: The effect of time on curing and degassing
Agent
Heat
Loss
Agent
Heat
Loss
A Agent
Heat
Loss
A
% W/m % W/m % % W/m %
63 A 32 52 15 23 17,5 15 0 17,1 12
75 A 40 46 18 40 16,9 -5 0 20,1 13
90 A 40 64 17 53 18 5 0 18,5 8
6 days curing Degassed
Product
Fresh

15
16
17
18
19
20
0 10 20 30 40 50 60
Percentage of foaming agent
H
e
a
t

l
o
s
s

o
f

b
u
r
i
e
d

p
i
p
i
n
g

[
W
/
m
]
Graph 4: Relationship between heat loss and foaming
agent

Graph 4 shows that there is a tendency of decreasing
heat loss with increasing foaming agent. This tendency
has seems weak. The spread is large over the entire
graph.
September 5
th
to September 7
th

117

RESULT BASED FUTURE DEVELOPMENTS
The tests have led to improvements to the flexible
plastic pipe products produced by Thermaflex. In the
near future we expect improvements in:
Production methods:
Starting June 2010 a change in the machine
configuration will be implemented as an extra step. The
new configuration improves temperature control in the
extruder, which leads to better cell structure.
Chemical and physical composition of layer material:
Up to a certain degree the anti-radiation agent
improves isolating values. With trial and error the anti-
radiation agent content is increased. Up till now the
maximum content has been limited by production
methods. Research is required to investigate maximum
desired value for insulating effects.
Cell structure and gap between foam and medium pipe:
In the tests we see variations in cell structure and gap
width. Future research will aim at acquiring more
detailed knowledge of these phenomena.
Up till now the heat loss performance on single pipes
has been measured and analysed. This has resulted in
an understanding of the heat loss principles in district
heating systems. Twin pipe systems will soon be
tested, analysed and evaluated.
In this paper only the heat loss of the Flexalen 600 pre-
insulated pipe product has been handled. Information
about the system can be read in [4] Heat loss
optimization of flexible plastic piping systems, life time
heat loss performance (Korsman et al) and [5] New
economical connection solutions (Engel).
CONCLUSION
The results of testing are reliable. Knowledge of the
product and production has led to promising
improvements of both.
Further research will certainly lead to further
developments.
ADDENDUM
Degassing
The Thermaflex pipe system is liable to the process of
degassing. Degassing causes the heat loss values to
rise over the products life time. Extra research on this
subject shows an average heat loss increase of
9 percent (range 513 percent) (graph 5). Heat losses
are calculated according to EN 15632.
15.0
17.0
19.0
21.0
23.0
25.0
27.0
29.0
31.0
50 70 90 110 130 150 170 190
Outer diameter pipe system [mm]
H
e
a
t

l
o
s
s

o
f

b
u
r
i
e
d

p
i
p
i
n
g

[
W
/
m
] Heat loss
Heat loss Degassed

Graph 5: Degassed heat loss values of a buried system at
a temperature difference of 60K
Examples of product improvement
Further research has led to product improvements.
Based on these improvement proposals Thermaflex
has been able to produce new pipe system samples.
As represented in graph 5 the new samples have a
heat loss decrease up to 16 percent compared to the
previous results.
15.0
20.0
25.0
30.0
35.0
40.0
50 70 90 110 130 150 170 190
Outer diameter pipe system [mm]
H
e
a
t

l
o
s
s

o
f

b
u
r
i
e
d

p
i
p
i
n
g

[
W
/
m
]
Heat loss Heat loss New

Graph 6: New heat loss values of a buried system at a
temperature difference of 60K
September 5
th
to September 7
th

118

ACKNOWLEDGEMENT
We would like to acknowledge P. van Rijswijk for his
dedication to all the heat loss measurements
performed during this research.
FURTHER INFORMATION
Questions concerning the paper may be addressed to:
Thermaflex International Holding B.V,
Veerweg 1
5145NS Waalwijk
The Netherlands
www.thermaflex.com
Liandon B.V.
Dijkgraaf 4
6920AB Duiven
The Netherlands
www.liandon.com
REFERENCES
[1] EN 15632 District heating pipes, Pre-insulated
flexible pipe systems, Requirements and test
methods
[2] P. Wallentn; Lund Institute of Technology,
Sweden; 1991, Steady-state heat loss from
insulated pipes
[3] E. van der Ven, F. Duursma, H. Korsman, I. Smits;
Paper on DHC, Tallinn; 2010, Verification of heat
loss measurements conducted on (semi) flexible
pipe systems
[4] H. Korsman; Paper on DHC, Tallinn; 2010, Heat
loss optimization of flexible plastic piping systems,
life time heat loss performance
[5] C. Engel and G. Baars, New economical
connection solution for flexible piping systems,
12
th
ISDHC 2010.
September 5
th
to September 7
th

119

COMPARISON OF COMPETITIVE (SEMI) FLEXIBLE PIPING SYSTEMS BY MEANS
OF HEAT LOSS MEASUREMENT
I.M. Smits
1
, J. Korsman
1
, J.T. van Wijnkoop
1
and E.J.H.M. van der Ven
2

1
Liandon B.V.
2
Thermaflex International Holding B.V.

ABSTRACT
Different types of pre-insulated pipes are tested on
their heat loss values. Three flexible pipes and a rigid
pipe are tested. The different heat loss values are
compared not only on absolute heat loss, but also on
their performance relative to the insulation surface. The
heat loss values are measured according to EN 15632
and published as declared values. The declared
values are calculated according to EN 15632 Annex
D1-D3.
The flexible pre-insulated systems, with PE and PE-X
foams, show a variance of up to 5 W/m in the heat loss
values. These absolute differences in the system are
caused by the outer casing dimensions of the pre-
insulated pipes. Recalculation to the same outer casing
diameters shows a slight advantage for the PE system
in service pipes of 32 and 63 millimetres
The flexible piping system with the PUR insulation
foam on the other hand performs better compared to
equally dimensioned flexible PE and PE-X insulation
foams.
Flexible pre-insulated pipes have a higher heat loss
compared to rigid pre-insulated pipes. Recalculation to
the same transport capacity [kg/s] and the same outer
casing diameter also shows that rigid pre-insulated
pipes perform better. However the fact that smaller
diameters show a smaller heat loss difference between
rigid and flexible pre-insulated pipes is interesting.
INTRODUCTION & OBJECTIVE
The objective of this research is to compare different
types of competitive (flexible and rigid) pre-insulated
pipes on their differences in heat loss values. The
comparison is based on an overall heat loss
measurement under similar conditions. Overall heat
loss is determined for different samples of pre-insulated
pipes, by using newly developed heat loss testing
equipment as described in Verification of heat loss
measurements by J.T. van Wijnkoop et al. [1]. The
heat loss data of these flexible pipes will be compared
with practical measurement on a rigid pre-insulated
pipe.
Notice that this study does not compare entire district
heating systems. For system comparisons see Heat
loss analysis and optimization of a flexible piping
system by J. Korsman et al. [2].
Where most studies only focus on one product this
study compares different types of flexible pre-insulated
pipes on their practical heat loss values and gives an
explanation of the practical heat loss values. It also
compares flexible pipes with rigid pre-insulated pipes
on an equal basis.
BRIEF METHOD DESCRIPTION
First a brief description of different types of flexible pre-
insulated pipes and a rigid pre-insulated pipe is given.
This chapter highlights the differences and similarities.
The different types of foam for plastic pre-insulated
pipes are described in a separate paragraph.
Secondly, the method of testing is briefly addressed.
Thirdly, the different types of flexible pre-insulated
pipes are tested on their absolute heat loss just after
production.
Since heat loss of pre-insulated pipes can increase
over time due to degassing of the insulation foam, a
gas analysis is performed on all test samples.
In the second paragraph the absolute heat loss values
of the different types of flexible pre-insulated pipes are
compared on the basis of service pipe dimensions. The
third paragraph defines a comparison on the basis of
insulation surface and service pipe dimension.
In the fourth paragraph the comparison of flexible
pre-insulated pipes versus a rigid pre-insulated pipe is
described. The comparison in the third and fourth
paragraph is based on declared values. The defined
conditions are: (1) thermal conductivity of soil: 1.0
W/m.K, (2) thermal transmittance factor of earth to
ambient air: 0.0685 m
2
.k/W and (3) soil covering:
0.8 m.
In the first paragraph absolute heat loss values are
compared on the basis of corresponding service pipe
dimensions. The second paragraph gives a comparison
based on equal transport capacity for flexible and rigid
September 5
th
to September 7
th

120

pre-insulated pipes. The third paragraph founds the
second comparison by adding a heat loss value based
on insulation surface.
Finally, both flexible and rigid pre-insulated pipes are
compared, resulting in conclusions concerning flexibility
versus heat loss behaviour.
PRE-INSULATED PIPES
This paper compares different types of pre-insulated
pipes and highlights their mutual similarities and
differences.
The flexible pre-insulated pipe systems are;
- Two different types of Cross linked Polyethylene
(PEX) service pipe with Cross linked
Polyethylene (PEX) insulation;
- One type of Cross linked Polyethylene (PEX)
service pipe with Polyurethane (PUR) insulation;
- One type of Polybutene (PB) service pipe with
Polyethylene (PE) insulation.
The rigid pre-insulated pipe system is;
- One type of Steel (ST) service pipe with
Polyurethane (PUR) insulation.
Firstly all types of pre-insulated pipes are functionally
explained. Secondly the different kinds of foam
production methods are described. All types of pre-
insulated pipes described are commonly available
products used for district heating purposes in Europe.
PEX service pipe with PEX insulation
A short description of the PEX/PEX systems is given.
Figure 1 shows the cross section view of the
PEX/PEX/PE pre-insulated pipe. A cross linked Poly-
ethylene (PE-Xa) service pipe with anti-oxygen barrier
(EVOH) is placed in a multiple layered low-density
cross linked poly-ethylene (PE-X) insulation foam with
a corrugated outer casing of high density Poly-ethylene
(HDPE). Because of the cross linking in both service
pipe and foam this product cannot be re-used.

Fig.1. Section view of the PEX/PEX pipe
PEX service pipe with PUR insulation
This paragraph describes the flexible PEX/PUR pre-
insulated pipe. Figure 2 shows the cross section view
of the PEX/PUR/PE pre-insulated pipe. A cross linked
poly-ethylene (PE-Xa) service pipe with anti-oxygen
barrier (EVOH) is placed in Polyurethane (PUR)
insulation foam with a corrugated outer casing of high
density poly-ethylene (HDPE). This product shows a
tight connection between the service pipe, the foam
and the outer casing. This product therefore cannot be
re-used once it is formed.

Fig. 2. Section view of PEX/PUR pipe
PB service pipe with PE insulation
The third type of pre-insulated pipe is a flexible PB/PE
pipe. Figure 3 shows the cross section view of the
PB/PE/PE pre-insulated pipe. A Polybutene (PB)
service pipe with anti-oxygen barrier (EVOH) is placed
in a low-density poly-ethylene (LDPE) insulation foam
with a corrugated outer casing of high density Poly-
ethylene (HDPE).
The PB service pipe makes it possible to use electro
fusion welding with a PB coupling. This makes a strong
bond. Corrosion is not an issue, because PB is inert
with water.
There is no connection between the PB and foam plus
outer casing. Therefore it is possible to re-use both
elements in own production. The complete product can
be re-used.
Information regarding the use of this product is given in
New economical connection solutions for flexible
piping systems (Engel) [7].

Fig. 3. Section view of the PB/PE pipe
Steel service pipe with PUR insulation
The last system described in this paper is the rigid
steel/PUR system. Figure 4 shows the cross section
view of the ST/PUR/PE pre-insulated pipe. A steel (St)
service pipe is placed in Polyurethane (PUR) insulation
foam with a smooth outer casing of high density Poly-
ethylene (HDPE). Because of the steel service pipe,
the complete system has to be mechanically welded.
Also, because of the combination of steel and water,
there is a potential risk of corrosion.
September 5
th
to September 7
th

121

This system is called a rigid system with a tight
connection between the service pipe, the foam and
outer casing. Once it is formed, this product cannot be
re-used in own process.

Fig. 4. Section view of Steel/PUR pipe
Foam production processes
The insulation foams described in this paper are made
of Poly-urethane (PUR) foam, Poly-ethylene (PE) foam
or cross linked Poly-ethylene foam (PE-Xa). These
foams have different properties. Some of these
properties influence the heat loss properties of the
complete product.
Polyurethane (PUR) foam
PUR foam is a thermo-set foam. It is made out of two
chemicals, a Poly-alcohol and an Iso-cyanate. These
materials react and the Polyurethane is formed. This
reaction is irreversible, so the material can never return
into its original chemicals. The blowing agent for this
kind of foam can be Carbon Dioxide, Nitrogen or
Hydrocarbon molecules, for instance Cyclopentane or
Butane.
If Hydrocarbon gases are used, these gases strongly
influence the heat loss performance of the pre-
insulated system. These gases have different thermal
conductivities compared with air. After production of the
foam an exchange with air starts. A product that is
freshly made contains a high percentage of
Hydrocarbon gases. At this point in time the product
will have the lowest heat loss possible. If the same
product is for instance three years old it contains more
air and less Hydrocarbon gases due to gas diffusion.
And so the product will have a higher heat loss
compared to the fresh product. The process of
degassing is described in research papers Long term
heat loss of plastic Polybutylene piping systems by S.
de Boer et al. [3].
Poly-ethylene (PE) foam
PE foam is a thermoplastic foam. Once it is formed, it
can go back to its original state by heating it above its
melting point. Because of this property, it is possible to
re-use these kinds of foams.
The foaming process to make PE foam is called the
physical foaming process. A Hydrocarbon molecule is
mixed into the PE matrix under high pressure in an
extruder. Once the material flows out of the extruder,
the pressure drop causes the expansion of the
hydrocarbon. The aggregation state of the molecule
changes from liquid into gas.
Examples of hydrocarbon gases that can be used are:
LPG, Butane or Isobutane. And just like the PUR foam
there is a degassing effect: the exchange of the
blowing agent with air will increase the heat loss of the
product. This effect is shown in Heat loss of flexible
plastic pipe systems, analysis and optimization by
E.J.H.M. van der Ven et al. [4].
Cross linked Poly-ethylene foam (PE-Xa)
Although cross linked PE foam (x-PE) is also made of
PE, there is a big difference compared to PE foam: the
type of Blowing agent.
The foaming process to make x-PE foam is called the
chemical foaming process. In this case a chemical is
mixed into the PE matrix. The blowing agent can for
instance be Azodicarbonamide. While heating the
matrix, the chemical starts decomposing and gases are
released. These gases are Carbon dioxide and
Nitrogen. The thermal conductivity of these gases is
more or less equal to the thermal conductivity of air. So
the aging effect of this product in relation to the heat
loss is less.
To make this foaming process possible, it is necessary
to connect the Poly-ethylene chains with each other.
This is called the cross link. The complete process to
make x-PE foam is called the chemical foaming
process with cross link. To make the comparison with
PE foam complete: this process is called the physical
foaming process without cross link.
The blowing agent is not the only additive that
influences the thermal conductivity of the foam and
therefore the heat loss properties of the pre-insulated
system. Also other additives can influence the thermal
conductivity of PE foam.
Nucleating agents will influence the cell structure of the
foam. As a basic rule: the finer the foam the lower the
thermal conductivity. With this additive the convection
part of the insulation material will be influenced.
Another additive that influences the thermal
conductivity is an anti-radiation additive. By using this
special kind of additive it is possible to create a
reflection of radiation energy.
HEAT LOSS TEST METHOD
This chapter briefly describes the test rig and test
method used to determine the absolute heat losses of
the different types of pre-insulated pipes.
September 5
th
to September 7
th

122

Test rig
The test rig is used to determine the absolute heat
losses. The test rig has been designed in compliance
with EN 15632 and the tests are carried out according
to ISO 8497 and EN 15632.
The physical part of the Thermaflex heat loss
equipment consists of three sections. The first is the
water cooled compartment in which all the tests are
performed. This compartment is kept at a constant
23 C during each measurement.
Heating probes are used as a heat source. These
heating probes are custom made by preparing a two
meter Thermaflex piping segment of all available
diameters. The third part of the heat loss equipment is
the control unit. This unit powers the probes and
regulates the temperature and reads out the
temperature and power values.
Method of testing
Different heating probes are used for the testing. The
probe with the appropriate diameter is inserted in a test
sample and inserted in the cooled test section.
The heat loss measurement is done by measuring the
energy required to keep the probe at a constant
temperature, by measuring the current at constant
voltage in the heating coils and calculating the power
consumption. Since the middle/testing coil is exactly
one meter in length the required energy represents the
exact heat loss through one meter of piping and
insulation in W/m. For this paper the heat loss is
determined for multiple probe temperatures.
Information
For more information concerning the test rig and
method of testing see the paper Verification of heat
loss measurements (J.T. van Wijnkoop et Al. [1])
Blowing agent analysis
All products that are involved in this paper have been
analyzed on quantity of blowing agent and type of
blowing agent. The following results were found:
- The samples of PEX/PEX I and PEX/PEX II
did not show any amounts of hydrocarbon
blowing agents;
- The PB/PE samples contained a quantity of
hydrocarbon blowing agent over 50 percent;
- The samples of ST/PUR and PEX/PUR
products contained a mixture of hydrocarbon
gases. These gases were analyzed. Both
product types contained approximately 95% of
blowing agent.
Measured samples
The table below contains all products and dimensions
of outer casing and service pipe that are treated in this
paper.
Table 1: Measured products
ST/PUR
Dc/Ds
PEX/PUR
Dc/Ds
PB/PE
Dc/Ds
PEX/PEX
II
Dc/Ds
PEX/PEX
I
Dc/Ds
-- -- 62A32 -- --
90DN25 -- 90A32 90A32 90A32
125DN50 -- 125A63 160A63 175A63
162DN80 162A110 200A110 200A110 200A110
Results of testing
All results have been extracted from measurements
carried out by the Thermaflex testing rig. The new
European standard EN 15632 has been used.
This standard describes in Annex D a method to
present the results of testing in end-use condition. This
means: the product is buried in soil. According Annex
D.3 the following general values are used for the
calculation:
- Soil covering
o 0.8 m
- Thermal transmittance factor of earth-air
o 0.685 m
2
.K/W
- Thermal conductivity of the soil
o 1.0 W/(m.K)
The heat loss is calculated using the following
formulas:
(1) Q
T
flow
T
surrounding
R
soil
R
flow
+
|
\
|
|
.
:=

W
m

(2)
Z
d
4
2
|
\
|
|
.
H + :=

m
(3)
Z
c
Z R
0
soil
+ :=

m
(4) R
soil
1
2 t
soil
ln
4 Z
c
d
4
|
\
|
|
.
:=

mK
W

All results are presented in W/m, measured and
calculated at a temperature difference of 60 Kelvin.
This temperature difference is derived from inner
service pipe temperature minus surrounding ambient
September 5
th
to September 7
th

123

temperature (70 degrees Celsius minus 10 degrees
Celsius).
Table 2: Results according to EN 15632 at a temperature
difference of 60 Kelvin.
Product

Type Heat Loss
Buried
system
W/m
System
(W/m.K)
ST/PUR 90DN25 11.6 0.042
ST/PUR 160DN80 16.0 0.033
PEX/PUR 162A110 22.3 0.049
PB/PE 63A32 15.2 0.038
PB/PE 90A32 12.8 0.044
PB/PE 125A63 22.0 0.056
PB/PE 200A110 27.4 0.068
PEX/PE II 90A32 16.6 0.057
PEX/PE II 160A63 17.6 0.055
PEX/PE II 200A110 31.1 0.073
PEX/PE I 140A32 12.5 0.057
PEX/PE I 175A63 17.6 0.059
PEX/PE I 200A110 28.8 0.051
COMPARISON OF FLEXIBLE PLASTIC
PRE-INSULATED PIPES
This chapter compares the flexible pre-insulated pipes.
The comparison is based on three diameters
representing the entire diameter range for plastic pre-
insulated pipes.
The comparison is expanded by evaluating the heat
loss in correlation to the outer casing diameter (resp.
the foam area).
In Table 1 the flexible plastic pre-insulated pipes are
defined. These are the products PEX/PEX, PEX/PUR
and PB/PE.
For more information concerning the PB/PE pre-
insulated pipes see Heat loss of flexible plastic pipe
systems, analysis and optimization by van der Ven et
al. [4].
The 32, 63 and 110 millimetre service pipes
First the absolute heat loss is displayed, followed by
the insulation area analysis.
Absolute heat loss
In this paragraph all absolute heat loss values are
compared for the 32, 63 and 110 millimetre service
pipes.
In Graph 1 the results are displayed for temperature
difference of 60 Kelvin.
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
90/32 160/63 200/110
Diameter service pipe [mm]
H
e
a
t

L
o
s
s

[
W
/
m
]
PB/PE PEX/PEX I PEX/PEX II PEX/PUR

Graph 1 Absolute Heat Loss 32, 63 and 110 mm service
pipe (dT = 60 K)

The products based on PE or PE-x foam show higher
heat losses for the 110 mm service pipe than the
system based on PUR foam. The difference is
approximately 20 percent.
The different test samples show a wide variance in the
diameter of the outer casing.
Therefore, only the results for the 32 millimetre service
pipe are comparable for PB/PE 90A32 and PEX/PEX II
90A32. For the 110 mm service pipe, a comparison can
be made between the PEX/PUR 200A110, PB/PE
200A110, PEX/PEX II 200A110 and PEX/PEX I
200A110.
Another difference in this comparison is the use of a
PB pipe or a PE-x pipe. PB and PE-x have different
thermal conductivities (0.19 W/m.K versus 0.40
W/m.K). However, this effect is already corrected by
using the Wallentn equation [5], as shown in (1).
Insulation area
To compare the different kinds of flexible pre-insulated
pipes on their performance, all outer diameters are
altered towards 90, 160 and 200 millimetres. The
September 5
th
to September 7
th

124

corresponding heat loss is calculated using the thesis
of Wallentn [5], as in (5).
o
i
( )
2t T
probe
T
casing
( )
s
ln
d
2
d
1
|
\
|
|
.
i
ln
d
3
d
2
|
\
|
|
.
+
1
c
ln
d
4
d
3
|
\
|
|
.
+
:=
Where:
- T
probe
/T
casing
= Probe / Casing temperature
- d
1
to d
4
= inner/outer diameters of service
pipe and casing
-
s
, i,
c
= heat coefficient of service pipe,
insulation and casing
Graph 2 represents the comparison on the basis of
the same outer casing.
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
90/32 160/63 200/110
Diameter casing/service pipe [mm]
H
e
a
t

L
o
s
s

[
W
/
m
]
PB/PE PEX/PEX I PEX/PEX II PEX/PUR

Graph 2 Relative Heat Loss 32, 63 and 110 mm service
pipe, all with an equal outer casing (dT=60K).
Result analysis
The flexible pre-insulated systems, with PE and PE-x
foams, show a variance in heat loss values. The
absolute differences in the system are caused by the
dimensions of the pre-insulated pipes and the quantity
and type of blowing agent that has been used. Also the
recalculation to the same outer casing diameters
shows an advantage for the PE foamed system in PB
service pipes of 32, 63 and 110 millimetres.
COMPARISON OF FLEXIBLE PLASTIC PRE-
INSULATED PIPES VERSUS A RIGID PIPING
SYSTEM
In this chapter the flexible pre-insulated pipes are
compared with a rigid piping system. The
comparison is based on diameter. The comparison
is expanded by evaluating the heat loss in
correlation to the foam area.
Rigid Pre-insulated pipes
The different systems and their corresponding
dimensions are represented in Table 1.
The rigid pipe product that has been tested
according to EN 15632 was the ST/PUR product.
First the absolute heat loss is displayed, followed by a
recalculation towards transport capacity and finally the
insulation area analysis.
Absolute heat loss
In this paragraph all absolute heat loss values are
compared. The rigid DN25 pipe service pipe is
compared with a flexible PB/PE-x service pipe with an
outer diameter (OD) of 32 mm. DN50 is compared with
OD 63 mm and DN80 is compared with OD 110 mm.
In Graph 3 the results are displayed for temperature
differences of 60 Kelvin.
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
DN25-PB32 DN50-PB63 DN80-PB110
H
e
a
t

L
o
s
s

[
W
/
m
]
PB/PE PEX/PEX I PEX/PEX II PEX/PUR ST/PUR

Graph 3 Absolute Heat Loss DN25/PB32, DN50/PB63 and
DN80/PB110 mm service pipe (dT = 60 K)).

The different test samples show a wide variance in the
diameter of the outer casing.
The heat loss for ST/PUR 160DN80 is much lower
compared to the heat loss of the 200A100 flexible
piping products.
Even the difference with the PUR based PEX/PUR
system is high (28 percent). For the PE and PE-x foam
based products the difference is even higher
(42 percent)
The heat loss for ST/PUR 90DN25 is more or less
comparable with the heat loss for PB/PE type 90A32
(9 percent). So it seems that for smaller sizes the
(5)
September 5
th
to September 7
th

125

difference in absolute heat loss is lower, compared to
the absolute heat loss difference for larger sizes.
Transport capacity
When comparing rigid steel service pipes with flexible
plastic service pipes there is a difference in transport
capacity for comparable diameters.
This paragraph calculates the amount of heat loss
when transporting water of 70 degrees Celsius through
one meter of steel DN25, DN50 and DN80
(k-factor = 0.07 mm, velocity = 1.0 m/s). Subsequently
the same amount of heat loss is used as a reference
for calculating the amount of water that can be
transported through a plastic pipe 32 and 63
(k-factor = 0.007 mm, velocity = 1 m/s). For these
calculations the thesis of Colebrook and White [6] is
used. The results of this calculation are displayed in
Table 3.
Table 3: Calculation results transport capacity
DN2
5
PB32 DN5
0
PB63 DN8
0
PB110
Velocity
[m/s]
1.0 1.072 1.0 0.81 1.0 0.76
Head
Loss
[Pa/m]
425 188 112
Flow
[kg/s]
0.64 0.57 2.33 1.68 5.35 4.86

Next the absolute heat loss is recalculated to an equal
flow per diameter. The basis is 0.57 [kg/s] for the
DN25/PB32, 1.68 [kg/s] for the DN50/PB63 and
4.86 [kg/s] for the DN80/PB110.
Insulation area
To compare the flexible pre-insulated pipes and the
rigid piping system on their performance, all outer
diameters are altered towards 90 and 160 millimetres
and compared on the same transport capacity. The
corresponding heat loss is calculated using the thesis
of Wallentn, as in (5). The steel DN80 and PB 110
has the same outer casing and is not recalculated.
Overall comparison
The comparison on basis of the same outer casing
and transport capacity is shown in Graph 4.
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
90/DN25-PB32 160/DN50-PB63 200/DN80-
PB110
H
e
a
t

L
o
s
s

[
W
/
m
]

Graph 4 Relative Heat Loss, all with an equal outer
casing and transport capacity.
Result analysis
Flexible pre-insulated pipes have a higher absolute
heat loss compared to rigid pre-insulated pipes.
Recalculation to the same transport capacity [kg/s] and
the same outer casing diameter shows that rigid pre-
insulated pipes perform better.
The reason for this difference is the relative small inner
diameter of the plastic service pipes. The low k-factor
can not compensate for the smaller diameter. Table 4
shows the steel versus plastic service pipe diameter
dimensions.
Table 4: Service pipe diameter dimensions
DN25 PB32 DN50 PB63 DN80 PB110
ID
[mm]
28.5 26.0 54.5 51.4 82.5 90.0
However the fact that smaller diameters show a smaller
heat loss difference between rigid and flexible pre-
insulated pipes is interesting.
CONCLUSIONS
This chapter briefly addresses each chapter and
outlines its conclusions.
Test samples
This paper compares different types of pre-insulated
pipes that have been randomly taken from the market.
The flexible pre-insulated pipes compared in this paper;
- PB/PE,
- PEX/PEX I,
- PEX/PEX II,
- PEX/PUR.
September 5
th
to September 7
th

126

The rigid system in this paper;
- ST/PUR.
Method of testing
All heat loss tests are performed on a test rig that has
been designed in compliance with EN 15632. The
tests are carried out according to ISO 8497 and
EN 15632.
Blowing agent analysis
All measured products are checked on type of gas and
gas content. The ST/PUR and PEX/PUR products
contain approximately 95 percent of blowing agent.
The PB/PE product range has a quantity over 50% of
blowing agent.
In the products of PEX/PEX II and PEX/PEX I no
Hydrocarbon gases were detected.
Comparison of flexible pre-insulated pipes
A fair comparison is difficult because of differences in
outer casing and other dimensions. These conclusions
are therefore only valid for the products that have been
tested for this paper.
In a buried condition the PB/PE pre-insulated pipe
shows for equally dimensioned pipes 90A32 and
200A110 the lowest absolute heat loss values for all
pre-insulated pipes based on PE or PE-x foam.
As mentioned before, the absolute differences in the
system are caused by the dimensions of the pre-
insulated pipes. Recalculation of the same outer casing
diameter shows also an advantage for the
PB/PE system in service pipes of 32, 63 and 110 mm.
See Graph 5.
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
90/DN25-PB32 160/DN50-PB63 200/DN80-
PB110
H
e
a
t

L
o
s
s

[
W
/
m
]

pipe, all with an equal outer casing and transport capacity
(dT = 60 K)
The flexible piping system with the PUR insulation
foam on the other hand performs better compared to
flexible PE and PE-X insulation foams with equal
dimensions.
Comparison of flexible piping system versus the
rigid pre-insulated pipes
Flexible pre-insulated pipes have a higher absolute
heat loss compared to rigid pre-insulated pipes.
Recalculation to the same transport capacity [kg/s] and
the same outer casing diameter shows that rigid pre-
insulated pipes perform better.
However the fact that smaller diameters show a smaller
heat loss difference between rigid and flexible pre-
insulated pipes is interesting.
To be comparable in heat loss some dimensions of the
flexible piping systems range need to be optimized.
However, other advantages of flexible pipe systems, for
instance the potential decrease of service meters
because of a curved layout-design, can partly
compensate the higher heat loss compared to the rigid
system (see Heat loss analysis and optimization of a
flexible piping system by J. Korsman et al. [2]).
ADDENDUM
Significant product improvement of the PB/PE/PE pipe
system has led to a decrease in heat loss [4]. Graph 5
is updated with these improvements resulting in the
comparison displayed in Graph 6. The new samples
are displayed under the name of PB/PE II.
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
90/DN25-PB32 160/DN50-PB63 200/DN80-
PB110
H
e
a
t

L
o
s
s

[
W
/
m
]
PB/PE PB/PE II PEX/PEX I PEX/PEX II
PEX/PUR ST/PUR

pipe, all with an equal outer casing and transport capacity
(dT = 60 K)
September 5
th
to September 7
th

127

FURTHER INFORMATION
Questions concerning the paper can be addressed to:
- Thermaflex International Holding B.V.
Veerweg 1
5145NS Waalwijk
The Netherlands
www.thermaflex.com
- Liandon B.V.
Dijkgraaf 4
6920AB Duiven
The Netherlands
www.liandon.com
ACKNOWLEDGEMENT
We would like to thank all involved employees of
Thermaflex Isolatie B.V. who made this research
possible (especially H. Leunessen and M. van Doorn).
Special thanks go to P. Blom and P. van Rijswijk for the
dedication they showed in carrying out all the heat loss
measurements during this research.
REFERENCES
[1] J. T. van Wijnkoop and E.J.H.M. van der Ven,
Verification of heat loss measurement, in Proc. of
the 12th International Symposium on District
Heating and Cooling, Tallinn, Estonia (2010).
[2] J. Korsman, I.M. Smits and E.J.H.M. van der Ven,
Heat loss analysis and optimization of a flexible
piping system, in Proc. of the 12th International
Symposium on District Heating and Cooling,
Tallinn, Estonia (2010).
[3] S. de Boer, J. Korsman and I.M. Smits, Long term
heat loss of plastic Polybutylene piping systems,
in Proc. of the 11th International Symposium on
District Heating and Cooling, Tallinn, Reykjavik
(2008).
[4] E. J .H. M. van der Ven and R.J. van Arendonk,
Heat loss of flexible plastic pipe systems, analysis
and optimization, in Proc. of the 12th International
[5] P. Wallentn, steady-state heat loss from
insulated pipes, Lund Institute of Technology,
Sweden, 1991
[6] C. F. Colebrook, "Turbulent flow in pipes, with
particular reference to the transition region
between smooth and rough pipe laws", February
1939
connection solution for flexible piping systems, in
District Heating and Cooling, Tallinn, Estonia
(2010).
September 5
th
to September 7
th

128

EFFECTIVE WIDTH THE RELATIVE DEMAND
FOR DISTRICT HEATING PIPE LENGTHS IN CITY AREAS
Urban Persson
1
, Sven Werner
1
1
School of Business and Engineering
Halmstad University, PO Box 823, SE-30118 Halmstad, Sweden

ABSTRACT
One key concept when assessing network investment
cost levels for district heating systems is the linear heat
density. In contrast to a traditional way of expressing
this quantity entirely on the basis of empirical data, a
recently developed analytical approach has made it
possible to estimate linear heat densities on the basis
of demographic data categories. A vital complementing
quantity in this analytical approach is the concept of
effective width.
Effective width describes the relationship between a
given land area and the length of the district heating
pipe network within this area. When modelling
distribution capital cost levels by use of land area
values for plot ratio calculations, there is a potential
bias of overestimating distribution capital cost levels in
low dense park city areas (e < 0.3).
Since these areas often include land area sections
without any housing, avoiding overestimations of
network investment costs demand some kind of
corrective mechanism. By use of calculated effective
width values, a compensating effect at low plot ratio
levels is achieved, and, hence, renders lower
anticipated distribution capital cost levels in low dense
park city areas.
INTRODUCTION
One key concept when estimating investment cost
levels for district heating systems is the linear heat
density, i.e. the quota of annually sold heat in a district
heating scheme and the trench length of the piping
system in this scheme (Q
s
/L) [1]. In contrast to a
traditional way of expressing this quantity entirely on
the basis of empirical data, a recently developed
analytical approach has made it possible to estimate
linear heat density on the basis of demographic data
categories [2]. A vital complementing quantity in this
analytical approach is the concept of effective width.
BACKGROUND
Effective width is a stand alone concept within district
heating theory, describing the relationship between a
given land area, A
L
, and the length of the district
heating pipe network, L, within this area. Hence, the
effective width becomes the width of an analogous
rectangle with the trench length as the length and
where the rectangle area is equal to the given land
area.
The concept was introduced by Werner [3] and has
been further elaborated recently in model estimations
of distribution capital cost reactions to decreased heat
demands in four north European countries [2].
Essential for calculations of anticipated investment cost
levels for future district heating systems, the effective
width constitutes an important model parameter
indicating levels of network extensions in given land
areas.
Since the concept of effective width itself is rather new,
with no previous analytical or statistical use, data on
effective widths are in principal non attainable within
national statistical sources. Effective width might be
regarded as an innovative model quantity with no
previous representation in the field of district heating
research.
AIM
The aim of this paper is to describe the concept of
effective width and outline the basic properties of this
quantity. On the basis of, although sparse, empirical
observations, preliminary statements concerning the
properties of effective width are made. The aim is
further to enlighten the theoretical environment in which
effective width contributes when applying demographic
quantities for estimations of district heating network
investment costs.
LIMITATIONS
Due to a limited amount of empirical data, in principal
less than 100 observations, the specific result values
and relationships accounted for in this paper must be
considered as preliminary. Although thorough in theory,
the concept of effective width needs to be supported
further by extended empirical data gathering. In order
to be able to produce solid and reliable estimations of
effective width values in different kinds of city areas,
such information is considered vital for future use of the
concept.
September 5
th
to September 7
th

129

EFFECTIVE WIDTH
Effective width is a measure indicating the district
heating network extension level within a given land
area. The quantity effective width, which is symbolised
by use of the letter w, with the unit metres, expresses
the ratio between land area and the total trench length
of the distribution network within a district heating
system [3]
w = A
L
/ L [m] (1)
Being in this way the result of explicit area and grid
properties, effective width can be used to describe
typical district heating properties in different population
density areas and hence, give information on
prerequisite conditions for future district heat
establishments.
THE CONCEPT
In order to introduce the concept of effective width, it is
necessary to first understand some basic principals
regarding the linear heat density. The concept of linear
heat density, being the division of total annually sold
heat in a district heating system and the total length of
the district heating piping network, indicates the level of
district heat distribution system utilisation. Furthermore,
linear heat density is a denominator parameter when
calculating district heating network capital costs.
L
Q
Density LinearHeat
s
= [GJ/m] (2)
As has been put out in [2], this traditional presentation
of the concept of linear heat density offers no entrance
for estimations of future district heating systems, since
none of the two quantities can be known for yet not
built systems, which is the fundamental reason for
reformulation of the expression by use of demographic
quantities. If combining the two concepts of population
density (p) and specific building space () into the city
planning quantity plot ratio (e), which is suggested in
[2], the concept of linear heat density can be
alternatively expressed as;
w e q
L
Q
s
= [GJ/m] (3)
The three new parameters, specific heat demand (q),
plot ratio (e) and effective width (w), are defined as:
q = Q/A
B
[GJ/m
2
a] (4)
e = p [1] (5)
w = A
L
/L [m] (6)
where
p = P/A
L
[number/m
2
] (7)
= A
B
/P [m
2
/capita] (8)
P = Total population [number]
A
L
= Total land area [m
2
]
A
B
= Total building space area [m
2
]
The concept of effective width hereby plays a key role
in the reformulation of the traditional expression for
linear heat density, and hence, constitutes a central
quantity in model estimations of the feasibility and
viability of future district heating network. If linear heat
density can be said to indicate the level of district heat
distribution system utilisation, the effective width
indicates the distribution system coverage of the land
area at hand.
THE PROBLEM
From a district heating distribution point of view it is
relevant to distinguish between two kinds of land area
low plot ratio situations. The land areas can, principally,
consist of either a wide dispersion of households
spread out over the whole area (A), or households can
be closely limited to only a fraction of the land area (B),
see figure 1.
Figure 1. Low plot ratio land areas, scenario A with wide
dispersion of buildings and scenario B with high
concentration of buildings.
In the first case (A), a district heating distribution grid
would have to cover all of the land area at hand in
order to deliver heat (at very low linear heat density),
while in the latter case (B), the grid could be narrowed
down to the limited area fraction. If, when conducting
district heating feasibility model analysis, plot ratios are
extracted by means of (5), it would be relevant and
recommended to somehow adjust the land area
magnitude in order not to include non-targeted area
fractions. An adjustment to reach this purpose can be
achieved in several different ways, of which Effective
Width compensation suggested in this paper is one
option.
September 5
th
to September 7
th

130

DATA AND VALUES
In the spring of 2009, the authors, both being lecturers
at Halmstad University in Sweden, initiated a pre-study
to be carried out by two Bsc-students at their
department [4]. The study was two-fold in regard of
gathered data. Partly it delivered previously assembled
and crucial data on plot ratios, land areas and trench
lengths in 39 detached house districts heating schemes
in Sweden [5], allowing estimations of effective widths
in these districts, see Figure 2, and partly own collected
data.
y = 27,802x
-0,3731
0
50
100
150
200
250
0,0 0,1 0,2 0,3 0,4
Plot ratio (e)
Effective
width (w)[m]
Detached
houses
Power
(Detached
houses)

Figure 2. Effective width as a function of plot ratio in 39
district heating schemes in detached house districts in
Sweden. Source: [5]

The own collected data of the study refers to data from
34 district heating schemes in multi-family housing
districts in the Swedish cities of Halmstad and
Gothenburg, see Figure 3.
y = 56,622x
-0,41
0
50
100
150
200
250
300
350
400
0 0,5 1 1,5
Plot ratio (e)
Effective
width (w) [m]
Multi-
family
houses
Power
(Multi-
family
houses)

Figure 3. Effective width as a function of plot ratio in 34
district heating schemes in multi family housing districts in
Sweden. Source [4]

On the basis of these results, and when combined in
one common graph, see Figure 4, a power function
were established and presented in [2]. Note that (e)
refers to plot ratio values, not to the natural logarithm
base (e);
15 . 0
8 . 61

= e w [m] (9)
As can be seen in Figure 4, the graph suggests a
convergence at effective width values at 60 meters for
plot ratio values above 1. This would indicate that the
relationship between high dense inner city land areas
and the length of the required piping grid in such areas
is constant.
Still, if plotted explicitly, the function does not converge
at any effective width value, no matter how far the plot
ratio value is extended, but the rate of divergence
decreases with higher plot ratio values. Since plot
ratios values above 3 are considered extremely rare,
effective width values within high dense inner city areas
(plot ratio values above 0.5) can be anticipated to be
found in the interval of 50 < w < 60 meters.
y = 61,838x
-0,1495
0
50
100
150
200
250
300
350
400
0 0,5 1 1,5
Plot ratio (e)
Effective
width (w) [m]
Detached
and MF
houses
Power
(Detached
and MF
houses)

Figure 4. Effective width as a function of plot ratio,
combination of 39 district heating schemes in detached
house districts and 34 in multi family housing districts in
Sweden. Datapoints merged from figure 2 and 3.

40
50
60
70
80
90
100
110
120
130
140
0 0,5 1 1,5 2 2,5 3
Plot Ratio (e)
Effective
Width (w) [m]

Figure 5. Effective width as a function of plot ratio by use
of eq. (9).

For plot ratio values below 0.5, on the other hand
(outer city area and park areas), the relationship is by
no means constant, but diverges rapidly with increased
effective width values as a consequence. At a plot ratio
value of 0.04 the effective width reaches a value of 100
meters, and the curve reveals that the increase of
effective width values at even lower plot ratio values
below 0.04 renders values above 100 meters and
beyond.
The graph characteristics of Figure 5 has significance
for estimations of district heat distribution capital cost
September 5
th
to September 7
th

131

levels in park areas, since these areas often also
include land area fractions without any housing, i.e. not
to be targeted by district heating networks. When using
crude statistical land area values for plot ratio
calculations, there is a potential bias of overestimating
distribution capital cost levels in these suburban areas,
since actual habitations plausibly only occupy parts of
the land area at hand. In these occasions, effective
width values arrived at by use of eq. (9). have a
compensating effect by rapidly increasing its value at
low plot ratio levels, and, hence, rendering lower
anticipated distribution capital cost levels.
CONCLUSION
The main conclusion from this analysis is that the
concept of effective width offers a new simple shortcut
for quick estimations of capital investments for heat
distribution in virgin urban areas.
This conclusion is especially valid if the effective width
has almost a constant value over a plot ratio of 0.5 as
preliminary stated from Figure 4. Further data collection
will show how true this new finding will be.

REFERENCES
[1] Frederiksen S. and Werner S, Fjrrvrme teori,
teknik och funktion (District Heating theory,
technology and function). Studentlitteratur, Lund
1993.
[2] Persson U. and Werner S, The Future
Competitiveness of District Heating, to be
published.
[3] Werner S, Fjrrvrme till smhus vrmefrluster
och distributionskostnader (Sparse district heating
heat losses and distributions costs). Report
1997:11, The Swedish District Heating Association.
Stockholm 1997.
[4] Netterberg H and Isaksson I, District Heating in
Slough. BSc thesis from Halmstad University,
Halmstad 2009.
[5] [Andersson S et al, Nulge Vrmegles Fjrrvrme
(The current situation for sparse district heating),
The Swedish District Heating Association, research
report FoU 2002:74. Stockholm 2002.
September 5
th
to September 7
th

132

INTEGRATING RENEWABLE ENERGY INTO LARGE-SCALE DISTRICT HEATING
SYSTEMS
Peter Begerow, Dr. Stefan Holler
MVV Energie AG, Mannheim, Germany

ABSTRACT
Renewable energy for heating is mostly used in small
systems for single-family houses. The existing district
heating networks are generally run by large heating
plants or combined heat and power plants fired with
fossil fuels.
To combine these two systems, a feasibility study was
completed with a focus on the district heating grid in
Mannheim, Germany, and with a focus on solar thermal
heat. Other renewable energy heat sources,
geothermal heat and heat from biomass, are included
for a comparison.
The study focuses on the heat price as a key figure to
analyse the economic feasibility. The technical
feasibility has been evaluated by using a simulation
model of a secondary district heating grid, which is
operated on a low flow temperature level of 70 C and
which is connected to a central solar thermal energy
plant. The paper describes which technical and
economic framework conditions are necessary for
implementing renewable energy into large-scale district
heating systems. The calculations show that in
comparison with other renewable heat sources solar
heat has the highest heat costs ranging from
7,7 ct/kWh to 14,5 ct/kWh depending on the plant size,
the solar fraction and the use of a storage system. The
major technical problems for integrating solar heat into
a heat grid are the pressure difference between the
flow pipe and return pipe and the low temperature the
flat plate solar collectors are working with.
INTRODUCTION
Based on the protocol of Kyoto and European
regulations a high reduction of CO
2
emissions in
Germany is necessary. To achieve those goals, an
expansion of renewable energy in the heat market is
required. In Germany the major aim to reach is a share
of 50 % renewable energy in the heat market by 2050.
Furthermore, 50 % of the renewable heat is supposed
to be contributed by a heating grid. [13]
To achieve these goals, different governmental as well
as local support mechanism and financial subsidies are
available. If the heat production is combined with an
electricity production, the major financial support is
based on the EEG (German law for renewable energy).
If the used technology just produces heat, a financial
aid from the BAFA (Federal Office of Economics and
Export Control, Germany) or KfW (bank under control
of the Federal Republic, Germany) is possible. There
are different regulations which have to be fulfilled by
the project in order to be eligible for those subsidies [8].
Main criteria are the size and type of the investor and
planner, the type of technology and the size of the heat
plant and of the storage tank.
Previous studies [2], [12], [16], [18] showed that solar
thermal energy is mostly used in single-family houses
and smaller heating grids combined with seasonal heat
storage systems. Those systems are still in
development and need financial support to be realized.
Most of these heating grids run with a lower flow
temperature and use either fossil fuels or heat from
biomass for an auxiliary heat generation. The largest
solar thermal district heat system in Germany is located
in Crailsheim. It covers an area of approx. 7300 m of
solar collectors with two buffer tanks with a combined
volume of about 500 m as thermal storage. In addition
a seasonal geothermal storage has been built which
will cover 50 % of the heat demand for about 2000
residents. This research project has heat production
costs without any financial support of about 19 ct/kWh.
This sum will be reduced depending on the possible
subsidies. [12] Reported technical difficulties were
mostly in the thermal storage technology. There were
little problems with the collectors as common flat plate
collectors were used which are commercially available
and used in large numbers in smaller systems.
The project in Crailsheim has shown the technical
feasibility of a system with a seasonal thermal storage,
but it also shows that considerable costs are involved.
Furthermore, if a single-family house will install a
seasonal storage to get a solar fraction above 50 %, it
needs more than 10 m of hot water storage
(depending on the building type and planned solar
fraction). However, in common buildings there isnt
enough room for that size of storage [17]. Those two
aspects show that the use of a heating grid could
significantly reduce the costs of the solar thermal
systems and could save space otherwise necessary for
a storage tank.
The paper will give an overview of how the expansion
of renewable energy in the heat market will be possible
September 5
th
to September 7
th

133

by integrating renewable heat into a district heating
network. A detailed simulation for a solar thermal
integration was done by using RETscreen [14] as
simulation software.
MATERIAL AND METHODS
Evaluation of heat demand
For planning a new heat production facility, the heat
demand of the connected consumers is necessary. If
those are existing households, the heat demand from
the past can be used for calculations. For newly built
houses the heat demand should be exactly calculated
with the standards named in DIN V 4108-6.
If this is not possible, the yearly heat demand can be
assumed by the given figures:
Table 1: heat demand [15]
Building size
[housing
units]
Heat demand
(room heating)
[kWh/ma]
heating demand
(hot tap water)
[kWh/ma]
1-2 72,3 20
More than 3 55,3 20

These figures can be realized in buildings constructed
between 2011 and 2020. [15]
Another factor for the planning of a heating grid is the
outlook into the future, because the payback period of
a renewable heat production facility is very long.
The following graph shows the expected change in
heat demand for Germany focusing on different factors
of influence:

Fig 1: development of heat demand [11]

Within the following simulation this development is not
further regarded. The major reduction within whole
Germany is based on renovation of old buildings, which
is not part of the simulated area and on the reduction
through population, which has not a direct effect on one
special housing area. The reduction through influence
of temperature has a share below 5 % within 15 years
and is therefore not included within the simulation.
Existing Systems
Between newly built and existing heat networks there
exist some main differences which have to be
considered. If the network is designed especially for the
renewable energy source, it can be technically
specialized (e.g. forced low return temperature for
building owners; special isolation of the used pipes).
Older heating grids on the other hand are normally
constructed for the heat production with fossil fuels and
are normally designed for higher temperatures.
Furthermore, in some heating grids a high temperature
is necessary either for thermal cooling systems (e.g.
absorption chillers) or for the heat transfer stations
within the houses which are built for high temperatures
(low flow temperatures need optimized heat transfer
stations [12]. In the following, the main aspects for the
integration of different sustainable heat generation
technologies are described.
Heat grid for renewable energy
For the integration of renewable energy into heat grids,
different possibilities for the connection exist.
Especially for the solar thermal energy production it is
assumed, that more than one heat plant will be
connected.
The three options are:
1. Taking water from the return pipe, heat it and
return it into the return pipe
2. Taking water from the flow pipe, heat it further
and return it into the flow pipe
3. Taking water from the grid out of the return
pipe and rise the temperature to the necessary
flow pipe value [3]
All of those options have some obstacles. The first
option is normally not welcome by the grid operator
because of higher losses in the system. The second
option is almost impossible for the use of flat plate solar
collectors; because the high flow temperature cannot
be further heated.
The third option shows the best possibility for
integration but has the obstacle with high pressure
differences between the flow pipe and the return pipe.
To evaluate the necessary pump work a first estimation
can be done with equation (1). It gives the pump work
W depending on the necessary heat flow Q, the
September 5
th
to September 7
th

134

pressure difference p and the temperature difference
T between flow and return. Included in this equation
is, the pump efficiency as well as the density and
the thermal capacity of water c
p
.

(1)
For the integration of renewable energy into heat grids
some aspects have to be regarded.
For solar thermal energy the flow and return
temperature of the network is a major problem. This
should be lower than in the existing district heating grid
and run with a temperature of about 60 C / 40 C. [6]
For using deep geothermal heat it depends on the used
technology. If it is combined with the electricity
production, the waste heat after the power plant is
normally below 80 C.
Another major obstacle is the variation of the heat
production and the demand if using solar thermal heat.
During the summer months, the solar radiation is at its
peak, but the heat demand has its peak during the
winter months. To cover a heat grid with a high solar
fraction, a long term thermal storage system is
necessary.
Description of selected site
For modeling the integration of solar thermal energy
into a district heating network, a yet to be built housing
estate was selected. This housing area is planned with
a district heating grid running at a flow temperature of
about 70 C. This area is connected with a heat
exchanger to the central heating grid of the city, which
is run with flow temperatures between 90 C and
130 C.

Table 2: heat demand selected site
Building size
[housing units]
Number of
buildings

Total heating
demand [MWh/a]
(room heating +
hot tap water)
1-2 135 1561
More than 3 111 585
Sum 246 2146

Table 2 gives an overview of the planned houses and
their heat demand. The whole heating grid will have an
length of about 1,3 km and the total heat demand will
be 2146 MWh/a.
Used Software
RETscreen is a program to make first feasibility studies
of all kind of green energy projects. In terms for solar
thermal heating, it uses an included weather database
to calculate the expected heat production. Furthermore
a product database is included with the necessary
technical parameters for many different solar thermal
collectors. The needed amount of heat can either be
calculated other ways or assumed by the software
depending on the amount and size of buildings.
Combining those input factors with others, the
simulation tool gives a recommendation of the used
number of solar collectors and the size of a thermal
storage system. If all input factors are included the
program calculates the yearly heat production and the
solar fraction. Beyond that, the program can be used to
include a second heating system for the remaining
needs to get the final payback period and the total
emissions. For the simulation of this paper the version
4 (November, 2009) of the named software was used.
Financial Calculation
The calculation of the heat costs is based on the net
present value method. For the internal rate of return the
given value was used, all other costs included and the
heat costs varied to get a net present value of zero.
This method gives the current heat price and a further
increase during the next years is included. This makes
it possible to compare the actual heat price to the given
values of other systems. For the economical calculation
in the conclusion of this paper, a competitive heat price
from now on was realized.
The named financial support which is included in the
calculations are subsidies on the capital cost. They
depend, like mentioned in the introduction, on different
aspects. A research project like the one in Crailsheim,
can get a higher support than commercial ones run by
large companies. [1]
Solar thermal heat production
For the heat supply of the given housing area, different
scenarios based on solar thermal energy were
developed. Using the RETscreen software tool the
technical parameters of the flat plate solar thermal
collectors, weather data from a climate database and
the given heat demand of the area was combined for
each scenario.
The scenarios differ in the necessary amount of
collectors needed to achieve a solar fraction of the total
heating demand of 50 % [scenario 1], a 100 % solar
heat production of the used hot tap water (which stays
constant throughout the whole year) [scenario 2] and a
September 5
th
to September 7
th

135

50 % solar fraction of the total heating demand without
a thermal storage system [scenario 3].
For the thermal storage a hot water system is
assumed, because those are state of the art and can
be used in most applications. Other systems have
more specific requirements to the geological situation
of the area. A geothermal heat storage for example
does not work in an area with a flow of the ground
water. For the simulation of scenario 2 a smaller
thermal storage compared to scenario 1 was assumed,
because there is no necessity for a seasonal heat
storage system.
For the simulation model a commercial solar collector
was taken (s. Table 3). It is a flat plate collector with a
anti-reflection glass and a gross area of about 2,6 m.
Its efficiency is 84,4 % (calculated according to EN
12975). The simulation uses specific given parameters.
Those are shown in Table 3.
Table 3: used input parameters for solar simulation
For a comparison of the different scenarios the heat
cost per kWh were calculated.
The calculation of the emissions is based on the
operation of the system and not on its total life cycle.
For solar thermal heat the CO
2
emissions only arise
from the used electricity for the necessary pumps.
Included in the calculation is only the pump energy for
the solar thermal collectors and, if necessary, to
increase the pressure for the integration into the
heating grid flow pipe. The CO
2
emissions for the
German electricity grid are given with 506 g/kWh.
For the calculation without a thermal storage system
[scenario 3] it was assumed that the produced solar
heat can directly be distributed throughout a district
heating network. This would make it possible to save
the investments of a seasonal heat storage system and
also reduce the losses within the thermal storage
system.
For those calculations the same heat amount was used
than in scenario 2. But in this case it is not possible to
cover 50 % of the heat demand of the total grid. Just a
small amount, for example the losses of the grid and
the base load, can be produced with solar thermal
technologies without a thermal storage.
Another option would be to integrate small systems into
the district heating grid. In this case the operator of the
grid would not run the facility by itself. The heat
producer could use a solar thermal collector for its own
heat demand but without a thermal storage system.
Instead of using an in-house thermal storage (what is
getting very large if a seasonal heat storage system is
used) the heating grid could be used. For the single
house technology an internal rate of return of 5 % was
used for the economic calculation (average percentage
of building credit [4]). Furthermore the financial support
is a little different because of different regulations for
large and small systems. In the following those two
calculations are named scenario 4 for the heat
production of a single-family house with a thermal
storage and scenario 5 for the calculation without a
thermal storage.
Summary of different scenarios:
Scenario 1
- Solar fraction of 50%
- Seasonal thermal storage included
Scenario 2
- 100% heat production of hot tap water
- Buffer heat storage included, but no seasonal
thermal storage
Annual heating energy
(calculated with given
method)
Scenario 1: 1076 MWh
Scenario 2: 493 MWh
Scenario 3: 1071 MWh
Scenario 4: 11,6 MWh
Scenario 5: 11,6 MWh
Flow temperature 67 C
Return temperature 45 C
Slope of collector 55
Azimuth of building -45 (southeast)
Type of collector WagnerSolar L20 AR
Storage capacity Scenario 1: 1000 l/m
Scenario 2: 100 l/m
Scenario 3: 1 l/m
Scenario 4: 100 l/m
Scenario 5: 10 l/m
Heat exchanger
efficiency
80 %
Miscellaneous losses 5 % if storage is used
(smaller grid)
8 % if integrated into
large grid
Pump efficiency
(for grid integration)
40 %
Time period 20 year
Internal rate of return Scenario 1: 8,5 %
Scenario 2: 8,5 %
Scenario 3: 8,5 %
Scenario 4: 5,0 %
Scenario 5: 5,0 %
Increase of heat price per
year
2 %
Financial support 30 %
September 5
th
to September 7
th

136

Scenario 3
- Same amount of heat produced than in scenario 1
- No storage; connected to large district heating grid
Scenario 4
- Solar fraction of 100% for a single family house
- Seasonal thermal storage included
Scenario 5
- Same amount of heat produced than in scenario 4
- No storage; connected to large district heating grid
Geothermal heat production
The geothermal heat can be used in various ways for
room heating. Using the shallow geothermal heat is
only possible in combination with a heat pump.
Therefore, for a large integration into heating grids the
deep geothermal energy is the favoured one.
Furthermore in the upper valley of the river Rhein
(Oberrheingraben) the geothermal heat can be used for
a combined heat and power production because of its
high temperature. In Germany this gives the possibility
to get a payment for the electricity based on the EEG
which grows for 3 ct/kWh if the heat is used as well.
For a comparison to the solar thermal heat a
geothermal power plant in Landau, Germany is used as
a reference.
This project began in 2004 and at the end of 2007 the
power plant started its first electricity production. The
first heat output was planned for 2009.
The power plant uses the ORC process (Organic
Rankine Cycle) to generate electricity. A drill hole with
a depth of 3000 m connects to thermal water with a
temperature with up to 160 C which is cooled down
during electricity production to 70 C. The whole yearly
energy output of the power plant is planned to be
22.000 MWh electricity and 9.200 MWh heat. One of
the major benefits of the geothermal heat production is
the base load which is always available. On the other
hand this gives the problem that the heat is also
available in the summer time and needs to be cooled
down in other ways.
The calculated emissions of the power plant are
0 g CO
2
/kWh because the electricity production has no
emissions and for the pumps the own electricity can be
used. [6]
Currently the power plant runs with a limited output due
to small earthquakes in the area of the drilling hole and
does not deliver heat until now. Additional geological
studies are done right now and a heat output should
start after they are finished.
Heat production with biomass
The heat production from biomass is technically very
similar to the fossil fuel powered heating plants.
Therefore the integration into existing district heating
grids is the easiest way compared to the other
renewable energy sources.
The exact technology depends on the used fuels and
therefore the economic calculation is mainly based on
the price development of the biomass.
The emissions of such a system are by way of
calculation zero, because the emitted CO
2
was firstly
bound by the biomass during its growing period. If the
biomass is planted in an area which was deforested for
that, the emissions are not zero any more. The former
forest was a CO
2
sink which does not exist anymore
and should be included in the calculation. Furthermore
the transport and processing of the biomass should be
included. [19]
Fossil fuels for comparison
In our days the district heating grid in Mannheim is fed
with heat from a fossil fuel fired CHP plant. The heat
prices from that system are much lower than the
renewable heat. Looking into the future it mainly
depends on the price development of CO
2
emissions
and the coal price. [8]
The emissions of such a system are very high, even if
the used heat is more or less waste heat. To reduce
those, a CCS technology can be implemented in the
future.
RESULTS
The results of the simulation are shown in Tab. 4 and
Fig. 2 and 3.
In conclusion the heat price is lower if the collector area
increases (economy-of-scale). Furthermore the use of
a district heating grid instead of a thermal storage
lowers the heat cost extremely.
For scenario 1 it is necessary to install a gross area of
3080 m solar thermal collectors. 1076 MWh heat can
be produced in combination with a 2820 m hot water
storage. The heating costs calculated with the given
framework conditions are 11,2 ct/kWh. To operate the
collector area, pumps are needed which consume
electricity. The emissions of that electricity are, based
on the produced heat, 7,9 g CO
2
/kWh.
In scenario 2, 1916 m solar thermal collectors need to
be installed. Combined with a hot water buffer storage
with a volume of 175 m, 494 MWh of heat can be
produced. The financial calculation over 20 years lead
September 5
th
to September 7
th

137

to heating costs of 14,5 ct/kWh. The emissions of such
a system are 8,9 g CO
2
/kWh.
If the heating amount of the 50 % scenario is used but
without a storage [scenario 3], and therefore without
those losses, a much smaller collector area is
calculated. However, the produced heat has to be used
directly within a large heating grid. For such a system
2542 m of solar thermal collectors are needed which
produce 1071 MWh/a. The smaller collector area and
the elimination of a storage system give heating costs
of 7,7 ct/kWh. On the other hand the emissions of such
a system are higher because of the necessary pump
energy for the pressure compensation. The total
specific emissions of that system are 12,2 g/kWh.
Under consideration of scenario 4, the heat costs are
13,8 ct/kWh within a single family house. If a heating
grid would be used for storage and therefore no large
thermal storage is necessary, the heat costs can go
down to about 11 ct/kWh. If the losses of the storage
system are included in the calculation, a smaller gross
area of collectors can be used. Combining all those
savings, the heat cost for a single-family house can go
down to 7,2 ct/kWh (scenario 5). This shows, that there
is a wide margin and a high potential of cost reduction
if a heat grid is used. But it has to be said, that those
heat costs are still much higher than from other heat
generating systems.
Table 4 shows the technical results and parameters for
each calculated scenario. Based on those figures the
financial and ecological calculation where made. Those
results are shown in Figure 1 and 3. For the renewable
technologies the increase of 2% of the heat costs can
easily be included in the calculation. For the fossil (and
the biomass) use, the heat price is highly dependent on
the fuel price development. Therefore a price range is
given on those systems. The reason for the range for
fossil CHP heat CO
2
emissions is that different
references are used.
Table 4: output parameters of simulation

Fig 2: CO2 emissions of different system [1], [5]

Fig 3: heat costs of different system [1], [20]
CONCLUSION
To make the solar thermal heat production economical
compared to the other systems, different aspects have
to be changed. In order to show the potential of cost
reduction for solar thermal heat generation a sensitivity
analyses has been carried out. The following
parameters have been varied in order to reach a heat
prices of around 3,5 ct/kWh in the beginning year. This
is the actual heat price for private customers in
Mannheim.
- Future heat price development
- Change of investment
- Amount of financial support
- Internal Rate of Return
scenario Collector gross
area [m]
Produced
heat
[MWh/a]
Storage
size
[m]
1 3080 1076 2820
2 1916 493,6 175
3 2542 1071 -
4 44 11,6 4
5 29 11,6 -
September 5
th
to September 7
th

138

Fig 4: influence on the heat costs of different factors

Figure 4 shows the influence of the different factors, if
the others stay the same. But it also shows, that by
changing just one aspect, a reduction of the heat cost
to 3,5 ct/kWh is only possible with lowering the capital
costs by 50% of the scenario without a thermal storage
[scenario 3].
Therefore a combination of different factors was done.
The capital costs are also influenced by the financial
support und were calculated separately.
In order to reduce the heat costs down to about
3,5 ct/kWh in scenario 1, a reduction of the capital cost
by 40% combined with financial support of 50% is
necessary, if the heat price will rise with 8% per year.
This is higher than shown in figure 4 and is based on a
high price assumption as reported in [10].
If a lower IRR is assumed (5 %), the capital costs have
to go down to 70 % and a financial support of 40% of
the investment is necessary.
For scenario 3, a rise of the heat price and the lower
IRR (5 %) just need a reduction of capital costs of 10 %
to achieve heat costs of 3,4 ct/kWh. In this case the
assumed financial support of 30% stays the same. This
shows that an economical use of solar thermal energy
within district heating could be achieved. The financial
support of 30% can be possible based on a KfW
program, the capital costs of 90% of the base
investment is possible within a feasibility study and the
5% IRR is an average figure for building loans.
The high requirements to make the solar heat profitable
show, that this technology is not advisable if other
renewable energy sources are available. Furthermore
for the technical integration a low temperature heating
grid is necessary.

Table 5: parameters for economical operation
Scenario 1 Scenario 3
Heat costs 3,5 ct/kWh 3,4 ct/kWh
IRR 5 % 5 %
Capital cost 70 % 90 %
Financial support 40 % 30 %
Heat price
development
8 % p.a. 8 % p.a.

For the near future it might get more interesting to look
on the biomass and geothermal heat, particular if the
heat is needed in a region where high temperatures in
the depth could be exploited or cheap biomass sources
are available. Further research in the solar collector
technology is needed to lower the capital costs and
equally within the thermal storage technology, as it
might get interesting in the future to include those even
in district heating grids with fossil fuels as heat source
to cover peaks in the demand and transfer a surplus
heat production from the summer into the winter
season.
NOMENCLATURE
W [J] work of pump
Q [J] heat flow
p [Pa] pressure difference (flow / return)
T [K] temperature difference (flow / return)
cp [J/(kg*K)] heat capacity
[kg/m] density
pump efficiency
REFERENCES
[1] Begerow, P.; Integration von erneuerbaren
Energien in Fernwrmenetze Eine technische
und wirtschaftliche Analyse aus Sicht eines
Fernwrmeversorgers, Diplomarbeit an der
Universitt Flensburg, MVV Energie AG,
Mannheim; 2010.
[2] Bodmann, M.; Mangold, D.; Nubicker, J.; Raab,
S.; Schenke, A.; Schmidt, T.: Solar untersttze
Nahwrmeversorgung und Langzeit-
Wrmespeicher; Forschungsbericht zum BMWA
Vorhaben; Universitt Stuttgart; 2005
September 5
th
to September 7
th

139

[3] Bucar, G.; Schweyer, K.; Fink, C.; Riva, R.;
Neuhuser, M.; Meissner, E.; Streicher, W.;
Halmdienst, C.; Dezentrale erneuerbare Energie
fr bestehende Fernwrmenetze;
Bundesministeriums fr Verkehr, Innovation und
Technologie; Wien, 2005. Page 15-16
[4] Deutsche Bank; Zinslandschaft;
https://www.deutsche-bank-
bauspar.de/de/media/Zinslandschaft.pdf; 2010.
[5] Fielenbach, H.; Ohl, G.; Schwarzburger, H.:
Effiziente Wohnwrme und hoher Komfort; GBG
Mannheimer Wohnungsbaugesellschaft mbH;
2009.
[6] Frey, M.; Milles, U.: Geothermische
Stromerzeugung in Landau; BINE Projektinfo
14/07; Karlsruhe; 2007.
[7] Heidemann, W.: Solare Nahwrme und saisonale
Speicherung; FVS LZE Themen; Berlin; 2005.
Page 36
[8] Kaltschmitt, M.; Streicher, W.; Wiese, A.:
erneuerbare Energien; Springer Verlag; Berlin;
2006. Page 29
[9] KfW: Programm erneuerbare Energien;
http://www.kfw-mittelstandsbank.de/DE_Home/
Service/Kreditantrag_und_Formulare/Merkblaetter/
KfW-Programm_Erneuerbare_Energien_ 270_
271_272_281_282.jsp; 2010.
[10] Klpsch, M.; Besier, R.; Wagner, A.: Reicht fr
Kunststoffmantelrohre die Standarddmmung?;
Euroheat&Power 38. (2009); issue 12
[11] Lutsch, W.: Neue Wege zur Marktumsetzung
solarer Nah- und Fernwrme; Fernwrme-, Klte-
und KWK-Versorgung: Entwicklungsstrategie;
AGFW; Frankfurt; 2009.
[12] Mangold, D.; Riegger, M.; Schmidt, T.: Solar
Nahwrmeversorgung und Langzeit-
Wrmespeicher; Forschungsbericht zum BMU
Vorhaben; Solites; Stuttgart; 2007. Page 14, 20
[13] Nitsch, J.; Wenzel, B.: Langfristszenarien und
Strategien fr den Ausbau erneuerbarer Energien
in Deutschland; Leitszenario 2009; BMU; Berlin;
2009. Page 53-57
[14] RETscreen Version 4; Natural Resources Canada;
http://www.retscreen.net; 2009
[15] Smolka, M.: kologisch-technische Auswirkungen
dezentraler Energieversorgungsszenarien mit
Blockheizkraftwerken in elektrischen Verteilungs-
netzen; Verlagshaus Mainz GmbH; Aachen; 2009.
Page 18
[16] Solarge: Marstal district heating Plant; Project
Summary; http://solarge.org/index.php?id=1235&
no_cache=1; 14.03.2010
[17] Sonnenhaus-Institut e.V.; http://sonnenhaus-
institut.de/wohnhaeuser.html; 2010.
[18] Ulbjerb, F.: Large-Scale Solar Heating; Hot|Cool;
3/2008; DBDH; Frederiksberg; 2008
[19] Watter, H.; Nachhaltige Energiesysteme;
Vieweg+Teubner; Wiesbaden, 2009. Page 168
[20] Vo, A.: Das Wachstumspotential der Nah- und
Fernwrme - wirtschaftliche und gesetzliche
Voraussetzungen fr den Ausbau; aus: Forschung
und Entwicklung Heft 10; AGFW; Frankfurt, 2005.

September 5
th
to September 7
th

140

SOLAR DISTRICT HEATING (SDH): TECHNOLOGIES USED IN LARGE SCALE SDH
PLANTS IN GRAZ OPERATIONAL EXPERIENCES AND FURTHER
DEVELOPMENTS
M. Schubert
1
, C. Holter
1
and R. Soell
1
1
S.O.L.I.D. Solarinstallationen und Design GmbH, Puchstr. 85, A-8020 Graz,
m.schubert@solid.at

ABSTRACT
S.O.L.I.D. installed three large scale solar plants for
feeding into the citys district heating in Graz in recent
years. These three solar plants have an annual heat
production of 15,8 PJ, the citys grid delivers 2800 PJ
per year. Therefore the integration of solar thermal in a
technical and economical feasible way has to meet the
requirements of Graz existing district heating grid,
which is one of the largest in Austria.
The first plant, at stadium Graz-Liebenau with
1.420 m, has been now for seven years in reliable
operations, with very good power output data.
AEVG Graz, the largest plant in Graz at 4.960 m,
feeds into the gas power station (maximum power of
250 MW) and from there the heat is distributed through
the district heating grid.
The latest plant, at Wasserwerk Andritz with currently
3.860 m, has a buffer storage of 60 m and the
planning for installation of a heat pump is completed.
The plant feeds into the district heating grid and
supports the room heating of a large office building.
This paper presents operational experiences about
three different ways for feeding solar thermal energy
into a large citys district heating grid. Recent
developments like buffer management for combined
district heating and room heating and integration of a
heat pump are outlined.
INTRODUCTION
For reasons of energy security and environmental
protection, the European Union has set a target of 1%
solar fraction in district heating in 2020 and of 5% in
2050 [1].
Solar thermal technology is widespread in the single
family house sector in most European countries. Mainly
for domestic hot water preparation (DHW), but also for
room heating (RH).
In multi-family houses and for heating grids, there are
not yet as many solar thermal plants and the market
begins to develop.
First solar thermal plants for district heating were built
in the 1970s in Sweden. Since then, various plants
have been built mainly in Austria, Denmark, Germany
and Sweden.
Most of these solar plants feed into rather small heating
grids or sub-grids with an annual heat delivery below
50 GWh
th
(180 TJ). In Denmark, this market was
growing rapidly in recent years and is now bigger than
the market for small-scale solar systems for single-
family houses.
In Graz, Austria, solar thermal plants feed into a large
scale heating grid with an annual heat delivery of
830 GWh
th
(2,99 PJ) and a maximum power of
382 MW
th
. Technical parameters and operation
strategies in large scale heating grids are different to
those in small scale grids and solar thermal technology
has to adopt to these circumstances.
Three solar thermal plants in Graz are presented and
the way they are integrated into the citys heating grid.
SDH PLANT DESIGNS IN GRAZ
1. Feeding directly into the district heating grid
plant at stadium Graz-Liebenau
This plant is located on the roof of an ice-skating hall
next to the citys football stadium (Fig. 1).

Fig. 1: Aerial view of solar plant Stadion Liebenau

September 5
th
to September 7
th

141

The return medium of the heating grid is heated up and
transferred to the flow (Fig. 2) [2]. The adaption of solar
thermal technology for the temperature and pressure
levels of the district heating grid were challenging. This
project was realized with standard large scale
collectors (1420 m collector area) of the Austrian
manufacturer kotech and temperature levels in the
district heating flow of above 70 C have to be reached
dependant on the ambient temperature.
During first operation years, detailed monitoring was
done on the plants performance. Dependant on
climate condition, the annual yield of the plant was
between 521 MWh/a and 569 MWh/a. This
corresponds to a specific yield of 370404 kWh/a per
square meter collector area. Also the return
temperature of the heating grid is of great importance
for the performance of the solar plant.

Fig. 2: Hydraulic scheme of solar feed-in at Stadion Liebenau

2. SDH connected to a large scale fossil fuel fired
station plant AEVG Graz
This is the largest solar thermal plant in Austria and it is
installed on four different buildings of the local
collection and recycling station (Fig. 3).
Situated next to the central heating plant, pressure
parameters are favourable for feed-in. Pressure is
higher in return and thus only valves are necessary and
no additional pumps for integration into the district
heating grid.

Fig. 3: Solar plant AEVG Graz
September 5
th
to September 7
th

142

3. SDH for combined room heating and district
heating with buffer and heat pump plant
Wasserwerk Andritz
As solar thermal systems cant always generate the
high temperatures as required for the district heating
grid, other applications were found for temperature
levels below 75 C (Fig. 4).
Solar heat at low temperature level is stored into a
60 m buffer tank and later used for room heating of an
office building (low temperature floor heating). The
buffer is also fed by district heating and thus decreases
the required connected load of the office building.

Fig. 4: solar thermal plant Wasserwerk Andritz
Even lower temperature levels in shoulder seasons and
in winter can be raised by a heat pump. The installation
is planned for the end of 2010. COPs above 4 are
expected, i.e. when heat from the collectors of 26 C is
heated up to 55 C for room heating.
ACKNOWLEDGEMENT
This work is supported by the EU in the project
SDHtake-off (IEE - Intelligent Energy Europe).
REFERENCES
[1] ongoing EU-funded project SDHtake-off
[2] Bucar, G., Schweyer, K., Fink, Ch., Riva, R.,
Neuhuser, M., Meissner, E., Streicher, W.,
Halmdienst, Ch. (2005), FEEt Bestehende fossile
oder teilfossile Fernwrmenetze Einbindung von
dezentraler Energie aus Erneuerbaren
Energietrgern Chancen und Hemmnisse,
Endbericht zu Energie der Zukunft
Forschungsprojekt No 807718 im Auftrag des
BMVIT, publisher: Grazer Energieagentur
Ges.m.b.h.
September 5
th
to September 7
th

143

BIOENERGY COMBINES IN DISTRICT HEATING SYSTEMS:
PROSPECTS FOR A FUTURE GROWTH INDUSTRY?
E. Axelsson
1
, A. Sandoff
2
, C. Overland
2

1
Profu, Gothenburg, Sweden.
2
Department of Business Administration, University of Gothenburg, Sweden.

ABSTRACT
District heating offers opportunities for integration of
bioenergy production (e.g. of biofuel). The aim of this
paper is to assess the environmental benefit and the
economic value of such integration, in order to evaluate
the prospect for bioenergy combines in district heating
systems. Since the detailed characteristics of the
district heating system are crucial for the feasibility for
integration of bioenergy production, the assessment is
based on four real district heating systems. The
environmental evaluation shows that the decrease in
green house gas emissions from a combine are in
proportion to the increase in output of CO
2
neutral
energy products. However, the CO
2
reduction per used
quantity of biomass is higher in conventional combined
heat and power production as long as marginal
electricity is related to high CO
2
emissions. Also the
economic evaluation show ambiguous results: two
cases had negative net present value even for low
discount rates, while the two other cases showed to be
more economically robust. In addition to this, a more
detailed analysis of the industrial conditions for the
integration shows a need for achieving a fit regarding
several operational, strategic and economic
circumstances for this type of business ventures. Two
important conclusions that can be drawn from this is
that: 1) not all district heating systems are suitable for
bioenergy combines 2) there are many barriers for a
wide spread adoption of bioenergy combines.
INTRODUCTION
District heating is a technology that receives increasing
interest as it has great potentials in several ways. One
unique characteristic of the district heating technology
is the use of low temperature energy flows for large
scale energy distribution. In contrast to other energy
transformation technologies (e.g. condensing power or
distributed gas heating), district heating can interact
with energy flows that otherwise do not have any
alternative use (e.g. industrial residual heat). Although
this is one of the competitive advantages of the
technology and a fundamental platform for its business
model, this can further enhance the scoop of the
business: by backward integration it is possible to
increase profitability in other industrial processes with
waste heat as a by-product.

One industrial branch that shows promising prospects
in this respect is bioenergy production, i.e. production
of various kinds of biofuel, biogas and solid biofuel.
Integration of bioenergy production to district heating
production eventuates in a bioenergy combine were the
residual heat from the bioenergy production can be
utilised for district heating. Moreover, the integration
can, in many cases, offer additional positive synergies,
e.g. regarding the use of steam and combustible
by-products.
The fact that worldwide bioenergy production as well as
the number of bioenergy products offered is increasing
is a result of changing demand, which in turn offers
new business opportunities. However, one of the great
issues with large-scale production of bioenergy
products is the growing concern over the negative
externalities (social and environmental aspects as well
as resource efficiency). Since energy production and
consumption shows strong path dependence [1], there
is an urgent need to develop and establish production
technologies that help minimize the negative
externalities. Utilizing the taiga and deciduous forest
resources in the Northern hemisphere for this purposes
is, arguably, a promising alternative. The majority of
these natural resources exist in harvested forests,
typically found in regions with, or suitable for, district
heating.
This paper investigates the prospects of using district
heating production as a base for bioenergy production
and its potential to become a wide spread technology.
For this purpose, we use data from four existing district
heating companies to which a bioenergy production
unit is fitted. By acknowledging the complexity of this
integrative business venture, it is possible to get
credible assessments of the magnitude in energy
efficiency, environmental gains and economic profits.
Equally important is the possibility to detect potential
limitations for bioenergy combines to become a
complement to district heating. Finally, conclusions are
made to acquire clues to important restrictions to a
wide spread adoption.
RESEACH DESIGN
We argue that prospects for becoming a future growth
industry are dependent on the environmental benefits,
economic attractiveness and fit with existing business
context. Hence, these three aspects of joint production
September 5
th
to September 7
th

144

are analysed. The environmental benefits are analyzed
with a system perspective on greenhouse gases (GHG)
emissions, taking into account both on and off site
consequences of introduction of an energy combine;
see Environmental evaluation below. Moreover, the
resource efficiency in the form of CO
2
reduction per
used quantity of biomass is evaluated for each
combine.
The economic benefits of the joint production set up
are analyzed through both a short and long-term
commercial lens. By using discounted cash flow
techniques as a base for this analysis, it is possible to
account for both the yearly consequences as well as
long term economic value; see Economic evaluation
below.
Fit with existing business context is analysed with
respect to input/output markets, production and system
configuration and general business conditions
dominant in the host industry. The analysis focus on
restrictions for short term fit; see Business context
evaluation.
Since the detailed characteristic of the district heating
system is paramount to the feasibility for integration of
bioenergy production, we base our investigation on four
real district heating systems in Sweden with different
compositions. The chosen systems are all of equal size
(500-600 TWh of yearly heat deliveries) established in
towns with 40 000 to 80 000 inhabitants. These
systems are in turn equipped with a bioenergy
production unit that best suits ruling company strategy
as well as operational characteristics and maximizes
energy efficiency. In order to capture the additional
values of these investments, evaluation of each
combine configuration is made in relation to a
reference case consisting of the existing system
(complemented with investments to maintain a
comparable level of production quality). The reference
and combine cases are further described in the
Description of the cases below.
Much effort was put into indentifying efficient technical
solutions that best take advantage of the site-specific
conditions in each system. This work included
everything from choice of equipment, appropriate size
of the integrated production unit and production
strategies over the year regarding output of heat,
electricity and other energy products. To identify
efficient technical solutions an integrative computerized
process was applied, including both the district heating
simulation software MARTES [2], and detailed spread
sheet calculations. In order to guarantee high quality
input data, representatives from these four companies
gave access to technical, environmental as well as
economic data.
Below follows a description of the environmental and
economic evaluation procedure. It is important to stress
that the input data for these assessments only include
the change resulting from the integration of the
bioenergy production. One implication of this approach
is that the environmental benefit of the heat produced
(for district heating) is not included, since one base
condition is that the heat deliveries are the same with
and without bioenergy production. Another implication
is that production units in the district heating system
that are not affected (e.g. base load and peak load
production units) are not included. This system
boundary is also pervading for the Description of the
cases to follow.
Description of the cases
The four district heating systems with reference and
combine cases, respectively, are presented in brief
below. The four objects for the evaluation are also
summarized in Table I. A more comprehensive
description can be found in ref. [3].
Table I. Overview of the reference and combine cases in
the four district heating systems. Economic and energy
data are given for both the reference and combine case,
separated with a slash (ref./combine).
CONFIGURATION
1 2 3 4
Heat deliv.
(TWh/y)
500 530 560 620
Ref. inv. Bio CHP None Bio CHP Bio CHP
Combine
technology
Pyrolysis
Enzymatic
hydrolysis
Acid
hydrolysis
Gasi-
fication
Products Bio oil Ethanol
1
Ethanol FTdiesel
2

ECONOMIC DATA, reference/combine
1 2 3 4
Inv. (M) 74/60 0/144 116/310 146/473
O&M (M/y) 2.3/2.8 0/8.8 3.6/15.8 6.1/11.1
ENERGY CONSUMTION, (GWh/year), ref./combine
1 2 3 4
Biomass 397/244 730/1537 470/1271 362/2970
Others 74/135
3
- - -
ENERGY PRODUCTION (GWh/year), reference/combine
1 2 3 4
Electricity 125/0 218/209 145/55 99/78
Biofuel 0/90 0/444 0/294 0/1336
Others - - 0/384
4
-
1
Besides ethanol also biogas and pellets is produced.
2
Also kerosene and nafta is produced.
3
Fuel oil (21/15) and industrial waste heat (53/120).
4
Biogas (0/114) and Pellets (0/270)

September 5
th
to September 7
th

145

System 1
In the current configuration of this system 15-20% of
the energy demand is covered with fuel oil, which
needs to be reduced. One interesting option could be
to convert biomass into bio oil by pyrolysis and then
use the bio oil in the existing oil boilers. Bio oil that is
not used within the system can be sold (e.g. summer
time). If no pyrolysis reactor is built, a conventional
biofuel fired combined heat and power plant (bio CHP)
will be invested in, building up the reference case.
System 2
In this system, there is no need for new production
units, rather there is a high production capacity,
allowing for integration of a bioenergy production unit.
System 2 has good access to biomass, but might have
difficulties to find a market for large quantities of by-
products. Based on these prerequisites, a suitable
combine technology could be cellulose ethanol
production with enzymatic hydrolysis aiming at high
yield and in-house use of energy by-products.
Regarding the O&M cost for the enzymatic process in
Table I, future enzyme price are assumed [4], With
todays prices, the enzymatic process will not be
profitable.
System 3
In System 3 there is a need for new production
capacity, which is represented by a bio CHP in the
reference case. This system has good access to a
large energy market, which enables output of other
energy products. Hence, a cellulose ethanol plant
based on acid hydrolysis can complement the
reference case investment to build up the combine
case.
System 4
This system is in many aspects similar to System 3, but
ethanol production is not in line with company strategy.
Moreover, System 3 has good access to peat, which
could supplement biomass for a large scale production
unit. Hence, gasification of biomass for production of
synthetic biofuel is evaluated for this system.
Environmental evaluation
The assessment of the environmental implication of
introducing a bioenergy production in an existing
district heating system focuses on changes in
emissions of green house gases (GHG). A system
approach for analysing the changes of GHGs is
applied. This means that besides changes of the direct
emissions on site, also the changes of emissions in
affected parts of the energy systems are included; see
Figure 1. For instance, production of biofuel in the
combines ads to the environmental benefit since fossil
fuels can be replaced, while reduced electricity
production has a negative impact to the environmental
benefit in accordance with marginal electricity
production.
GHG
Direct GHG emissions
GHG
GHG
DH system with
or without bioenergy
production
Power
system
Production,
distribution and
use of biomass
Production,
distribution and use
of transportation
f uel

Fig.1. Illustration of the applied system approach for
assessing the changes of GHGs.
In the assessment, all GHGs of significance are
included [3]: carbon dioxide (CO
2
), dinitrogen oxide
(N
2
O) and methane (CH
4
). For all energy carriers, life
cycle emissions are considered, i.e. both combustion
emissions and well-to-gate emissions such as
emissions from fuel extraction, processing and
transportation. Also leakages are considered when
applicable. How the GHGs for the relevant energy
carriers are assessed are described in brief below, a
more thorough description can be found in [3].The
adopted life cycle GHG emissions associated with
changes in consumption/production of the energy
carriers are summarized in Table II.
Table II. Emission factors for included energy carriers.
ENERGY CARRIER LIFE CYCLE EMISSION
(kg CO2 eq./MWh)
Biomass 14-17
1

High emission elec. (E1) 800
Low emission electricity (E2) 260
Pyrolysis oil 292
Ethanol 307
FT diesel 277
Fuel oil 312
Biogas 207
Pellets 286
1
The lifecycle emission of biomass is dependent on how
the biomass is used in the energy combines (e.g.
hydrolysis for fermentation or gasification)
Biomass
The energy input in all four combines is in the form of
biomass. Production, distribution and use of biomass is
related to GHG emissions. The GHG emission from the
use of biomass differs depending on how the biomass
is used. Combustion raises emissions of both methane
September 5
th
to September 7
th

146

and N
2
O (the CO
2
emission are assumed to be neutral
from a climate perspective), while hydrolysis and
fermentation is not assumed to raise these emissions.
Hence, the net lifecycle emission of biomass differs
between 14-17 kg CO
2
eq./MWh fuel.
Electricity
In all district heating systems, the electricity production
decreases as a consequence of introducing the
combine (see Description of the cases). Any change in
electricity production is assumed to be compensated by
changes in marginal electricity production. For
instance, if the electricity production decreases by 85
GWh/year, it is assumed that other producers will
increase their production by 85 GWh/year. To assess
the environmental impact of this, the decrease has to
be multiplied with a emission factor for marginal
electricity.
There are many opinions regarding the emissions of
marginal electricity. Here we have used a high and a
low level, based on dynamic response for electricity
production with two different developments over a long
time period [5]. By using a high and low figure, the
impact and importance of changes in electricity can be
illustrated in a clear way. For the high figure, the
reference case in [5] is used where lifecycle emissions
of marginal electricity are about 800 kg/MWh
el
. This
marginal electricity is denoted E1 hereon. With more
stringent environmental targets the electricity
production can be carbon lean [5] implying that the long
term lifecycle emissions would be about 260 kg/MWh
el
,
denoted E2 hereon.
Biofuel
As seen in Table I, the evaluated bioenergy combines
have various biofuel products as output. In System 1
pyrolysis oil is produced. The pyrolysis oil is assumed
to replace fossil fuel oil (but is categorized as an biofuel
herein). If lifecycle emissions are regarded according to
the approach in ref. [6] for both pyrolysis oil and fossil
fuel oil, the net GHG reduction for replacing fuel oil with
pyrolysis oil is 292 kg per MWh of pyrolysis oil exported
from the combine. Also the amount of fuel oil used
differs in the combine case from the reference case in
System 1 (see Table I). The net life cycle GHG of this
fuel oil is set to 312 kg/MWh.
In systems 2 and 3 ethanol is produced, which is
assumed to replace gasoline with net GHG reduction of
307 kg per MWh of ethanol reaching the market.
In System 4, three biofuels are produced: Fischer
Tropsch (FT) diesel, nafta and kerosene. All three
products are assumed to replace fossil transportation
fuel with the net GHG reduction of 277 kg/MWh. The
possible leakage of methane from the gasification
process is assumed to be negligible.
Biogas and pellets
In the energy combine of System 3, also biogas and
pellets are produced. The biogas is assumed to be
used as a transportation fuel to replace both petrol and
diesel. The net GHG reduction for replacing fossil
transportation fuel with biogas is set to 207 kg/MWh
including life cycle emission and gas leakage in the
production. The pellets are also assumed to replace
fossil fuel, in this case oil with a net GHG reduction of
286 kg/MWh pellets.
Resource efficiency
With the emission factors in Table II and the energy
flows of the reference and combine case in Table I, the
environmental benefit of the energy combine can be
assessed. However, if biomass is assumed to be a
limited resource from a sustainability point of view, it
makes sense to evaluate the use of biomass from an
efficiency perspective. Hence, the resource efficiency is
assessed as the net GHG reduction potential (in kg
CO
2
eq.) per used quantity of biomass (in MWh). By
comparing this key figure for the reference case with
the combine case for each system, the resource
efficiency of the combines can be evaluated.
Economic evaluation
In order to analyze whether an investment adds
financial value we rely on a standard discounted cash
flow (DCF) model estimating the net present value
(NPV) for each project so that:
( ) ( )
=
+ =
n
t
t
t
r CF NPV
0
1 / (1)
where CF
t
denotes the net cash flow in year t, r is the
future weighted cost of capital and n is the number of
years included in the cost-/benefit analysis. The cash
flow at year 0 indicates the initial outlay. Concerning r,
the weighted cost of capital (WACC), we do not
predetermine a specific hurdle rate; instead we analyze
value added for three different levels of discount rates.
We do so because any statements on the actual
riskiness of the project or an estimation of the WACC
for the companies are outside the reach of this study.
As stated before, when estimating cash flows the point
of departure is a reference object. That is, our NPV
calculations only address the differences in cash flows
between the reference and the bioenergy combine; this
for two reasons. First, only the incremental cash flows
are relevant in a DCF analysis. For instance, in the
case of System 3 they already decided that they would
at least build a combined heat and power (CHP)
facility, and the question is if they gain from making
additional investments in a bioenergy production unit.
Second, by focusing on the differences we do not need
to consider the cost structure in the reference case, it is
treated as a given. Besides simplifying the analysis,
September 5
th
to September 7
th

147

academic access is facilitated as there is no need to
reveal sensitive information.
Table III. Assumptions made for non-site idiosyncratic input
and output prices (/MWh).
Ethanol 78 Biomass 19
FT-diesel 78 Fuel oil 57
Kersone 78 Pellets 25
Nafta 52 Electricity 47
Biooil 47 Electricity excise 0.5
Biogas 68 Electricity certificate
1
21
1
Premium paid to producers of renewable electricity.
Cash flows
The initial outlay is assumed to take place in full at year
0. Yearly operational cash flows are projected by first
estimating an operational cash flow for the first year. As
cash flows are the products of price and quantity, this
estimation is based on the technical analysis in order to
obtain energy flow estimates (see Table I), and then
multiply them with price estimates, to which we add
out-payments for operation and maintenance. We
extrapolate this operational cash flow over the 20 year
long investment horizon with a three percent yearly
growth rate (adjusted for the fact that green certificates
are obtained for fifteen years only). All cash flows are
conservatively assumed to occur at the end of each
year. Next, we add tax payments (assuming an
effective tax rate of 26,3%), tax discounts from
depreciation (according to Swedish tax code), changes
in working capital (approximated by dividing the
difference between in-payments and out-payments of
year t by 12 and subtracting the corresponding value
from year t-1, save for the last year where the
difference is set to zero) and a terminal value (5% of
the initial outlay). Initial outlays are determined by
consulting [7] [19]. Our price assumptions for non-site
idiosyncratic inputs and outputs are presented in
Table III. For translation between different currencies
the following exchange rates were used: 9.6 SEK/ and
6.5SEK/USD.
Sensitivity analysis
We then control the robustness of the NPV estimates
through sensitivity analysis; that is, we examine how
the cost-/benefit analysis is affected when changing a
variable at the time, holding all else equal. We do this
in two steps for each system. First, we illustrate the
changes in estimated NPV by changing yearly in-
payments, yearly out-payments, initial outlay and
terminal value respectively. Second, we show how
yearly in-payments and out-payments respond to price
changes.
By this sensitivity analysis, we can to some degree
compensate for the uncertainty that surrounds our
estimates of initial outlays and terminal value, and we
can see for what potential price changes extra concern
is warranted. Certainly, a drawback with the sensitivity
analysis is that it is just a ceteris paribus analysis and
does not take into consideration the potential
covariance of variables, for instance between ingoing
biomass and outgoing biofuel.
Business context evaluation
The environmental and economic analyses of a joint
production operation act as a starting point for the
business context analysis. A wide-spread adoption
demands not only indications of environmental benefits
and economic profits, but must also offer a fit with the
existing business context. Even though the degree of fit
is defined on company level we will not analyze it as
such. Rather we use the business context of the
studied systems in order to put together a compilation
of restrictions and barriers to a wide-spread adoption.
The magnitude and importance of these will give
important indications of the short term possibilities of
realizing environmental benefits and economic profits
in making bioenergy combines a future growth industry.
The restrictions and barriers are identified through the
fit with existing input/output market situation, production
and system configuration and general business
conditions, (i.e. strategic focus and capacity to absorb
additional risk) dominant in the host company.
ENVIRONMENTAL BENEFITS
As already stated in the Research design, the
environmental benefit from integrating bioenergy
production into an existing district heating system is
assessed as the reduction of GHGs from a system
perspective. As also explained, the net difference
depends on the reference case as well as the
composition of the energy combine. In Figure 2, the
GHG reduction for the included parts of the reference
case and energy combine case of System 3 is
displayed. In the reference case (left bar in Figure 2)
a combined heat and power (CHP) plant biomass is
converted into heat (for district heating) and electricity.
The amount of heat is the same in both the reference
and combine cases and, hence, not considered in the
evaluation of GHG reduction. However, the production
of electricity will change and the system consequences
of that is, as stated, considered by including two
different assumptions for marginal electricity. Assuming
that marginal electricity is related to about 260 kg CO
2

eq./MWh
el
(E2), the electricity produced in the
reference case results in a yearly reduction of 38
Mtonne (dark blue bar to the left in Figure 2). If the
emissions of marginal electricity instead is assumed to
be 800 kg/MWh
el
(E1), the emission reduction would
increase by 78 Mtonne/year (light blue bar) to be in
total 116 Mtonne (dark + light blue bar = E1). The
handling of the biomass is related to GHG emissions
September 5
th
to September 7
th

148

(see Environmental evaluation) and, hence, there is a
negative bar of 8 Mtonne for biomass. To sum up, the
net GHG reduction in the reference case is 30 or 108
Mtonne CO
2
equivalents depending on assumptions for
the marginal electricity.
The combine case of System 3 has lower electricity
production than in the reference case (see Description
of the cases). Consequently, the GHG reduction from
the electricity production is also lower, which is seen as
lower dark and light blue bars for the combine case;
middle stacked bar in Fig. 2. Moreover, the negative
bar for biomass is larger for the combine since more
biomass is used in this case. In the energy combine,
however, bioenergy products such as biofuel (ethanol
in this system), biogas and pellets are produced. As
already explained, these energy products are assumed
to replace fossil fuels and the resulting GHG reduction
from the combine is significant: 188 or 217 Mtonne CO
2

eq. with carbon lean (E2) and carbon intense (E1)
electricity production, respectively.

-200
-100
0
100
200
300
Reference Combine Difference
Electricity, E1-E2*
Electricity, E2
Pellets
Biogas
Ethanol
Biomass
G
H
G
r
e
d
u
c
t
i
o
n

(
M
t
o
n
n
e

C
O
2
e
q
.
/
y
r
)
* additonal emission
reduction/change if
electricity is related
to high CO
2
emissions
Net reduction (E2/E1):
30/108 188/217 158/109

Fig. 2. GHG reduction in System 3 for the reference case,
combine case and the net difference for converting to the
combine.
The dark blue bars are related to marginal electricity
associated to low GHG emission (E2). The additional
emission reduction/change if electricity is related to
high GHG emissions (E1E2) is indicated by the light
blue bars. The total emission/change for E2 is given by
the sum of light blue and dark blue bar.
The implication in terms of GHGs of integrating
bioenergy production in System 3 can be visualised by
moving from the left bar in Figure 2 to the middle bar.
Consequently, the difference of the two bars shows the
GHG implication of converting to an energy combine in
System 3, which is presented in the right hand bar in
the figure. The change from the reference to the
combine case gives rise to GHG reduction from the fuel
products (green bars) However, the electricity
production decreases, implying decreased reduction
(emission increase) and, hence, negative bars for
electricity. As can be seen in the figure, the net GHG
reduction from introducing an energy combine in
System 3 is 158 or 109 Mtonne/year depending on the
assumption for marginal electricity (E2 and E1,
respectively).
The equivalents to the right hand bar in Figure 2 for all
four systems are shown in Figure 3. As can be seen,
the reductions of GHGs are significant in systems 2-4,
especially if the electricity is associated with low
emissions (E2, dark blue bar only). In System 1, the
environmental benefit is negative, even if the marginal
electricity is CO
2
lean.
Significant environmental benefits, as displayed for
systems 2-4, are expected since the combines in these
systems use more biomass, which eventually replaces
fossil fuel in the system approach applied (in system 1
less biomass is used which explains the negative
results for this system). However, if biomass is
assumed to be a limited resource from sustainability
point of view, the use of biomass should also be
evaluated from an efficiency point of view. As explained
in the Environmental evaluation, one measure of
resource efficiency is the GHG reduction potential per
used quantity of biomass. This key figure is presented
in Figure 4 for both the reference case and the
combine case for the four district heating systems
evaluated.

-150
-100
-50
0
50
100
150
200
250
300
350
400
System 1 System 2 System 3 System 4
Others*
Biofuel
Elec., E1-E2
Elec., E2
Biomass
G
H
G
r
e
d
u
c
t
i
o
n

(
M
t
o
n
n
e
)

* biogas
and pellets
Net reduction (E2/E1):
-2/-69 124/119 158/109 321/309

Fig. 3. Environmental benefit from introduction of energy
combines.

As seen in Figure 4, the energy combines are less
resource efficient than the reference cases (generally a
biomass fired CHP plant) if the marginal electricity is
associated with high CO
2
emissions (E1, dark + light
blue bar). However, if the marginal electricity is
associated with low CO
2
emissions (E2, dark blue bar
only), the combines are more resource efficient than
the reference cases. As also can be seen, the resource
efficiencies do not differ dramatically between
systems 24. System 1, however, shows lower
resource efficiency, which can be explained by the fact
that a major part of the produced pyrolysis oil is
consumed internally in the system instead of replacing
fossil fuel off site.
September 5
th
to September 7
th

149

0
50
100
150
200
250
R
e
f
.
C
o
m
b
.
R
e
f
.
C
o
m
b
.
R
e
f
.
C
o
m
b
.
R
e
f
.
C
o
m
b
.
E1-E2
E2
R
e
s
o
u
r
c
e

e
f
f
i
c
i
e
n
c
y

(
k
g

C
O
2
e
q
.
/
M
W
h

b
i
o
m
a
s
s
)

Fig. 4. Resource efficiency of biomass quantified as
GHG reduction per used quantity of biomass.
ECONOMIC VALUE
Whether the cost/benefit analyses return positive NPVs
depend largely on the hurdle rates assigned to them. In
Table IV a summary of the economic results are
presented including the initial outlay, the expected free
cash flow for the first year and estimated NPVs for 4, 7
and 10% discount rates, respectively. With the
exception of System 1, where the bioenergy combine is
actually cheaper than the reference plant, marginal
initial outlays vary between M 140 and 330, and
expected cash flows for the first year of operations
between M -3 and 57. The largest addition to existing
cash flow (both in absolute and relative terms) comes
from the bioenergy combine investment in System 4.

Table IV. Summary of cost/benefit analyses for adding a
bioenergy combine to the reference investment in the
studied systems.
1 2 3 4
Initial outlay (M) - 13.9 144 194 327
Cash flow (My)
-3.4 18.8 15.7 57
NPV (M) for different discount rates
4%
-40 76 -62 362
7%
-27 29 -89 207
10%
-19 -4 -108 101

As also can be seen in Table IV, only two projects are
value adding at a 4% discount rate, and System 4 is
the only one that can bear a 10% discount rate. The
results for System 1 are a bit upside down, since
compared to the reference case the investment cost
and net cash flows are negative for the combine.
System 3, perhaps being the weakest of cases
analyzed, will not show positive figures for any positive
discount rate.

For robustness control purposes, sensitivity analyses
are performed, here presented for System 3. Figure 5
illustrates the estimated NPV consequences from
changes in marginal cash flows, disaggregated into in-
payments, out-payments, initial outlays and terminal
value.

-300
-250
-200
-150
-100
-50
0
50
100
-30% -20% -10% 0% 10% 20% 30%
C
h
a
n
g
e

i
n

N
P
V

(
M
,

1
0
%

d
i
s
c
.

r
a
t
e
)
Change in cash flows
In-payments Out-payments
Initial outlay Terminal value

Fig. 5. Estimated changes in NPV (M) for System 3 as
a result of percentage changes in cash flows assuming a
10% discount rate.

A percent change in either of these, results (ceteris
paribus) in a NPV change, as indicated in the figure. It
is clear that the project is most vulnerable for changes
in in-payments followed by out-payments. Assuming a
hurdle rat of ten percent, a 20% average increase in
yearly in-payments would result in an increase in NPV
of about 100 million. Correspondingly, a 20%
increase in yearly out-payments result in a NPV
reduction of 84 millions. Fig. 5 also show that the
cost/benefit analysis is not very sensitive to changes in
initial outlay and leave no visible mark for changes in
terminal value. The order of importance of NPV impact
of cash flow changes are similar in the other three
systems, where in-payments being the most important
ones.

-25%
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
25%
-30% -20% -10% 0% 10% 20% 30%
C
h
a
n
g
e
i
n
m
a
r
g
i
n
a
l

i
n
-
p
a
y
m
e
n
t
s
Price change
Ethanol Biogas Pellet

Fig. 6. Estimated percentage changes in in-payments for
System 3 as a result of percentage changes in input
prices.

September 5
th
to September 7
th

150

Having established the sensitivity to changes in cash
flows it follows naturally to examine also to what
degree different cash flows changes with respect to
changes in underlying prices. In Figure 6, the relation
between marginal in-payments and prices of ethanol,
biogas and pellets are shown for System 3. It is clear
that ethanol is by far the most important bioenergy
product, where a 20% increase in prices renders a 12%
increase in in-payment.

-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
-30% -20% -10% 0% 10% 20% 30%
C
h
a
n
g
e
i
n
m
a
r
g
i
n
a
l

o
u
t
-
p
a
y
m
e
n
t
s
Price/unit cost change
Biomass Electricity O&M

Fig. 7. Estimated percentage changes in out-payments for
System 3 as a result of percentage changes in input
prices/unit costs.

Similarly, Figure 7 shows how out-payments vary with
input prices. Inputs included in the figure are biofuel,
operations and maintenance (O&M) and electricity
7
.
Not surprisingly, biofuel is the key input, where a 20%
price change results in a 10% change in out-payments,
which in Figure 5 translates to a 42 million change in
NPV.
The sensitivity analyses of System 3 show that minor
changes in underlying factors can result in significant
changes in the NPV estimates. However, a not
insignificant part of the indicated variability in cash
flows should be hampered by the offsetting effects
driven by the probable covariance between prices for
biomass and bioenergy products. To be noticed is that
the order of importance of the inputs in the other three
systems show a similar ranking, where biofuel and
biomass price being the two most important ones.
FIT WITH EXISTING BUSINESS CONTEXT
The environmental and economic evaluations indicate
that the integration of bioenergy production into
medium sized district heating systems can be
associated with both environmental and economic
benefits, but the picture is mixed and ambiguous. From
an environmental point of view, the results are coherent
across all systems: the absolute environmental benefit
of bioenergy production is in proportion to the use of
biomass, since increased use of biomass implies
increased output of CO
2
neutral energy products.
However, from a resource efficiency point of view,
biomass should not be used to replace transportation
fuel as long as the marginal electricity is related to high
CO
2
emissions. One important explanation to the
coherent environmental profiles of the different
bioenergy combine solutions is similar resource
efficiency for the four technologies evaluated. Hence,
our results suggest that it is possible to find different
energy combine with similar resource efficiency.
However, these similarities in resource efficiency do
not indicate similarities in economic attractiveness. In
fact, the economic evaluation seems to suggest that
some bioenergy production technologies are not
currently economic viable for integration with district
heating system. Furthermore, the results indicate that
not all district heating systems are suitable for
integration with a biofuel production unit. Despite being
of the same size, use the same raw material and being
evaluated only on marginal effects on the economic
situation, differences in district heating system
characteristics have a profound impact on the
economic possibilities of energy combine integration. In
this study we have matched every system with a
combine solution in order to maximize the site-specific
opportunities in each system. This opens of course the
possibility that there exist other matches with less
resource efficiency but higher economic profitability.
Even if this can be the case, we would like to point out
that one of the starting points of this study was to base
in-data on the conditions of real systems. This includes
taking various kinds of restrictions into consideration.
Even though these restrictions vary, the ones
prominent in this study can be grouped into four
different categories:
- Proximity to input resources
- Proximity to customers or infrastructure for
transporting the finished products
- Existing production and system configuration
- Dominant business conditions
Proximity to input resources
Some combine solutions (such as the one for System
4) demand huge amounts of biomass. This requires
large areas of regional biomass recourses and little or
no competition over it. Import by sea is an alternative
but it requires production sites close to a harbour.
Proximity to market for the finished product
The production of biogas is one example of both the
importance of proximity to customers and to

7
The electricity in out-payments corresponds to the electricity
used in the bioenergy production unit. In Table 1, only the net
electricity export is displayed.
September 5
th
to September 7
th

151

infrastructure. Only relying on local demand for biogas
is considered too challenging at present time.
Existing production and system configuration
Investments in bioenergy combines are seldom green
field but, as we have shown earlier, have to be adapted
to suit existing heat volumes, demand curves, system
configurations and also production site layout. In one of
the systems, the production site was too small to house
the large amounts of biomass necessary for achieving
an economic profitable size of an ethanol operation.
Dominant business conditions
The results of the study show that two business areas
have an evident influence on the type of bioenergy
combine investments the companies carry out: 1) the
strategic framing of the district heating company and 2)
the risk that these investments innate. Concerning the
first, many of the municipally owners use the utilities to
enhance and to some extent even realize the
environmental visions that are formed and expressed
on the political level. Examples of these found among
the companies represented in this study include;
phasing out fossil fuels, use of local waste resources
and visions of a fossil free cities based around locally
produced bioenergy fuels. When present, strategic
framing has a visible effect on limiting the number of
available alternatives for integrates production.
As stated, the second area that has an significant
influence on the type of bioenergy combine that these
companies consider is the risk that these investments
innate. Due to the municipal ownership, these
companies are inherently dependent on stable
business conditions. The ability to absorb negative
results is strongly limited. The added business risk of
bioenergy production must, if needed, be able to be
absorbed by cash flows from existing operations or a
strong capital base. In principle, this can be done in two
ways, either by keeping the investment relatively small,
or by only accepting business propositions with cash
flows that can be made relatively stable.
In Fig. 8, the operational risk of the investment can to
some extent be visualized by the size of the marginal
cash flows of the different investments. The investment
in system 4 stands out not only because it is the largest
one but also because its in-payment comes from one
source only. If the price correlation with biomass is
high, this might not be a large problem. However, it is
interesting to note the relatively small positive cash flow
available from existing operations in Systems 4, and
also for System 3. If the company carries through with
the evaluated investment, it will dramatically change its
operational risk profile and over-all business focus.
-20
0
20
40
60
80
100
Biofuel
Biogas
Pellets
Biomass
Industrial waste heat
Electricity
Electricity certificate
O&M
Yearly Cash Flows (M)
In-payments
Out-payments
Free cash flow

Fig. 8. Marginal cash flows (in-payments/out-payments)
for each system in comparison to free cash flow from
existing operations in 2007 (shaded bar).

The considerable positive free cash flow of system 2
from its existing operations is explained by the
companys sell of hydropower. Although irrelevant for
the value of this investment, it could function as a
general safeguard against negative results, due to
unfavourable relation between biofuel and biomass
prices.
The investment in system 1 was not profitable
according to the valuation earlier. Despite this, it is
worth pointing out that the risk of this investment
should be low since it uses its own products as input. It
too has, relatively speaking, a strong free cash flow
from its current operation that will decrease the risk of
ending up in the red.
CONCLUSIONS
The results of the bioenergy combine analyses show
that there are indications for both environmental gains
and added economic value of such investments.
However, these benefits seem to be limited by several
operational, environmental and economic
circumstances present in these systems. First, these
investments are dependent on the need for making
major changes in current production layout, typically
the need for new or altered production plants. This
limits the available window of opportunity. There are
also several limitations related to operational
characteristics, availability of input resources and
suitable product markets. A closer investigation of
existing governance situation also shows that these
investments often are made to fit owner strategies
regarding environmental goals of the local energy
system. Finally, the municipally ownership typically
limits the risk appetite which also limits available
investments. The doubtful short term environmental
benefit is a more general objection based on the
valuation of the current marginal power production.
September 5
th
to September 7
th

152

Never the less, it will hamper the potential for wide-
spread adoption of bioenergy combines.
These circumstances lead us to conclude that not all
biofuel production technologies are suitable for all
district heating system. Our economic analyses also
indicate that not all district heating systems are suitable
for bioenergy combine production. In fact the barriers
are so many that it is reasonable to assume they will
effectively reduce the number of systems adopting this
operational design in the near future.
ACKNOWLEDGEMENTS
The main funding for this project is provided by
Fjrrsyn, which is a research program organized by the
Swedish district heating branch agency. Additional
funding is also received from the project Pathways to
Sustainable Energy Systems.
We kindly thank the representatives from each district
heating system for a good cooperation and for
providing us with technical and economic data of their
systems. Without these inputs, the work would not
have been as solid as it is.
We also thank Karolina Nilsson and John Jonsson
(both at Profu) for their valuable contribution to the
work.
REFERENCES
[1] M. Odenberger, F. Johnsson, Pathways for the
European electricity supply system to 2050, Int. J.
of Greenhouse Gas Control, 2010, Vol. 4:2, pp
327-340
[2] J.Sjdin and D. Henning, Calculating the marginal
costs of a district-heating utility, Applied Energy,
2004, Vol. 78:1, pp 1-18.
[3] E. Axelsson, C. Overland, K. Nilsson, and A.
Sandoff, Bioenergikombinat i fjrrvrmesystem,
Fjrrsynsrapport 2009:11.
[4] T. Brandberg, Senior researcher at SEKAB E-
technology, Personal communication, 2009.
[5] H. Skldberg and T. Unger, Effekter av frndrad
elanvndning/elproduktion. Elforsk report (2008).
[6] IVL, Miljfaktabok fr brnslen, IVL Rapport B
1334B-2 (2001).
[7] Svebio, Kraftvrmeutbyggnad 2007-2015, Svebio
repport 2008-03-31.
[8] H. Hansson, S-E. Larsson, O. Nystrm, F. Olsson
and B. Ridell, El frn nya anlggningar - 2007,
Elforsk repport no 07:50 (2007).
[9] M. Zakrisson, Internationell jmfrelse av
produktionskostnader vid pelletstillverkning,
Masters thesis no 29 2002, SLU.
[10] A. Hang and S. Ilic, En frstudie fr bioetanol
produktion i Bors, Masters thesis at Institutionen
Ingenjrshgskolan, Hgskolan i Bors (2008).
[11] M. Lantz, Drivmedelsproducenters
betalningsfrmga fr energigrdor, Milj- och
energisystem, LTH (2006).
[12] J. Benjaminsson and A. Dahl, Uppgradering av
biogas, Presentation at Temadag uppgradering
av biogas, Gteborg (2008).
[13] I. Granberg, Project leader at Jnkping Energi,
Personal commication (2008).
[14] M. Tijmensen, A. Faaij, C. Hamelinck, and M. van
Hardeveld, Exploration of the possibilities for
production of Fischer Tropsch liquids and power
via biomass gasification, Biomass and Bioenergy
2002, Vol. 23.
[15] I. Johansson, S. Larsson and O. Wennberg,
Torkning av biobrnslen med spillvrme,
Vrmeforskrapport 881 (2007).
[16] E. Sandvig, G. Walling, R. Brown, R. Pletka, D.
Radlein, and W. Johnsson, Integrated Pyrolysis
Combined Cycle Biomass Power Systems,
Repport of Alliant Energy, Iowa, USA (2003).
[17] H. Thunman, F. Lind, and F. Johnsson Delstudie
energikombinat, Elforskrapport, 2008.
[18] NREL, Research Advances Cellulosic Ethanol,
NREL (2007).
[19] P. Sassner, M. Galbe, and G. Zacchi, Techno-
economic evaluation of bioethanol production from
three different lignocellolosic materials, Biomass
and bioenergy 2008, Vol 32.

September 5
th
to September 7
th

153

SEA WATER DISTRTICT COOLING FEASIBILITY ANALYSIS FOR TALLINN
A. Hani
1
, I. Britikovski
1
, H. Voll
1
and T.-A. Kiv
1

1
Tallinn University of Technology, Department of Environmental Engineering, Estonia

ABSTRACT
In this paper sea water district cooling feasibility
analysis for Tallinn is presented. It has become more
and more interesting to study alternative solutions for
public buildings A/C cooling due to relatively high
electrical energy prices. Besides economical aspects
technical and environmental sides must be considered.
INTRODUCTION
The first large district cooling systems were developed
during the 1960s in Hartford (1962) and California
(1965) in United States [10]. The first systems in
Europe were La Defense (1967) in France and in
Hamburg (1968) Germany [1]. In the beginning of the
70`s the first system in Japan (Shinjuku) was built [3].
However, because of the energy crises in the end of
70`s, the District Cooling development was slow and no
new large systems were built. Until the end of 80`s
when many new large systems were opened for
example Kioi-cho, Nishi-Shinjuku in Japan and Trigen
Trenton in United States. Also the first district cooling
system of the Nordic countries was installed in Norway.
Operation started in 1989 in Baerum, near Oslo. The
first system in Sweden was built in 1992 in Vsters [2]
and since then the district cooling in Sweden has
developed rapidly. Since the 1990s the establishment
of commercial district cooling systems has increased
rapidly worldwide. Nowadays, more than 20 countries
have a commercial district cooling system and this is
expected to increase rapidly [4].
The sea water (SW) district cooling is based on large
natural cold water source. Enough cold water is
accumulated in lakes, seas, oceans, rivers, etc [8].
Lowering the coolant temperature with sea water is an
alternative to conventional electrical energy consuming
chillers [5]. The system working principle is quite similar
to geothermal energy production which is used in
heating systems [6]. Until now the sea water district
cooling is quite conservatively expanded around the
World.
SW DISTRICT COOLING PRINCIPLE
The temperature in conventional cooling water network
is between +4+7
o
C so applicable the sea water
temperature should be below +5
oC
. Despite that
compressor based cooling can be used in case cooling
water temperature is higher than mentioned [9]. It is
also important to locate the district cooling station near
to energy source.
Typical SW district cooling system principle is indicated
in Figure 1. The system consists of three main
sections:
- Cold sea water pumping;
- Cooling plant with heat exchangers;
- Standard cooling distribution network.

Fig. 1 SW district cooling principle schematic

Heat exchangers allow usage of the soft water in
distribution network while problematic salty sea water
handling will be done in open central circuit.
Environmental impact study is required before any of
the projects will be executed. Large sea water
quantities have to be available to minimize pumping
impacts. In addition to evaluation of the deep zone cold
water pumping, the analysis of recycling the sea water
back to lower sea water zone with higher temperatures
should be made.
Following factors shall be considered before system
design [7]:
- Minimum altitudes between heat exchangers and
water resource level should be designed;
- Centralized district cooling plant (heat exchangers,
pumping station and chillers) is less expensive
than decentralised system;
- Centralized system has less maintenance
problems.
DESIGN PARAMETERS
Temperature of the sea water varies during different
seasons and distance from the coast.
September 5
th
to September 7
th

154

In following Table 1 and Figure 2 the relations of the
SW parameter can be found.

Fig. 2 Temperature and SW depth relation

Tab. 1. SW parameters

From previous studies has been found that cooling
demand exceeds significantly when the outdoor
temperature exceeds 16 C (see Figure 3).

Fig. 3 Ambient temperature and cooling power relation
CASE STUDY TALLINN COSTAL AREA
The study was carried out to research feasibility to
build to the Tallinn costal area cooling plant and district
cooling network with total capacity of 19 MW. Project is
interesting to public buildings which have lower balance
temperature and due to that higher cooling demand.
Study was carried out to construct:
- Cooling plant with 4 water chillers;
- Sea water pumping station (free cooling, pre-
cooling) with 5 heat exchangers;
- District cooling network to customers.
In Tallinn costal area is 21 potential customers whose
cooling demand is app. 19,2 MW. Simultaneous factor
0,85 is assumed. Cooling demand will be covered with
water chillers and SW free cooling. Calculations of 21
public buildings information are presented in Table 2.
Cooling load is calculated 120 W/m
2
(building no 17
cooling load 60 W/m
2
). In calculations was not
considered residential area cooling load due to different
usage profile compared to public areas.
Tab. 2. Preliminary cooling demand calculation
Build.
no
Build.
height,
m
Storeys
above
ground
Public
area m
2

Cooling
demand
kW
1 24 6 8764 1052
2 24 6 18870 2264
3 24 6 1458 175
4 24 6 3564 428
5 24 6 5780 694
6 18 5 5198 624
7 11 2 2340 281
8 18 5 8775 1053
9 24 6 2268 272
10 24 6 2430 292
11 24 6 10260 1231
12 24 6 5049 606
13 24 6 4860 583
14 20 5 24500 2940
15 16 4 4250 510
16 19 5 11200 1344
17 - 4 37221 2233
18 19 5 10500 1260
19 19 5 2200 264
20 19 5 5250 630
21 22 5 3700 444
100% 19,2 MW
85% 16,0 MW

Dist.
from
coast, m
Depth
(sea),
m
Annual
aver.
temp,
o
C
Min
temp,
o
C
Max
temp,
o
C
500 20 7,5-8,5 2,5 17,5
1500 25 5,5-6,5 1,5 17
3200 30 4-5 1 14-16
4000 35 3,5-4 <1 8-15
5500 40 3,5 <1 6-8
September 5
th
to September 7
th

155

Tab. 3 Main technical parameters for design the system
Max cooling demand 19,2 MW
Ambient temp.calc. 27
o
C
Simultaneous factor 0,85
Cooling station capacity 18 MW
Annual average cooling consumption 21600 MWh
Supply water temp 6
o
C
Return water temp (max consumption) 16
o
C
Return water temp (min consumption) 13
o
C

Water chiller cooling
Centralized cooling plant contains up to 4 water chillers
to gain flexibility of the system. Also it is possible to
construct the cooling plant step by step according to
consumers interest and cooling energy demand.
System contains four 4500 kW water chillers with
centrifugal compressors. It is possible to adjust the
cooling power of the unit between 3004500 kW which
makes the system more energy (el) efficient during the
partial load period. The condenser has to be produced
from titan or similar resistant material due to fact that it
is being cooled with sea water. In the following Table 4
are indicated technical parameters for water chillers.
Tab. 4 Water chiller parameters
Cooling power 4x4500 kW
Refrigerant R-134a
Condenser temp. 28
o
C
Seawater (SW) supply temp. 18
o
C
SW return temp. 24
o
C
SW flow (each unit) 215 l/s
Evaporator temp. 3
o
C
COP full load 7
COP partial load 12
District cooling supply temp. 6
o
C
District cooling return temp. 16
o
C
District cooling flow (each unit) 115 l/s

Free-cooling
When sea water temperature is lower than return
temperature from the network free-cooling through heat
exchangers can be used. Optimum logarithmic
temperature difference shall be app. 1,5
o
C. Five heat
exchangers with capacity of 3600 kW are selected,
which assure whole cooling plant capacity (18 MW) in
case sea water temperature is below 5
o
C. Maximal
pressure drop in both circuits is selected 0,85 bar. Heat
exchanger parameters are indicated in Table 5.
Tab. 5 Free-cooling heat exchanger parameters
Heat exchangers capacity 5x3600 kW
Sea water (SW) supply temp. 4,5
o
C
SW return temp. 10
o
C
SW flow 130 l/s
District cooling supply temp. 6
o
C
District cooling return temp. 16
o
C
District cooling flow 72 l/s
Max pressure drop 0,85 bar

Tab. 6 Coolant parameters
Sea water (SW)
SW temp 1,5-18
o
C
Max pressure 6 bar
District cooling liquid
Temperature 10-18
o
C
Max pressure 10 bar

Sea water cooling
The sea water is supplied through insulated 800mm
pipes to pumping station using sea water gravity. Three
pumps (max 1080 m3/h) with frequency converters are
installed using parallel scheme to suction pipe. Sea
water pressure is ca 1,5 m and pumps will add 2 bars
to overcome self-cleaning filters, heat exchangers and
condensers pressure drop. Frequency converters are
used to lower energy consumption during partial load.
Cooling plant operation modes
Cooling plants are designed to have three different
operation modes:
- SW temperature < 5
o
C. Completely free-cooling;
- SW temperature 512
o
C. Pre-cooling with SW +
compressor cooling;
- SW temperature > 12
o
C Only compressor cooling
(free cooling heat exchangers are equipped with
bypasses).
District cooling network
Supply (forward) water temperature is designed 6
o
C.
Return water temperature between 1316
o
C (see
Figure 4).
September 5
th
to September 7
th

156

Due to fact that summer period soil temperature in
1,5 m depth is 10
o
C it is not necessary to insulate the
return pipe of the district cooling network. Supply pipe
is insulated with 10 cm nowadays heat insulation
material.

Fig. 4 District cooling network temp

CONCLUSION
The sea water (SW) district cooling has until year 2000
quite modestly developed among different countries
around the World. Due to the fact that energy prices
have raised rapidly more and more researches for free
energy resources are carried out. Wind power, heat
pumps, solar energy and sea water have obtained
huge attention.
SW district cooling is centralized and will have
advantages like less pollution, less maintenance
problems and in perspective also economic benefits.
Current feasibility analysis was done in Tallinn costal
area to define possible cooling plant load, potential
consumers and technical possibilities.
Due low costal area it is possible to locate the cooling
plant near to sea water. Further studies should add
some more economic aspects to the technical solution.
Problematic is to develop the district cooling network in
Tallinn area (existing tunnels and subways will ease
the process).
Most of the new built or renovated public buildings
have high cooling demand due to glass walls and high
internal heat loads. In present research 21 buildings
with only public area were included (total cooling
demand 19,2 MW). The cooling demand rises
considerably when ambient air temperature exceeds
16
o
C. Sea water temperature 5
o
C can be found in
depth of 3540 m.

The cooling plant shall have three operational modes:
- Free-cooling;
- Pre-cooling + compressor cooling;
- Compressor cooling.
Optimization of the proposed system should be carried
out in further studies.
REFERENCES
[1] Vadrot, A. and Delbes, J, (1999). District Cooling
Handbook a Survey of Techniques, equipment and
Choice of System. European market Group.
Number of pages 208.
[2] Feldhusen. H, Francesc. M. R, (2001). "District
Cooling-Present Market Assessment," Master,
Kungl Tekniska Hgskola, Stockholm division of
Applied Thermodynamics and Refrigeration. pp 52.
Stockholm.
[3] Euroheat and Power, (2003). District Heat in
Europe Country by Country/2003 Survey. Brussel
Belgium.
[4] Mildenstein, B. S. P, (1999). District Heating and
Cooling Connection Handbook.
[5] Gosney. W.B, (1982). Principles of Refrigeration.
Cambringe University Press. Published by the
press syndicate of the University of Cambridge.
[6] Westin, P. E. H., (1999). Production Technologies
in District Cooling Systems and the Importance of
Local Factors. New Energy Systems and
Conversion-NESC 99.). pp 6.Osaka.
[7] Westin, P. E. H., Karlson, B., and Lundqvist, P,
(1999). Straategies and Methods For Increasing
the Capacity of District Cooling Systems.20th
International Conferenss of Refrigeration, IIR/IIF.).
pp 1-8. Sydney.
[8] Nordell, B., and Skogsberg, K, (2002). Snow and
ice storage for cooling applications.Winter Cities
2002.Japan Aomori. Lule University of
Technology
[9] Eliadis, C, (2003). Deep Lake Water Cooling A
Renewable Technology. Number of pages 3.
[10] Morris, A.P, (1995). The Road to Lockport:
Historical Background of District Heating and
Cooling. Ashrae Transactions: Symposia.
[11] Arvidson, J, Asplund, A-L, Birgerrson, E, (1997),
Cold production uning low temperature waste
heat,. Kungl tekniska hgskolan Kemisk
apparatteknik. Pp 54, Stockholm

September 5
th
to September 7
th

157

ANALYSIS FOR THE OPERATION BEHAVIOR AND OPTIMIZATION OF CHP
SYSTEM IN DISTRICT HEATING AND COOLING NETWORK
Yong Hoon Im
1
,

Hwa-Choon Park
1
, Byung-Sik Park
1
and Mo Chung
2
1
Cogen. & Boiler Research Group, Building Energy Research Center,
Korea Institute of Energy Research, Korea
3 Mechanical Eng. Dept., Yeungnam Univ., Korea

ABSTRACT
A simulation program for analyzing the effects of the
networking operation of existing DHC system in
connection with CHP system on-site is to be discussed
in this study. The practical simulation for arbitrary areas
with various building compositions is carried out for the
analysis of operational features in both systems, and
the various aspects of thermal network operation are
highlighted through the detailed assessment of
predicted results. The intrinsic operational features of
CHP prime movers, gas engine, gas turbine etc., are
effectively implemented by realizing the performance
data, i.e. actual operation efficiency in the full and part
loads range.
For the sake of simplicity, a simple mathematical
correlation model is proposed for simulating various
aspects of change effectively on the existing DHC
system side due to the networking operation, instead of
performing cycle simulations separately. The empirical
correlations are developed using the hourly based
annual operation data for a branch of the Korean
District Heating Corporation (KDHC) and are implicit in
relation between main operation parameters such as
fuel consumption by use, heat and power production. In
the simulation, a variety of system configurations are
able to be considered according to any combination of
the probable CHP prime-movers, absorption or turbo
type cooling chillers of every kind and capacity. From
the analysis of the thermal network operation
simulations, it is found that the newly proposed
methodology of mathematical correlation for modelling
of the existing DHC system functions effectively in
reflecting the operational variations due to thermal
network operation. The effects of intrinsic features of
CHP prime-movers, e.g. the different ratio of heat and
power production, various combinations of different
types of chillers (i.e. absorption and turbo types) on the
overall system operation are discussed in detail with
the consideration of operation schemes and
corresponding simulation algorithms. The various
aspects of system configuration in terms of CHP
system optimization are also discussed.
INTRODUCTION
In Korea, the district heating and cooling (DHC) system
gains share of the market steadily and it amounts to
12.3% on the basis of the total number of households
at the end of 2008 [1]. The annual heat sales, via DHC
network, in 2008 have reached 16,676 thousand Gcal
and it increased by about 5% on average after 2001.
Considering the trend of new-town development in
metropolitan areas and newly developing residential
areas on a large scale, it is generally expected to show
a clear increasing trend of DHC systems on the market
for the time being. Furthermore, the relevant changes
of circumstances such as the long-term expectation for
high prices of fossil fuels and the imminent realization
of UNFCCC around the world will help the CHP and
DHC system tighten its grips on the forthcoming
heating and cooling market [2]-[3]. Among the several
merits of DHC systems against separate heat & power
(SHP) or central heating system, the distinctive feature
of being able to construct the networking system with
the neighbouring DHC systems certainly deserves to
receive attention from the view point of efficient use of
energy resources and operation costs reduction [4].
However, the effectiveness of networking operation of
CHP and neighboring DHC systems is strongly
influenced by the conditions of energy consumption
behaviours and corresponding operation scenarios on
both sides. The different pattern of energy consumption
in new demand areas is highly desirable for creating
synergy effects by networking operation. In addition,
the different operation strategy of CHP system with that
of DHC network can also improve the effectiveness of
networking operation. The optimal system configuration
of the CHP system with networking operation certainly
differs from that of stand-alone CHP system not to
mention the operation characteristics. Since the heat
flows in the network are bi-directional, the appropriate
modelling for the mutual effects on each system is
highly required for the accurate estimation of the
networking operation.
The main purpose of this study is to examine the
feasibility of the network operation of the CHP system
on-site with the existing DHC system in terms of
efficient use of primary energy and reduction of the
operation cost. In this study, a simulation program is
developed for analysing the thermal networking
process between the existing DHC system and the
CHP system for the newly developing area. The effects
of thermal networking on the existing DHC system
operation are implemented using mathematical
September 5
th
to September 7
th

158

modelling with empirical correlations for main operative
parameters. The intrinsic features for the CHP prime
movers is modeled using the actual performance data
of operation efficiency in full or part load conditions.
The specific features of the newly developed program
in simulation of thermal networking process in district
heating is described in terms of the energy load
prediction and operation simulation of various system
configurations with CHP prime movers and types of
cooling chillers. The unit energy load model for various
buildings by use, e.g. apartment, hotel, hospital,
buildings for business and commercial use etc, is
introduced for the accurate prediction of energy loads
for newly developing area. The effects of intrinsic
features of CHP prime movers, e.g. the different ratio of
heat and power production, various combination of
different types of chillers (i.e. absorption and turbo
types), on the overall system operation are also
discussed in detail in the following.
MODELLING FOR NET-WORKING OPERATION
1. Modelling of CHP system operation
In the previous studies [5][9], a simulation tool for the
optimal design of the CHP system had been
developed, which is composed with three different
modules of energy load prediction, operation
simulation, and economic analysis modules as shown
in Fig. 1. The main goal of the simulation is to draw an
optimized system configuration for a given target area
by the systematic analysis of the physical and
mechanical behaviour of the CHP system and
corresponding operational cost structure. In principle,
the analysis is performed on hourly basis for a year.
The unit energy load model for a variety of building
types (e.g. apartment, commercial building, office
building, department store, hospital etc.) has been
developed for different types of energy loads, i.e.
heating, cooling, electricity and hot water [10]-[13]. In
energy load prediction module, the hourly, annual
energy demand for a target area is predicted using the
unit energy load models.

Fig.1. Relationship between load, operation and economic
analysis modules [9]
Fig. 2 and Fig. 3 show examples of the daily unit
energy load model of heating for the apartment and
hourly unit energy load model of electricity for the office
building respectively.
The annual hourly unit energy model can be obtained
by synthesizing the daily and hourly unit energy load
models [5]. The final annual hourly energy consumption
for given building compositions and corresponding
scale is to be predicted with the input of the total areas
for respective buildings since the unit energy load
models have been developed by normalizing the
statistical energy consumption measurement data with
the corresponding building areas. The example of
annual hourly energy consumption for the apartment is
shown in Fig. 4.

Fig. 2. Daily unit energy load model for the apartment

Fig. 3. Hourly unit energy load model for the office building

In the operation simulation module, a variety of CHP
system configurations can be considered in terms of
types of prime-movers for the CHP system (e.g. gas
engine, gas turbine, combined CHP, flexible electricity
gas turbine), its capacity, and facility types for cooling
(if cooling load is available) [6]. In the operating
simulation of the CHP system, it is noted that the
physical or mechanical operation results such as fuel
consumption, heat supply, electricity produced by CHP
etc. are calculated by using the operation performance
data for the real products of CHP system, or cooling
facility instead of performing thermodynamic cycle
simulations for respective facilities separately. In order
September 5
th
to September 7
th

159

to implement the schemes, the performance data for
the commercial products, operation efficiency in full
and part load condition, has been extensively
investigated and the database has been realized on the
simulation program.
One can consider a variety of CHP system
configurations with various CHP prime movers and
types of cooling chillers. If the type of CHP prime
movers is being selected, the capacity of it is to be
determined in the form of any percentage on the basis
of the maximum value of annual hourly electricity
demand. Then, the feasible options, which can match
the condition entered by the user, are compiled
according to the relevant algorithm as shown in Fig. 5.

(a) Heating load

(b) Electricity load
Fig. 4. Prediction of annual hourly energy consumption for
the apartment

Fig. 5. Parametric entry of option for CHP system product

The user is to select the most desirable one among the
list of options by referring to the technical specification
for each option such as the unit capacity of the product,
the number of units, and the load factor in terms of unit
capacity. When an option is selected by the user as
described above, its corresponding technical data for
CHP product will be linked automatically in the
subsequent operation simulation procedures. The
settlement of the system configuration for the cooling
system can also be performed in a similar manner by
providing the data for the ratio of being in charge of
turbo or absorption type chillers.
2. Modelling of DHC system for networking
operation
In contrast with small cogeneration or CES system, the
DHC system is not authorized to sell the electricity to
the customer directly in Korea [6]. As a result, the
operation mode differs from that of cogeneration or
CES system, i.e. the facilities are operating depending
on the heat loads, and CHP facilities stop operating
during summer to reduce waste heat production.
Instead, the hot water load during the summer season
is usually supplied from incinerators nearby, or heat
only boilers (HOB). However, the operation schemes of
DHC system for stand-alone operation are bound to be
modified to some extent by networking operation with
CHP system on-site and the appropriate modelling for
such an effect of networking operation on DHC system
is a key element for a reliable prediction of the
operation behaviours due to thermal network operation.
In this study, the changes of operation schemes and
corresponding variations for physical or mechanical
aspects on existing DHC system side have been
realized by employing mathematical correlations for the
sake of simplicity. The mathematical correlations for
energy productions as a function of energy
consumption are developed based on the annual
operation data of a branch of Korea District Heating
Corporation (KDHC). By applying a simple, but credible
empirical correlations instead of performing an
additional cycle simulation for the existing DHC system,
the calculation load and the complexity from the
standpoint of simulation are considerably alleviated.
The procedure to obtain the correlations for energy
production in terms of energy consumption are given
as follows,

The required data for the establishment of the
mathematical correlation is given by,
Annual, heat and electricity production and the sales
per day according to the facilities of heat production
(CHP, HOB, Incinerator)
Annual, fuel consumption per day according to the
facilities of heat production (CHP, HOB)

September 5
th
to September 7
th

160

The functional form of the mathematical correlation is
given as follows,

) , ( P H f F = (1)
Where,
F: Fuel consumption
H: Heating load
P: Electricity load

(a) Time vs. events

(b) Events vs. events
Fig. 6. Illustrative diagram for the correlation between
energy production and fuel consumption
Fig. 6 shows the illustrating diagram for the
mathematical correlation between energy production
and consumptions. For any time t
1
, an optimized
operation scenario already exists and corresponding
heat and electricity production, and fuel consumption
has been fixed according to the operation scenario and
for any time t
2
, it is the same as above. On the basis of
the operation data for a year, the behaviour of system
operation can also be described between dependent
variables (e.g. F: Fuel consumption, H: Heat
production, P: Electricity production). In the correlations
between dependent variables, the time t is reflected
with implicit manner and the meaning of time t in the
correlations is a certain time, not a specific time during
the year as in the original data. For example, if the
DHC system is requested to produce more heat
according to the request from CHP system to the
amount of Q A , the heat load of DHC system can be
regarded to be changed from Q
1
to Q
2
, i.e. Q
2
=Q
1
+ Q A .
Then, the operation behaviour for DHC system at the
moment can be estimated simply from the
mathematical correlations by simply referring the value
of F
2
*, corresponding to Q
2
* and P* corresponding to
F
2
*. It means that one can reconstruct the operation
behaviour of the DHC system as a function of
sequential time reflecting the effects of thermal energy
networks. The correlations for the heat and electricity
production vs. fuel consumption are shown in Fig. 7.
(a) Electricity production vs. fuel consumption

(b) Fuel consumption vs. Heat production

Fig. 7. Developed correlations for the energy productions
vs. fuel consumption

SIMULATION OF THE THERMAL NETWORKING
OPERATION
1. Operation Conditions and Schemes
The operation of the overall system should be carried
out by the order of priority of operation for the various
heat sources. In this study, the basic schemes in order
of priority for supplying the energy demands in newly
developed area are established as shown in Fig. 8,
September 5
th
to September 7
th

161

1. CHP system operation in A
2. Thermal networking operation using CHP in B
3. HOB operation in A
4. Thermal networking using HOB in B

Fig. 8. Schemes of the networking operation

2. Test Case & Energy Loads Prediction
On the basis of the operation schemes for various
available heat sources as described above, the
analysis for the operation behaviour of network
operation of both systems with those of respective
system is performed for two distinct test cases of
residential buildings only, and a group of non-
residential ones.

(a) Case A

(b) Case B

Fig. 9. Comparison of energy load prediction: heating load

(a) Case A

(b) Case B
Fig. 10. Comparison of energy load prediction: elec. Load

Fig.11. Comparison of energy load prediction: cooling load
(Case B)

For case A, the area is only comprised of residential
purpose buildings, i.e. apartments, whereas for case B
it is comprised of non-residential purpose buildings
such as commercial buildings, offices, hotels, and
hospitals. The annual hourly energy load data is
estimated by using the energy load prediction module.
The comparison of predicted energy loads, in the form
of the annual distribution and the cumulative curve, are
given as shown in Fig. 9 to Fig. 11.
3. Operation Simulation Results
For the test case comprised of only residential
buildings, the cooling load is reflected on the electricity
load by assuming that it is covered by the air
conditioner or electric fan in individual houses.
Consequently, the aspect of efficient utilization of the
recovered waste heat during the summer is supposed
to be a decisive factor in the optimization of the CHP
system.

September 5
th
to September 7
th

162

(a) Gas engine

(b) Gas turbine
Fig. 12. Typical pattern of heating load and recovered
waste heat for CHP prime-movers

It is easy to see the typical consumption pattern of
heating and hot water for residential houses in Korea
as shown in Fig. 12. A large variation of heating load is
observed in heat consumption rate and the optimal
design of CHP system with such a large variation is
more difficult than with a relatively regular consumption
pattern. The typical annual operation results of the
respective CHP prime-movers, gas engine and gas
turbine, is also shown in Fig. 12. There is a large
difference in the recovered waste heat prediction for
respective CHP arising from the intrinsic feature for gas
turbine, i.e. higher heat to power ratio of gas turbine
against gas engine.
Figure 13 shows the annual thermal energy supply and
demand operating conditions for two distinct CHP
prime-movers. First of all, the quantity of recovered
waste heat from CHP is not large enough to cover the
whole heat demand in the winter, so that most of heat
demand is covered by HOB operation on-site. It is
noted that the heat supply from DHC network seldom
occurs during the winter. This is mainly due to the fact
that it is also short of heat energy in existing developed
areas during the winter. Of course, it is a probable
scenario to operate the HOB in existing DHC system to
produce the required amount of heat energy for newly
developing area. However, it does not actually happen
because the operation of the HOB on -site has priority
over that of the HOB in existing DHC system according
to the operation schemes.

(a) Gas engine

(b) Gas turbine

Fig. 13. Annual thermal energy supply and demand
operating condition for newly developing area

(a) Gas engine

(b) Gas turbine

Fig. 14. Annual electricity supply and demand operating
condition for newly developing area

The thermal network operation is observed to take
place mainly during the intermediate seasons. It shows
that most of heat demand is covered by the thermal
networking heat supply and it results in bringing down
September 5
th
to September 7
th

163

the rate of operation for HOB on-site considerably.
From the view point of system operation efficiency, it
has a very positive impact in that the rate of operation
of CHP in DHC system increases to some extent.
However, in case of supply of surplus heat to existing
DHC system as shown for gas turbine, it is vice versa.
It is noted that the heat flow of thermal network can be
bi-directional for the gas turbine as shown in Fig. 13.
The annual supply and demand operating conditions
for electricity are shown in Fig. 14. A comparatively
good electricity-tracking operation is observed for both
CHP prime-movers and the supply from the grid tends
to increase during the summer due to the peak of the
electricity demand.

Fig. 15. Variation of electricity production on existing DHC
system side due to thermal networking operation

(a) Gas engine

(b) Gas turbine

Fig. 16. Detailed variation of electricity production on the
existing DHC system side according to CHP prime mover

Fig. 15 shows the variations of electricity production on
existing DHC system side due to thermal networking
operation. It is interesting to note that a minor increase
of the electricity production for existing DHC system is
observed during the intermediate seasons. This is
caused by the increased rate of operation of CHP in
existing DHC system due to thermal networking
operation.
The detailed variation of electricity production on the
existing DHC system side is given as shown in Fig. 16.
The net increase of electricity production for gas
engines is larger than that of gas turbines. This is
because of the intrinsic feature for gas engine CHP
system of smaller heat to electricity ratio than that of
gas turbine, which induce that more heat is supplied to
on-site by the thermal network and consequently
increase the rate of operation of CHP in DHC system.

(a) Gas engine

(b) Gas turbine
Fig. 17. Annual LNG consumption rate for newly
developing area according to the CHP prime mover

The net amount of LNG consumption for newly
developing area is given for different CHP operations
according to the heat source facility as shown in
Fig. 17. It is noted that the composition of LNG
consumption for respective heat source facility varies
considerably. Since there are various special discount
schemes for LNG price in promotion of energy efficient
facilities such as CHP, cooling chillers based on
cogeneration system etc. the reliable estimation of LNG
consumption according to their usage is crucial for the
assessment of economics for the scenarios.
September 5
th
to September 7
th

164

In case B, the analysis of operation characteristics for
drawing optimal system configuration becomes much
more complex due to the existence of cooling load. The
simulation results for gas turbine with 50% of
absorption type cooling (i.e. 50% turbo type cooling)
are given in the following.
Fig. 18 and Fig. 19 show the annual heat load and
operating conditions of thermal energy supply and
demand for on-site.
As shown in Fig. 9, where the heating loads for case A
and B are compared, the heating load for a group of
non-residential building composition is much smaller
than that of residential building composition. As a
result, the waste heat recovered from gas turbine
operation is sufficient enough to encompass the whole
heat loads in case B as shown in Fig. 18. In terms of
thermal networking operation, there is a great change
in the pattern of system operation in that a large
amount of surplus heat energy is available even in
winter not to mention the intermediate seasons. This
means that a large amount of heat is flowing toward the
existing DHC system side as shown in Fig. 19, and
there will be serious effects on the operation of existing
DHC system.

Fig. 18. Heating load and recovered waste heat for gas
turbine CHP

Fig. 19. Annual thermal energy supply and demand
operating condition for newly developing area

(a) Heat

(b) Electricity
Fig. 20. Variation of operating conditions due to thermal
networking operation on the existing DHC system side

The effects of surplus heat energy on the operation
conditions for the existing DHC system side are shown
in Fig. 20. First of all, the considerable reduction for the
rate of CHP system operation during the intermediate
season is observed and it is also expected that the rate
of operation for HOB is to be reduced in the winter as
much as the amount of heat supply from the CHP on-
site.

Fig. 21. Detailed variation of electricity production due to
thermal networking operation on the existing DHC system
side

Consequently, the heat production on the existing DHC
system side is reduced to some extent as shown in
Fig. 20 (a) and it brings about the reduction of LNG
consumption for DHC system. In terms of electricity
production as shown in Fig. 20 (b), there is a minor
variation for the production of it in winter despite the
considerable thermal networking operation. It means
that the CHP system on DHC system side is in full
operation during winter regardless of thermal
networking operation and the shortage of heat energy
September 5
th
to September 7
th

165

is covered by operating HOB. In other words, only the
operation of HOB on the DHC system side is affected
by the thermal networking operation in winter. As
shown in Fig. 20 (b) and Fig. 21, the production of
electricity on the existing DHC system side during the
intermediate seasons is certainly decreasing due to the
supply of surplus heat from CHP on-site, which results
in the diminution of the rate of operation for the CHP on
the existing DHC system side.

Fig. 22. Heat balance of operating the absorption chillers

The operating characteristics for cooling load are
described in the following with Fig. 22. It shows the
heat balance of operating the absorption chillers. The
cooling load exceeding the supply capacity from
recovered waste heat is modelled to be covered by
providing auxiliary heat for absorption chillers by direct
gas combustion. The cooling load assigned to turbo
type chillers is dealt with as an electricity load
converted according to the COP of the corresponding
product of turbo chillers.

Fig. 23. Annual LNG consumption rate by use for newly
developing area of a grope of non-residential buildings
The LNG consumption with the cooling load for newly
developing area is predicted as shown in Fig. 23. Due
to the lower level of heating loads for non-residential
buildings, operation of HOB facility is only permissible
in a limited period even in the winter. It is also noted
that a portion of LNG is consumed to provide auxiliary
heat for absorption chillers by direct gas combustion in
case of shortage of heat from recovered waste heat.
The effects of cooling system configuration on the
network operation characteristics are assessed in detail
as follows:

(a) Absorption type 80%

(b) Absorption type 20%
Fig. 24. Heat balance of operating the absorption chillers
for different responsibility by absorption type cooling
The heat balance of absorption chillers for different
ratio of responsibility by absorption type cooling is
shown in Fig. 24. In case of 80% absorption type
cooling, the recovered waste heat is not sufficient
enough to handle the assigned cooling load, so an
auxiliary heat source, such as direct gas combustion, is
needed to cope with the full absorption cooling load.
Whereas, when the 20% absorption type cooling load
is concerned, the required amount of heat for the
absorption chillers can be supplied only by the
recovered waste heat as shown in Fig. 24. The
remainder of total cooling load is covered by turbo type
cooling system.


Fig. 25. Annual electricity supply and demand operating
condition for newly developing area for different
responsibility by absorption type cooling
September 5
th
to September 7
th

166

The different operation characteristics in terms of
electricity demand and supply is given in Fig. 25. It is
noted that the electricity demand during the summer
increases considerably as the ratio of absorption type
cooling is decreasing. This peak of electricity during the
summer is due to the consumption of electricity for
operating turbo type chillers. From the view point of
design of the CHP system configuration in the
simulation, the cooling load is an important parameter
to be considered carefully, because the capacity of
CHP system is given in the form of any percentage of
the peak value of electricity, i.e. the maximum value of
the annual electricity consumption rate per hour.
Therefore, the criteria for defining the CHP capacity is
to be varied depending on the amount of cooling load
assigned to turbo type chillers. The respective LNG
consumption patterns depending on the ratio of
absorption cooling load are compared in Fig. 26. It is
interesting to note that the composition of fuel
consumed by use is substantially changed according to
cooling load treatment during the summer. The results
confirm that the effects of various aspects of
configuration for CHP and cooling system on the
prediction of operational parameters (e.g. fuel
consumption rate by use) are properly realized in the
simulation program.
By using the simulation approach as presented in this
study, the optimal design of the CHP system in
networking operation with DHC system can be carried
out since one can access the detailed physical data
regarding the whole operation of the network system
such as annual rate of fuel consumption for respective
systems (e.g. CHP, HOB, Chiller etc), annual
production of electricity, heat, and the amount of heat
exchange etc. Along with the appropriate cost
structures for fuel, product sales (heat and electricity)
and the estimation of capital cost, civil construction,
and O&M costs etc, one can also make the
assessement for the economic feasibility of various
scenarios. However, the detailed economic analysis for
the test cases and the procedures to determine the
optimized CHP system configuration based on it will
not be described in this paper due to the page
constraints. These tems will be discussed in further
studies.


Fig. 26. Annual LNG consumption rate by use for different
responsibility by absorption type cooling
CONCLUSION
A simulation program that predicts the energy loads for
a mix of buildings and estimate the operational
characteristics for networking operation of existing
DHC system with CHP system on-site is developed.
The distinctive features of this simulation approach can
be summarized as follows,
The unit energy load models are developed for
accurate prediction of energy consumption by use
accroding to any combiation of building type and scale.
A simple mathematical correlation for reflecting the
variations of the network operation on an existing DHC
system side is newly proposed for the sake of simplicity
and efficient simulation process.
The performance data for the commercial products,
operation efficiency in a full and part load condition,
has been extensively investigated and the database
has been realized successfully on the simulation
program.
The operational characteristics of thermal networking
operation has been assessed in terms of system
configurations for the CHP and the cooling facility as
follows.
According to the intrinsic features of the CHP prime
movers such as gas engine and gas turbine etc, the
aspects for the supply of surplus heat is progressing in
different manners by and large. For a gas engine, the
on-site is almost short of heat so that the prediction
results indicate that the additional operation of CHP on
the exisiting DHC system side is induced in the
intermediate seasons. Whereas, surplus of waste heat
recovered from gas turbine CHP is supplied toward the
existing DHC system side. As a result, the amount of
electricity production is being decreases to some
extent.
In case of a group of non-residential buildings, the
heating load reduces considerably. Therefore, it is
probable that the heating load can be covered by only
the recovered waste heat from on-site even in the
winter. Due to the heat flow toward the DHC system
September 5
th
to September 7
th

167

side in the winter the rate of operation of HOB will be
decreased.
The thermal energy exchanges via the network and
the corresponding changes in operation on both sides
are prevailing in intermediate seasons in case of similar
heat consumption patterns on both sides.
The operation of cooling system on the newly
developing area is verified not to have much effects in
terms of thermal networking operation. However, the
significant changes in the LNG consumption patterns
by use are observed according to the ratio of
responsibility by absorption chillers for the cooling load.
The various aspects of system configuration in terms of
CHP system optimization are discussed with the
development of a simulation program in this study. It is
verified that the physical and mechanical mechanisms
concerned with the thermal networking operation has
been appropriately modeled from the assessment of
operational behavior for test cases.
ACKNOWLEDGEMENT
The author gratefully acknowledges the financial and
technical supports for the research from the Korea
District Heating Corporation (KDHC).
REFERENCES
[1] Korea Energy Management Corporation, Statistics
for district heating and cooling enterprise in Korea,
2009.
[2] A. Marbe, S. Harvey, Opportunities for integration
of biofuel gasifiers in natural-gas combined heat-
and-power plants in district-heating systems,
Applied Energy, 2006, Vol.83, pp. 723-748.
[3] C. Weber, I. Heckl, F. Friedler, F. Marechal, D.
Favrat, Network synthesis for a district energy
system: a step towards sustainability, Computer
Aided Chemical engineering, 2006, Vol. 21, pp.
1869-1874.
[4] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D.
Blumberga, District heating and market economy
in Latvia, Energy, 1999, Vol. 24, pp. 549-559.
[5] H. C. Park, M. Chung, S. H. Kim, Development of
system simulator for community energy system,
Report to Ministry of Industry, 2003.
[6] Y. H. Im, H. C. Park, M. Chung, A study of optimal
heating supply systems for the newly developing
area in the vicinity of DHC system supplying area,
Report to Korea District Heating Corporation, 2006
[7] Y. H. Im, M. Chung, H. C. Park, Feasibility study
for small size cogeneration systems in the
metropolitan areas of Seoul, Final Report to SH
(Seoul Housing) Corporation, 2008.
[8] M. Chung, H. C. Park, Development of a energy
demand estimator for community energy systems,
Journal of the Korean Solar Energy Society, 2009,
Vol 29, pp. 37-44.
[9] M. Chung, H. C. Park, Development of a software
package for community energy system assessment
Part I: Building a load estimator, Energy, in
press.
[10] H. C. Park, S. S. Lee, D. J. Kim, Development of
energy models for department stores, Korean
Journal of Air-Conditioning and Refrigeration
Engineering, 2003, Vol. 15, pp. 1088-94.
[11] H. C. Park, M. Chung, Building load models for
hotels in Korea, Journal of the Korean Solar
Energy Society, 2009, Vol. 29, pp. 48-57.
[12] H. C. Park, Development of weighting factors for
variables associated with hourly energy
consumption pattern for hotels in Korea, SAREK
(Soc. Air-conditioning, Ref., Engineers of Korea)
Winter Annual meeting, 2002, pp. 76-82
[13] H. C. Park, Analysis of energy loads for hospital
buildings, SAREK journal, 2002, pp. 1088-93.
September 5
th
to September 7
th

168

IMPROVED PRIMARY ENERGY EFFICIENCY OF DISTRICT HEATING NETWORKS BY
INTEGRATION OF COMMUNAL BIOMASS-FIRED COMBINED HEAT AND POWER
PLANTS WITH BIOMASS PYROLYSIS
T. Kohl
1
, N.A. Pambudi
2
, T. Laukkanen
1
and C.-J. Fogelholm
1
1
Aalto University, Dept. of Energy Technology, Espoo, Finland
1
Corresponding Author: Thomas Kohl, e-mail: thomas.kohl@tkk.fi
2
Semarang State University, Semarang, IndonesiaAbstract

ABSTRACT
This paper investigates the influence of the integration
of communal biomass-fired combined heat and power
plants with wood-pyrolysis on the plants energy
balance and product distribution. Further the proposed
integration concepts influence on the environmental
performance of the connected district heating network is
pointed out. The environmental performance is
evaluated by means of the primary energy factor and
the CO
2
emission coefficient. For this evaluation, the
European standards EN 15603 and EN 15613-4-5 are
applied and modified.
The concept comprises the integration of a simple
pyrolysis model and of a steam dryer with a base case
combined heat and power plant. The yearly plant output
is calculated by applying a multiperiod model of the heat
duration curve. The work shows that, by co-generation
of valuable pyrolysis product, operation hours and
electricity production can be considerably improved.
The integration also clearly improves the district heating
networks primary energy efficiency and lowers its
carbon dioxide emissions significantly.
INTRODUCTION
The European Unions carbon dioxide mitigation goals
and plans to reduce energy import dependency require
action towards a more sustainable energy supply that is
based on renewable energy sources available in the
member states. Biomass is discussed controversially
due to its wide range of upgrade possibilities from
power, heat, cooling to chemicals and transportation
fuels. Among others, EU directives 2001/77/EC
(promotion of electricity produced from renewable
energy), 2004/8/EC (promotion of
cogeneration) and 2003/30/EC ( promotion of the
use of biofuels) state that the use of biomass for
energy purposes should be expanded on a sustainable
base.
However, the increased use of biomass is expected to
raise prices for biomass which will negatively influence,
among others, the economy of communal biomass-fired
combined heat and power (CHP) plants a technology
that is currently competitive to fossil energy production.
Furthermore the scarcity of the biomass available
demands most efficient use of this resource.
As shown in a previous study [1] it looks promising to
integrate biorefinery processes, that are linked to
transportation fuel production, with CHP plants, since
CHP plants can provide both a source for high
temperature heat needed for thermal conversion of
biomass as well as the district heating network (DHN)
as a sink for sensible heat that would usually be
rejected in stand-alone biofuel refineries. It has been
further worked out that the integrated production of
interstage products, such as liquid fast pyrolysis product
(often referred to as woodoil) and wood pellets, have
several advantages: Firstly, the products are
independent from the transportation fuel market
developments since they can be seen as a universal
input for different upgrading processes to e.g. biodiesel,
ethanol, methanol, hydrogen or other chemicals
production but they can also be directly combusted for
power and heat generation. Secondly, they increase the
biomass energy density making it more sustainable for
transportation to central plants required for economic
fuel production. Thirdly, technologies applied for such
pre-processing are relatively simple and robust, thus
keeping investment cost and system complexity on a
reasonable level and making it therefore also interesting
for local small-scale solutions.
In this paper, outgoing from a base case, we simulate
the retrofit integration of wood fast pyrolysis with an
existing wood-fired CHP plant. The aim is highest
possible pyrolysis product generation using the free
boiler capacity in part loads under the condition that the
district heat (DH) demand is still fulfilled.
With help of a multiperiod model of the DHNs heat
duration curve, the work shows the influence of the
integration on plant operating hours, electricity
production and biomass throughput. In addition the
effects on the DHNs primary energy factor and CO
2

emission coefficients are studied as well. The primary
energy factor and the CO
2
emission coefficient are
calculated according to European standards EN 15603
[2] and 15316-4-5 [3], applying a modified power bonus
method. However, no cost estimation is given, since the
focus of the work was to find out if this integration
September 5
th
to September 7
th

169

concept is possible within the operational limits of the
CHP plant.
In the following, first the used multiperiod load model is
described. Further a brief introduction to biomass fast
pyrolysis is given and it is shown how the process has
been simulated and integrated. Then the modification of
the European standards is explained, results are
presented and finally restrictions of the work and
options for further improvement are discussed.
DISTRICT HEATING LOAD
The CHP plant chosen has been integrated into a virtual
DHN. Therefore yearly data of a real DHN has been
scaled so that the CHP plant provides 60% of the hourly
peak demand of the DHN when on full load. The CHP
plant is assumed to be shut off at 50% load which
corresponds with 30% demand in the DHN. As stated in
[4], those are common operating parameters for
communal solid fuel-fired CHP plants.
In order to represent the yearly production of the base
case plant a multiperiod load model was developed.
One full load and five part load levels have been chosen
to represent the heat duration curve. The pyrolysis
integrated CHP plant is represented by 7 part load
levels since lower DH loads can be supplied, as
explained later. Operating time periods per part load
level are set of equal length and -together with the full
load period- match the total operation hours and yearly
DH generation of 94.5 GWh as shown in figures 1a and
1b. For each load level, fuel input and pyrolysis yield are
then iterated matching the required DH output. DH
demand not provided by the CHP plant is assumed to
be generated in oil-fired heat-only boilers with a thermal
efficiency of 0.85
Heat Duration Curve - Multiperiod Model
Base Case
0
5
10
15
20
25
0 50 100 150 200 250 300 350
Time [d]
D
i
s
t
r
i
c
t

H
e
a
t

L
o
a
d

[
M
W
]
District Heat Load Real CHP DH Load
Multiperiod Model DH Load

Fig. 1a: DH Load Multiperiod Model - Base Case

Heat Duration Curve - Multiperiod Model
Integrated Case - Lower Loads
0
5
10
15
20
25
0 50 100 150 200 250 300 350
Time [d]
D
i
s
t
r
i
c
t

H
e
a
t

L
o
a
d

[
M
W
]
District Heat Load Real CHP DH Load
Multiperiod Model DH Load

Fig. 1b: DH Load Multiperiod Model Integrated Case

CHP PLANT INTEGRATED WITH WOOD PYROLYIS
Wood Pyrolysis Model
Biomass fast pyrolysis is the thermal conversion of
biomass in the absence of oxygen at temperatures of
approximately 500 C and pressures close to
atmospheric [5]. The basic idea of the pyrolysis unit is
derived from the bioliq process developed by the
Forschungszentrum Karlsruhe (FZK). There, fast
pyrolysis is applied in order to yield a high share of
liquid pyrolysis product. Biomass is indirectly heated
and pyrolysed with sand in an inert atmosphere at a
temperature of about 500 C. Subsequently, the
pyrolysis gases are condensed and the liquid fraction
(also referred to as wood oil) is mixed with the coke and
forms the so-called bioslurry which leaves the plant as
the final product. In this work we use data published by
FZK [6] and hence assume that 90% of the biomass
energy is converted into bioslurry whereas 10% accrues
in gaseous form. The pyrolysis gas is thought to be co-
fired in the boiler and hence its energy is subtracted
from the fuel input into the boiler.
As pyrolysis requires a low fuel moisture content of
approximately 10% [5] a dryer must be integrated as
well. Indirect steam drying is applied, since this also
allows the regulation of the DH load. As explained later,
regulation is necessary since the enthalpy of the steam
flow after the modification exceeds the demand of the
DHN and hence must be adjusted.
The wood pyrolysis process is modelled as follows: The
heat of pyrolysis of wood is set to 1.87 MJ/kg (moisture
content 10%) using data for pine derived from [7].
Therewith the pyrolysis yield is calculated from the heat
extracted from the flue gases.
CHP Plant Base Case
A base case CHP plant with a bubbling fluidized bed
boiler (shown in fig. 1) has been simulated in full and
September 5
th
to September 7
th

170

part load using the thermal power plant simulator
Prosim. Performance data of the CHP plant and fuel
input specification is derived from [4] and [8],
respectively and given in table 1. The higher heating
value (HHV) is calculated by the simulation software.

Table 1: Base Plant Input Specification and Performance
at Design Load
Ultimate analysis C 50.64 O 42.22
H 6.10 N 0.16
Ash 0.8 S 0.08
Moisture 50 % HHV 18.8 MJ/kg
Fuel input 26 MW
High pressure steam 60 bars Condenser pressure 0.69 bar
District heat output 16.5 MW Electrical efficiency
el 0.243
Power output 6.3 MW Power to heat ration 0.381
Plant Performance
510C
Simulation model input data - design load
Wood Fuel
Steam Cycle

Fig. 1: CHP Plant Base Case
CHP Plant Integrated Wood Pyrolysis
The modified CHP plant is illustrated in Fig. 2. In order
to provide heat for the pyrolysis process, the heat must
be extracted from the flue gases leaving the fluidized
bed reactor (numbered 3 in Fig.2) boiler at 850 C. The
required amount of flue gas is split off (18) after the
fluidized bed reactor. As in the FZK process, those flue
gases are thought to heat up sand to 550 C (which
would provide the heat for the pyrolysis process by
cooling down to 450 C) [6]. The flue gas thereby is
estimated to cool down to 480 C. The flue gas is then
mixed back (20) into the main flow before the
economizer. The heat extraction needed for biomass
fast pyrolysis process is modelled by help of an
additional evaporator (19). 90% of the biomass energy
on a lower heating value base will form pyrolysis slurry
whereas 10% accrues as pyrolysis gas. The energy
carried by the pyrolysis gas reduces the biomass fuel
input as explained above (Wood Pyrolysis Model).
CHP plant Integrated Steam Drying
The dryer is modelled as a steam tube dryer. Paying
attention to the retrofit situation, live steam is extracted,
throttled to 10 bars and further cooled to 190 C by
spraying in the saturated water leaving the dryer. Drying
of biomass to low moisture contents requires
temperatures far above the saturation temperature at a
given pressure due to the hygroscopic properties of
biomass. Heat consumption of the dryer has been
estimated to 2750 kJ/kg water evaporated [9]. Wood
and hot flue gases are led in the dryer (24). If heat is
available from the flue gases, those are cooled down to
120 C and together with the fully condensing steam
provide the heat needed for the drying process. Dried
wood leaves the dryer at wet bulb temperature. For the
drying process live steam is extracted (21), throttled
(22) to 10 bars and further cooled to 190 C by spraying
in condensate (23) leaving the dryer. The dryer
condensate is throttled to 2 bars (26) and send to the
feedwater tank (15). Flue gas temperatures of 120 C
September 5
th
to September 7
th

171

are considered not to cause sulphur corrosion
problems, especially not with low-sulphur wood fuels.
Certainly this design specification must be reconsidered
in case of changed fuel properties.
The maximum pyrolysis production for each load point
is restricted by the maximum steam extraction rate and
by the boilers maximum burning power. The maximum
possible pyrolysis yield logically requires highest
possible fuel input since heat must be provided both, for
drying and pyrolysis. Conversely, this means that the
steam enthalpy exceeds the demand of the DHN. This
is because the boiler temperature is controlled by
means of the evaporator- and superheater tubes in the
boiler walls. If now, the heat input in the boiler is kept on
a higher level as usual the water amount needed to
dissipate the heat from the boiler walls is only
decreasing to a certain amount (resulting from a
reduced temperature after the economizer).
Consequently, in order to match the DH load, this heat
must now be dissipated in the pyrolysis heat
exchanger (19) or in the dryer (24). By iteration the DH
load is matched by adjusting dryer load, correlated split-
off to the pyrolysis heat exchanger and fuel input. In all
cases the boiler load (characterised by the fuel heat
input) is restricted to 100%. So, the overload back-up
capacity of the boiler is maintained. With this setup the
pyrolysis yield constantly increases with the decrease of
the DH levels down to 60%. The maximum flow off the
dryer (and thus its capacity) is be restricted by the
pressure prevailing in the feedwater tank, which in turn
is given by the extraction pressure after the turbine
stage (11). The pressure decreases with falling live
steam parameters and steam massflow. Hence, there is
a pressure dependant maximum enthalpy flow that can
be fed into the feedwater tank until saturation state is
reached for the mixture of the condensates from the DH
exchanger (13) and the dryer (24). In order to overcome
this restriction the feedwater tank pressure has been
increased load-dependently to a maximum of 2 bars
matching its design pressure. However, due to the
reason mentioned above, for loads below 60% the heat
that would need to be dissipated in the dryer (in order
to match the DH load) would result in such a high dryer
condensate heat flow which again would bring the
feedwater beyond saturation state. Hence for those
cases the boiler load is gradually decreased, resulting in
lower pyrolysis yields. The lowest DH load level that can
be represented is 28.6% of the plants full load.
Compared to a minimum load of 50% in the base case
which is given by the minimum fuel input required for
stable combustion conditions in the boiler-, the
integrated process offers possibilities to increase the
operating hours of the CHP plant considerably.

Fig. 2: CHP Plant with integrated pyrolysis and steam drying
September 5
th
to September 7
th

172

APPLICATION OF THE PRIMARY ENERGY
CONCEPT ACCORDING TO EN 15603
Primary Energy Concept
The EU standard EN 15603 [2] handles the energy
performance of a building as a whole and gives
guidelines how energy use and production of a building
shall be calculated. In order to aggregate the different
forms of energy produced and used within the building,
primary energy (PE) and CO
2
emissions are
accumulated and expressed by means of primary
energy factors (PEF) and CO
2
emission coefficients,
respectively. PE is energy that has not been subjected
to any conversion or transformation process [2]; it is
hence not yet extracted from the source. In the PE
approach described in EN 15603, all energy carriers
involved in the generation process are retraced to their
sources and all energy needed to deliver the final
energy product are aggregated to the total PE
consumption and CO
2
emissions. Thus the PE
approach applies the holistic principles of life cycle
assessment to an energy rating procedure. By retracing
energy consumption to the source, the system
boundaries automatically include the whole world, and
thus depict the real impact of the system concerning
energy consumption and CO
2
emissions.
Primary Energy Factor
The total primary energy factor is the sum of all PE input
to the energy system divided by the useful energy
delivered at the system border. It thus describes how
much PE input is needed in order to obtain one unit of
energy used and can hence be seen as an inverted
efficiency.
In standard EN 15316-4-5 [3] more detailed guidelines
for the calculation of PEFs of DH systems are defined.
According to EN 15316-4-5 PEFs can be calculated for
a certain part of the energy system. In this study the
system boundary comprises the power plants and the
DHN.
The PEF of the DHN has been calculated applying the
power bonus method. If yearly demand data of the DHN
and the generation data are known, the PEF of the DHN
can be calculated by applying the so-called power
bonus method. The power-bonus method is derived
from the energy balance of the building which can be
written as:
. , El DH DH
i
F i F
f P Q f E f + =
,
where E
F
, Q
DH
and P are the heat of the fuels used, DH
and power co-generated respectively. f
F,i
, f
DH
and f
El.
are
the PEFs of the fuels used, the DHN and of the co-
generated power.
In the power bonus method f
El.
is defined as the PEF of
the electricity that is thought to be replaced by the
power generated in the CHP plant (for instance, in this
study the average power generation efficiency in
Finland is used). This allocation pays attention to the
fact that the co-generated electricity is more sustainable
due the CHP process high overall efficiency. The PEF
of the DHN can thus be determined according to;

DH
El
i
F i F
DH
Q
f P E f
f
. ,

=

As production of products other than electricity is not
defined in EN 15316-4-5 the power bonus method has
been extended by regarding the produced pyrolysis
slurry as a bonus as well. The PEF of the pyrolysis
integrated CHP plant is thus calculated as:

DH
Pyro Pyro El
i
i F i F
DH
Q
f E f P E f
f

=
. , ,

In this study PEFs as shown in table 2 have been used:

Table 2: Primary energy factors and CO
2
emission
coefficients for fuels and products
kg/MWh
fBM
2
1.09 cCO2/BM
2
14
fOil
2
1.35 cCO2/Oil
2
330
fEl.
1
3.11 cCO2/El.
1
270
fPyro
1
1.28 cCO2/Pyro
1
14
1
: value is calculated,
2
: value is taken from EN 15603, Annex E

Fuels assumed to be used are wood logs for the CHP
plant and fuel oil for the heat-only/backup boiler(s) and
their PEFs are taken form annex E of EN 15603. The
PEF of electricity production in Finland has been
derived from [10]. The PEF of pyrolysis slurry in a
stand-alone unit has been calculated assuming a flue
gas dryer (which is considered as the drying technology
most likely to be applied) with an energy consumption of
3300 kJ/kg water evaporated [9] and a heat of pyrolysis
of 1.87 kJ/kg [7]. Although the standard asks for more
detailed analysis of the energy chain as e.g.
consideration of transport, transmission and other
processing should be included, this has not been
implemented into this study since those factors are
assumed not to differ between integrated and separated
production of pyrolysis oil.
September 5
th
to September 7
th

173

CO
2
Emission Coefficient
The CO
2
rating is done by calculating CO
2
emission
coefficients (c
CO2
) that quantify the total amount of fossil
fuel derived CO
2
, emitted to the atmosphere, per unit
delivered energy. As for the primary energy factor the
system boundary comprises of power plants and DHN.
Also the power bonus method is applied for calculating
the DHNs specific CO
2
emissions. For the sake of
completeness it must be mentioned that CO
2
-equivalent
emissions of other greenhouse gases can optionally be
included. However this has not been implemented into
this study, due to a lack of data. Similarly to the PEF the
CO
2
emission coefficients for the base case are
calculated as:

DH
El CO
i
i F CO i F
DH
CO
Q
c P c E
c
. , , , ,
,
2 2
2

=
.

And for the modified plant as:

DH
Pyro CO Pyro El CO
i
i F CO i F
DH
CO
Q
c E c P c E
c
, . , , , ,
,
2 2 2
2

=

.

E
F,i
, E
Pyro
, P and Q
DH
represent heat in fuels, heat in
pyrolysis slurry and co-generated electricity and DH
respectively. Accordingly, c
CO2,F,i
, c
CO2,Pyro
, c
CO2,El.
and
c
CO2,DH
are the related CO
2
emission coefficients. The
corresponding values are given in table 2.
RESULTS
In table 3, three simulation cases are presented: the
base case (case 1), pyrolysis integration with the same
operation hours (case 2) and the maximum pyrolysis
slurry production (case 3) with prolonged operation
hours and a DH load as low as 30% (matching 18% of
the total DH load). It can be seen from the table that for
all cases the DH output is the same for the 100-50%
operating points. This results, in the first two cases, in
an identical total DH output of 70.85 GWh. This
corresponds with 75% of the total yearly DH load. Due
to steam extracted to the dryer, the enthalpy flow
through the turbine in part load is decreased, which
results in a lower electricity production in part load for
the cases 2 and 3. Already for the second case
pyrolysis slurry with an energy content in the same
range as the DH load can be produced. Fuel input,
which is defined as wood burned in the boiler and wood
entering the dryer for subsequent pyrolysis, increases
with falling load for load levels 60% and higher. In those
cases the boiler combustion power is 100%, but it is
decreased for lower load levels as explained above. If
operation hours are extended by supplying lower DH
loads with the CHP plant (case 3), total pyrolysis slurry
production can be increased by approximately 55%,
electricity production by 7.8% compared to the base
case. Further DH production is increased by
approximately 14.7%, covering now 86% of the total DH
demand. This directly decreases the fossil fuelled
backup power as shown in table 4. The needed backup
heat is almost cut in half. Together with the additionally
produced electricity this substantially improves the
primary energy factor to 0.68 which certainly will have a
positive influence on the PEF of the buildings connected
to the DHN. For case 2 the improvement is marginal.
The CO
2
emission coefficient changes somewhat
controversially by increasing in the 2
nd
case. This is
because the loss in electricity bonus cannot be
compensated by the produced pyrolysis slurry, since the
CO
2
emission coefficients differ widely. However for
case 3 specific CO
2
emissions become even negative.
The negative value is very unlikely to reach and can be
explained with the not fully accounted fuel production
chain. Nevertheless, it is obvious that the DHNs CO
2

emission factors can be considerably reduced with the
presented integration concept.

September 5
th
to September 7
th

174

Table 3: Results Multiperiod Model
CHP DH Load [%] 100 90 80 70 60 50 40 30 total
Time [h] 2440 530 530 530 530 530 - - 212 days
Fuel Input [MW] 25.90 23.19 20.39 17.47 14.58 11.91 - - 109.56 GWh
Power [MW] 6.29 5.64 4.91 4.06 3.22 2.54 - - 26.13 GWh
District Heat [MW] 16.50 14.85 13.20 11.55 9.90 8.25 - - 70.85 GWh
CHP DH Load [%] 100 90 80 70 60 50 40 30 total
Time [h] 2440 530 530 530 530 530 - - 212 days
Fuel Input [MW] 25.90 36.49 44.24 52.42 60.21 53.88 - - 194.13 GWh
Power [MW] 6.29 5.54 4.69 3.71 2.88 2.25 - - 25.45 GWh
District Heat [MW] 16.50 14.85 13.20 11.55 9.90 8.25 - - 70.85 GWh
Pyrolysis Slurry [MW] 12.21 21.16 30.60 39.58 36.76 - - 74.31 GWh
CHP DH Load [%] 100 90 80 70 60 50 40 30 total
Time [h] 2266 633 633 633 633 633 633 633 279 days
Fuel Input [MW] 25.90 36.49 44.24 52.42 60.21 50.97 39.56 28.88 256.62 GWh
Power [MW] 6.29 5.54 4.69 3.71 2.88 2.27 1.71 1.18 28.17 GWh
District Heat [MW] 16.50 14.85 13.20 11.55 9.90 8.25 6.60 4.95 81.26 GWh
Pyrolysis Slurry [MW] 0.00 12.21 21.16 30.60 39.58 33.79 26.11 19.00 115.46 GWh
Base Case - Case 1
CHP + Pyrolysis - Prolonged Operation Hours - Case 3
CHP + Pyrolysis - Case 2

Table 4: Results - PEF and CO2 Coefficient
Base Case
Case 1
CHP + Pyrolysis
Case2
CHP + Pyrolysis -
Prolonged Operation
Case 3
Required Backup Power MWh 27.8 27.8 15.5
Total PEF [-] 0.80 0.79 0.68
CO2 Coefficient kg/MWh 38.6 42.1 -5.3

CONCLUSION AND DISCUSSION
The work shows that by integration of a CHP plant with
wood pyrolysis operation hours can be increases by
30%, a valuable product can be co-produced and PEE
as well as the CO
2
emission coefficient of the DHN can
be substantially improved. As next steps more
comprehensive data of the fuel supply chain should be
implemented to get more realistic values that will
approve the trend shown with this work. The process
can be further improved by integrating heat that is set
free during the condensation of the pyrolysis liquid and
gaseous product. The heat is available in a
temperature range from approximately 500 C to 25 C
and could hence be used for steam superheating,
feedwater preheating, but also for DH generation. This
integration is not a simple task since many plant
parameters influence each other. The heat integration
must be carried out together with a pinch analysis to
assure an energy efficient integration.
Another open question is the influence of the real
pyrolysis gas on the combustion temperature and flue
gas properties. In order to gather more details of the
pyrolysis process a simple pyrolysis model is currently
under development. Together with the power plant
model the integration can be further optimised aiming
for highest PEE along with low CO
2
emission
coefficients.
Further an economic analysis should be carried out in
order to show potential economic benefits. The
integration itself seems to be viable a statement that
is supported by a press release from June 2009 where
boiler manufacturer Metso and forestry company UPM
announced the development of a new viable fast
pyrolysis process benefitting from the integration with a
CHP plant [11].
Concerning the European standards used for
evaluation, it can be said that the power bonus method
can be easily adapted to a polygeneration concept
September 5
th
to September 7
th

175

yielding heat, electricity and pyrolysis slurry. It very
likely can also be extended to other possible biorefinery
products as long as those are energy products. This
expansion option could be implemented into the
standard.
However the most difficult question remains how the
PEF of other, less common co-products should be
determined. In the case of pyrolysis slurry it is not
possible to find good average production efficiencies
since the technology is not yet on the market. But, how
is co-generation of cooling evaluated?
In general it can be said that the implementation of the
process will be strongly dependant on investment cost
and on the market value of the product. The product
value is currently difficult to predict and also its future
price development will be strongly dependant on the
use of biomass in the future.
Summarising it can be said that even though many
questions still need to be answered, this works shows
clearly that the integration of communal CHP plants
with wood pyrolysis is beneficial concerning the
connected DHNs PEF and CO
2
emission coefficient.
Vice versa it also shows that CHP plants can play an
important role in the sustainable bio-refineries of the
future.
ACKNOWLEDGEMENTS
This work is part of the Primary Energy Efficiency
project of Nordic Energy Research.
The funding of the Graduate School of Energy Science
and Technology (EST) is gratefully acknowledged.
REFERENCES
[1] Kohl, T., Jrvinen, M., Fogelholm, C.J.,
Gasification and biorefinery in combined heat and
power plants, Proceedings of the 11th
Cooling, Reykjavik, Iceland, 2008
[2] EN 15603:2008, Energy performance of buildings.
Overall energy use and definition of energy
ratings, European Committee for Standardization,
CEN, Brussels.
[3] EN 15316-4-5:2007, Heating systems in buildings.
requirements and system efficiencies. Part 4-5:
Space heating systems, the performance and
quality of district heating and large volume
Systems, European Committee for
Standardization, CEN, Brussels, 2007
[4] Savola, T., Modelling biomass-fuelled small-scale
CHP plants for process synthesis optimisation,
Doctoral Dissertation, Helsinki University of
Technology, Espoo 2007
[5] Bridgwater, A.V., 2000, Fast pyrolysis processes
for biomass, Renewable and Sustainable Energy
Rev., 4(1), pp. 1-73.
[6] Henrich, E., 2007, The status of the FZK concept
of biomass gasification, 2nd European Summer
School on Renewable Motor Fuels, Warsaw.
[7] Daugaard, D., Brown, R., Enthalpy for Pyrolysis
for Several Types of Biomass, Energy & Fuels
2003, 17, 934-939
[8] http://www.ecn.nl/phyllis: PHYLLIS is a service
provided by the Energy Research Centre of the
Netherlands ECN, 17.9.2009 [selected
subgroups: untreated wood birch and
fir/pine/spruce].
[9] Brammer, J., Bridgwater, A., Drying Technologies
for an integrated gasification bio-energy plant,
Renewable and Sustainable Energy Reviews 3
(1999) 243-289
[10] Dones, R. et al, 2004, Life Cycle Inventories of
Energy Systems: Results for Current Systems in
Switzerland and other UCTE Countries, ecoinvent
report No. 5, Paul Scherrer Institute Villigen, Swiss
Centre for Life Cycle Inventories, Dbendorf, CH,
p.170.
[11] N.N., press release on
http://www.metso.com/news/newsdocuments.nsf/w
eb3newsdoc/C89A8AC3F77ABD29C22575CF003
111F5?OpenDocument&ch=ChMetsoWebEng
[12] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D.
Blumberga, District heating and market economy
in Latvia, Energy, 1999, Vol. 24, pp. 549-559.
[13] H. C. Park, M. Chung, S. H. Kim, Development of
system simulator for community energy system,
Report to Ministry of Industry, 2003.
[14] Y. H. Im, H. C. Park, M. Chung, A study of optimal
heating supply systems for the newly developing
area in the vicinity of DHC system supplying area,
Report to Korea District Heating Corporation, 2006
[15] Y. H. Im, M. Chung, H. C. Park, Feasibility study
for small size cogeneration systems in the
metropolitan areas of Seoul, Final Report to SH
(Seoul Housing) Corporation, 2008.
[16] M. Chung, H. C. Park, Development of a energy
demand estimator for community energy systems,
Journal of the Korean Solar Energy Society, 2009,
Vol 29, pp. 37-44.
September 5
th
to September 7
th

176

[17] M. Chung, H. C. Park, Development of a software
package for community energy system assessment
Part I: Building a load estimator, Energy, in
press.
[18] H. C. Park, S. S. Lee, D. J. Kim, Development of
energy models for department stores, Korean
Journal of Air-Conditioning and Refrigeration
Engineering, 2003, Vol. 15, pp. 1088-94.
[19] H. C. Park, M. Chung, Building load models for
hotels in Korea, Journal of the Korean Solar
Energy Society, 2009, Vol. 29, pp. 48-57.
[20] H. C. Park, Development of weighting factors for
variables associated with hourly energy
consumption pattern for hotels in Korea, SAREK
(Soc. Air-conditioning, Ref., Engineers of Korea)
Winter Annual meeting, 2002, pp. 76-82
[21] H. C. Park, Analysis of energy loads for hospital
buildings, SAREK journal, 2002, pp. 1088-93.

September 5
th
to September 7
th

177

CHP OR POWER STATION? QUESTION FOR LATVIA
D. Blumberga
1
, G. Kuplais
1
, F. Romagnoli
1
and E. Vigants
1
1
Riga Technical University Institute of Energy Systems and Environment
Kronvalda boulv. 1, LV-1010 Riga, Latvia

ABSTRACT
This paper presents aspects and problems of the
Latvian energy-system connected to the choice of the
CHP and/or power stations for the future national
energy strategies. In the light of the last EU directive in
the subjects of Renewable Energy Sources (RES) the
share of electricity produced from RES at the moment
is attested on the value of 42.4 % but should be
increase to 49.3%. In the same time the share of
renewable energy resources in the final energy
consumption for 2020 should reach the level of 40%
from 30%.
Dependence on imported energy sources, growth of
electricity prices, the need to support local producers
are the main reasons for the use of new renewable
energy technologies in the Latvian energy sector to
implemented in refurbished energy supply system.
Several methods fro the evaluation of the best strategy
are explained.
This apaper summarizes the application of the Energy
Indicators for Sustainable Development (EISD) as good
tool for analyzing trends, setting energy policy goals
and monitoring progress. The results from the
application of a multi-objective optimization regarding
the implementation of the landfill biogas in the biogas
treatement plant Daibe are reported.
1. INTRODUCTION
The structure of energy user in Latvia is characterized
by high energy consumption in households, public and
service sectors, comparing with relatively low
consumption in rural and industrial sector. In light of
this situation, for the power sector development,
special tasks are required in connection to the choice
of the more adequate energy resources in order to
ensure the best energy production and supply.
Consequently question on which direction address the
main efforts for the energetic national improvement is
still actual: CHP or Power station?
If fuel, which is used to produce heat and electrical
energy in Latvia, is taken into account, the dominant
one is gas [1] and consequently appear evident how
the Latvian dependence on foreigner energy supplies
(mainly from Russian) is not only a weak point in
connection to the energetic sustainability but can serve
as a convenient way of exerting economic pressure [2].
In CHP station this dominance is almost total and very
high in district heat supply boiler houses. As it well
known Latvia is a great consumer of imported fossil
from one side but in the same the share of renewable
energy resources is one of the highest of Europe.
The use of specific energy resource depends on
energy supply policy, and total consumption of energy
resources depends on development of every type of
energy resources in regions. Now there is unjustified
high proportion of fossil fuel in state energy balance
which is possible to reduce by a beginning of active
use of local fuel in regions. The EU directive also
requires that Latvia in the year 2015 would generate
49% of electric power from renewable resources
(currently it is 45%) [3]. This is supportable, but the
power supply of Latvia cannot be let out of the sight
and this issue is already problematic.
Latvia has some electricity production from co-
generation plants and some from hydro-power plants.
However, the production of electricity from the hydro-
power plants fluctuates a great deal from year-to-year.
The rest of the electricity for consumption is imported
from the neighbouring countries.
In order to understand the role played by CHP and
power plant it is fundamental to understand the actual
situation in Latvia for thermal energy where more than
a half of Latvia district thermal energy is distributed and
consumed mainly in Riga.
Latvian heating primarily is performed on a centralized
basis and after the used of wood energy the natural
gas imported from Russia is the main source (see
Fig. 1).

Fig. 1. Main resources used for local and individual heat
supply [3]
September 5
th
to September 7
th

178

This type of heating supply scheme means that
consumers are grouped and heating is performed from
heat source which is intended for the consumer group.
About 70 % out of this thermal energy volume is
produced in the cogeneration cycle (only in Latvenergo
owned CHP, and Rigas siltums) and around 30% of
centrally supplied heat energy is produced in Riga CHP
plants and boiler houses. Of course as main fuel in
Riga natural gas is used approximately 98% of thermal
energy is produced from natural gas (CHP plants and
boiler houses together) [3].
As for heat supply outside of Riga, the dominant
thermal energy is produced in boiler houses with
relatively high proportion of local fuel usage (as shown
in Fig.1). Outside of Riga CHP heat production rate
does not exceed 5%.
2. EXISTING ENERGY SITUATION IN LATVIA:
SHORT OVERVIEW
During the recent past central (large) power plants in
Latvia supplied roughly 65% of the total annual power
demand - distributed energy resources (DERs) covered
36%, but the rest were received as import supplies
from Estonia, Lithuania and Russia (mainly) [4].
Regarding fuel sources Latvia has no real fossil-fuels of
its own and the consumption must be imported.
However Latvia uses the domestic renewable-energy
resources hydro-power and biomass.
Table I. primary energy-consumption in Latvia in the
year 2007 [1]

PJ %
Natural gas 56.92 27.8
biogas 0.32 0.16
Biodiesel 0.07 0.03
Oil products 73.33 35.8
Fuelwood 48.47 23.7
Hydroenergy 9.84 4.8
Import of Electricity 10.80 5.3
Import Coal and coke 4.36 2.1
Wind 0.19 0.09
Biodiesel 0.07 0.03
Total 204.6

The use of primary energy for the gross energy-
consumption in Latvia can be seen in Table 1. The
share of renewable energy in the gross energy-
consumption is made up of fuelwood and hydro energy.
That means that there is approximately a total of 30%
of the gross energy-consumption from renewable-
energy sources in Latvia in the year 2007.
The most important domestic renewable-energy
resource in Latvia is biomass in the form of fuelwood:
in fact approximately 45% of Latvia is covered with
woods and this substantial area makes wood a
significant potential as a resource for energy supplies.
Even though the share of renewable is one of the most
large Europe the EU directive fixes the target of 40%
share of renewable energy resources in the final
consumption in 2020.
This means that the increase is not feasible without the
need of refurbishment and/or construction of energetic
infrastructures.
The fact that Latvia has domestic renewable-energy
resources makes it interesting because the utilisation of
the domestic fuels would be sustainable both from an
environmental and an economic point of view.
Latvia has comparatively well developed power, natural
gas supply and district heating systems, and as a
consequence the electricity is basically produced by
hydro power plants and by cogeneration plants, which
are operated according to district heating demand, and
part of electricity is imported (fig. 2). Consequently the
main objectives of the Latvian energy policy now are to
ensure sustainable accessibility to necessary energy
resources and security of supply in order to favourite
the economic growth and improve quality of life, to
ensure environmental quality retention and meet the
objectives set in the Kyoto protocol of UN FCCC and
Latvian Climate Change Program.
Electricity Supply in Latvia
0
1
2
3
4
5
6
7
8
2000 2001 2002 2003 2004 2005
Year
E
l
e
c
t
r
i
c
i
t
y

a
m
o
u
n
t
,

b
i
l
l
i
o
n

k
W
h

imported electicity
wind generators
small HPS
small CHP
CHP
HPP

Fig. 2. Electricity supply in Latvia (Source: state JSC
Latvenergo, Ministry of Economics, Central Statistical
Bureau)
2.1 Lack of energy sources for electricity
The main domestic electricity capacity consists of 1517
MW of hydro and 520 MW [5] of thermal (CHP units in
Riga) all of which is controlled by the state company,
Latvenergo. The generating potential mainly consists of
three hydro power plants (HPP) on the Daugava River,
hence directly dependent on the rivers water flow. Due
to small reservoirs, utilization rates are low and the
September 5
th
to September 7
th

179

production is quite seasonal following the water flows.
The amount of power produced by the Daugava river
HPP cascade is average 2.62.8 TWh [1] annually,
reaching in the years, rich by spring floods and rain
even 4.5 TWh [ [5].
More in detail the three HPPs, located on the river of
Daugava, form a sort of cascade with the relative
capacity of: Plavinas 870 MW, Kegums 263 MW and
Riga 402 MW.
Almost two thirds of hydro electricity is produced in the
spring month of March, April and May. In this period the
supplies are from the hydro plants. In the high demand
winter season amount of electricity generated by hydro
plants is relatively low.
Looking the electricity supply statistics [1] the national
production of electricity is around 10.0 PJ where the
9.8 are produced using hydro energy and 0.2 PJ
produced by wind energy. The net electricity import
(including the amount of energy exported) is around
10.8 PJ approximately the 50% of the national supply.
These figures shows the lacks of energy sources in the
national system and seems reasonable to foreseen a
more large fraction of other energy sources for the
production of electricity, the main question is on which
methodology base this strategy .
2.2 Well organized and developed DH system
Latvian heating primarily is performed on a centralized
basis consequently consumers are grouped and the
heat is supply from heat source which is established for
a certain consumer group. The heat source power,
depending on type of consumer group, varies from the
range of kW to several hundred of MW. In general
lower power can correspond to building groups,
individual houses or even apartments heating.
Residential and separate heating of individual houses
belongs mainly from the decentralized heating. One of
the benefits of district heating is centralization of heat
load, which gives a possibility to increase the heat
source power and to form basis for the development of
cogeneration power. For large heat consumers in
Latvia (mainly heating systems in large cities like Riga)
large cogeneration plants are installed. The customers
who are not connected to a district heating cannot be
provided from this system. In the other regions far from
the big cities the heat supply system is mainly based
on district heating, consequently it means that that
there possibility for a CHP development.
CHP plants cover only a part of the total heat load. The
rest of the load is covered by the peak load boilers.
This means that following the total heat capacity of the
source, the potential heat capacity of cogeneration
should be assessed quantitatively. Heat capacity of
cogeneration plant has to be selected so that operating
equipment with this capacity heat production by
cogeneration would be a maximum.

Fig. 3. Heating energy distribution by cities in Latvia

If we are looking at the district heating division of Latvia
a huge difference can be seen in quantity of heat
supply in Riga and the rest of Latvia (see fig. 3)
Two large CHP plants, Riga TPP-1 with an installed
electric capacity of 144 MW and Riga TPP-2 (390 MW),
are located in Riga [5]. CHP plants are the main heat-
generating sources of heating networks of Latvian
capital. Power is produced mainly in cogeneration
mode, according to the heatload curve.
During the heating season, when there is a substantial
demand for heating and hot water, Riga CHP plants
produce approximately 80% of the total annual
production volume, while during summer the volume of
production reduces [5].
Nowadays Riga CHP plants cover about 20% of the
total annual power demand of Latvia [5] .
The main fuel used in Latvia biggest cities is natural
gas and the rates of thermal energy are 75% - 85% [3].
In Riga and other cities where most part of the heat is
produced in cogeneration cycle, the increase of rates
was not so high and currently (in the autumn of 2009)
heat rates are lower that in the cities where wood chips
are used.
From the thermal energy point of view seventy percent
of the heat in Latvia is supplied from district-heating
systems either from boiler houses or co-generation:
37% of the district heating in Latvia was produced by
means of co-generation plants [6]. This means that
63% of the district heating is produced in boiler houses
[6]. This means that there is potential to replace some
of the heat plants with co-generation units (Eighty
percent of the district heating in Denmark is supplied
from CHP [6]).
As for heat supply outside of Riga, the dominant
thermal energy is produced in boiler houses with
September 5
th
to September 7
th

180

relatively high proportion of local fuel usage. Outside of
Riga CHP heat production rate does not exceed 5% [3]
(combined heat and power plant up to 4 MW of power
operating in Bauska, Valmiera, Ogre, Vangazi,
Daugavpils, Jelgava, Dobele, Grobia, Saldus,
Ventspils, Ozolnieki, dai, Lielvrde and Cesis).
3. METHOD FOR EVALUATION
In connection to achieving sustainable development on
global scale the correct and judicious use of resources,
technology, appropriate economic incentives and
strategic planning at the local and national levels is
required. Therefore, choosing energy fuels and
associated technologies for the production, delivery
and use of energy services, it is essential to take into
account economic, social and environmental
consequences. The research on criteria and/or
indicators in order to understand the best energetic
choice for Latvia is the first step for a correct energy
planning.
There are several methodologies that can be chosen to
identify the most suitable indicators, and in the same
time the choice is related and strictly connected on
what the planning and consequently analysis is based
on.
3.1 Criteria and indicators
The methodologies can be chosen using several
methodological tools and approach such us: multi-
criteria or multi-objective optimization (MOO) [7],
energy indicators for sustainable development (EISD)
[8], Life Cycle assessment (LCA) [9, 10]. Each of these
methodology start from different point of views and
bases: MOO methodology is connected to best
optimization choice of a certain number of variables
that optimize certain objectives, EISD methodology
aims to evaluate (and consequently increase) the
concept of sustainability based on social, economical
and environmental indicators, LCA aims to figure out
the global environmental load of a process and/or
product taking into account the entire outflows and
inflows connected (in terms of energy, substances and
emissions), in this last case the indicators change
depending on type of Life cycle assessment methods
choosen.
A summary of the factors that can influence CHP
development in Latvia has been proposed in previous
papers. A. Volkova et al. [2] identify four main factors:
political, geographical-climatological, legislative and
technological.
In general the total amount of electricity produced in a
cogeneration regime and condensing mode depends
on constructive solutions (e.g. technical solution for the
biogas collectors), availability of source used (mainly
gas), constructive parameters of cogeneration plant,
parameters of heat energy consumers, heat load
duration curve, duration of heat energy consumption
levels, behaviour of energy end users, installed
capacity, energy efficiency of technologies,
development of demand side management factor, and
other factors.
3.2. Methodologies: EISD method and MOO method
In the following paragraph the algorithm of ISED core
set tool, included in the conceptual framework used by
United Nations Commission on sustainable
development (CED), is shown. After is also shortly
reported the MOO methodology
The EISD is an analytical tool developed which can
help energy decision and policymakers at all levels to
incorporate the concept of sustainable development
into energy policy. EISD core set is organized following
the conceptual framework used by United Nations
Commission on sustainable development (CSD).

Fig. 4. set of core EISD [8]

There are 30 indicators, classified into three
dimensions (social, economic and environmental) and
grouped in 7 big themes. There are four social
dimension indicators: three of them represent equity
(accessibility, affordability, disparities) and one health
theme (safety). The set of energy indicators of
economic dimension consists of 16 indicators. There
are nine environmental dimension indicators in the
EISD core list. The scheme of core EISD indicators is
presented in Fig. 4. The priority areas for energy sector
analysis in Latvia can be were selected based on the
main EU energy policy directions. These priority areas
are as follows:
- Energy use.
- Energy intensities.
- End-use intensities of economic branches.
- Energy security.
- Environmental energy impacts.
The next Fig. 5 shows the linkages among the
indicators selected for energy policy analysis in Baltic
States. Relevant policy actions based on analysis
September 5
th
to September 7
th

181

conducted in the previous sections are defined based
on targeted indicators.

Fig. 5. Linkages between indicators and relevant policy
actions based on the targeted indicators [8]

Multi-objective optimization (MOO), also known as
multi-criteria optimization, particularly outside
engineering, refers to finding values of decision
variables which correspond to and provide the optimum
of more than one objective. Unlike in single objective
optimization (SOO), which gives a unique solution,
there will be many optimal solutions for a multi-
objective problem. Multi-objective optimization involves
special methods for considering more than one
objective and analyzing the results obtained [7].
Often, the various objective functions conflict with each
other (i.e., optimizing one of them usually tends to
move another towards undesirable values), for solving
such models one needs to know how many units of one
function can be sacrificed to gain one unit of another,
but this trade-off information is not available. In other
words, one is forced to determine the best compromise
that can be achieved.
In the following paragraph an example of MOO applied
to the evaluation of possibilities to utilize landfill biogas
for electricity production in one of Latvias landfills.
4. TESTING OF LANDFILL GAS PRODUCTION
The improved biogas is one of the cleanest fuels with a
little impact on the environment and human health [11].
One of the advantages of biogas injection into natural
gas grid is the fact that natural gas grid connects a
place of biogas production (usually in rural areas) with
densely populated areas. It allows new consumers to
use gas. In this way it is possible to increase the biogas
production in remote areas not being worried about use
of produced biogas. The main barriers for improved
biogas injection are the high costs of improvement and
grid connection. Grid injection is limited by location of
biogas production and improvement sites, which have
to be close to natural gas grid [12].
Problems are connected with biogas utilization in
cogeneration plants (CHP) since there are no
possibilities to find heat energy consumers, which in
turn resulted with low efficiency landfill power plants
almost all over Latvia.
Due to high electricity feed-in tariff there is an
economical motivation for power plant operation with
low efficiency. For electricity produced in renewable
energy power plants with nominal capacity of up to
4MW high feed in tariff has been transposed in Latvias
legislative acts.
The development of Latvias landfill sites is at the
crossroads. On one hand it is economically feasible to
operate CHP just for electricity production, but on the
other it is important to use natural resources on full
value by producing from biogas the maximum amount
of heat energy. In first case it means that there is no
need for waste sorting in landfills, but in the other it is
important to sort both before waste collection and in
landfills.
Utilization of landfill biogas in Latvia is based on energy
production in power station placed close to landfill for
different reasons. One of the most important reasons is
financial state support of small scale power stations (4
MWe) from renewable energy resources. Such kind of
support prevents both, development of waste sorting
and utilization of refuse derived fuel in cement
production, and biogas improvement to cover needs
transportation sector or to connect to natural gas grid.
In the following is shortly reported the methodology
regarding the optimization model of biogas use in
landfills in Latvia in connection to the data collected
from landfill Daibe. After the analysis only two of the
independent parameters have been chosen: quality of
biogas (characterized by heat value), and technological
equipment (characterized by electrical capacity).
This optimization model for biogas utilization in landfills
includes four modules and is based on technological,
climate and economical sub models.
Results of economical optimization show that in case of
low biogas quality (4 kWh/m
3
) the optimal installed
capacity is 2.2MW. In case of biogas quality of 5
kWh/m
3
, optimal installed capacity is 2.8MW, and
3.4 MW in case of high biogas quality (6 kWh/m3).
September 5
th
to September 7
th

182

Fig. 6. Diagram of the economical optimization[12]

Results of technological optimization show that, the
higher the installed capacity, the shorter the operation
time of equipment. If assumed that operation time of
the equipment could be 5 up to 10 years, then the
installed capacity can be 0,5MW and higher.

Fig. 7. Equipment operation time vs installed capacity [12]

Results of the climate sub model show that the higher
is installed capacity, the greater the reduction of
greenhouse gas emissions. Besides that, it is not
possible to reach extremis by using two objective
functions (heat value of biogas and installed capacity),
which have been used in case of economical and
technological sub models, and it is necessary to
introduce another objective functions.

Fig. 8. Diagram of the environmental optimization [12]

Model of power production in landfill shows that feed-in
tariff stated as financial support today in Latvia allows
to reach economically feasible projects even in case if
cogeneration unit is operated in power station regime
(generates only electricity). Results show that state
policy needs corrections to improve energy efficiency of
biogas utilization for energy production.
5. COMMENT AND DISCUSSION
The use of CHP instead of conventional plant will
always improve energy efficiency and will reduce CO
2

emissions significantly, in Latvia there is potential to
replace some of the heat plants with co-generation
units (comparing with Denmark where approximately
80% of the district heating in Denmark is supplied from
CHP [6].
Hence promotion of high-efficiency cogeneration (CHP)
based on a useful heat demand is a priority with regard
to saving primary energy, avoiding network losses and
reducing emissions, in particular of greenhouse gases
[2].
Of course the choice of the fuel is fundamental in order
to reach the target required from the last EU directive in
terms Renewable Energy Sources (RES).
More use of energy from biomass in terms of woodfuel,
biogas, landfilled gas and biofuels seems to be a good
direction in order to displace the part of energy sources
given by the imported natural gas.
The use of wood in the energy sector (through the
production of heating and electricity) must become not
only an objective for the development of the energy
supply system, but it must also become part of
strategies for economic development and for the
improvement of the import/export balance of the
country. These measures can succeed in not only
developing local production and job creation, but if can
also stimulate and increase the potential export.
It is particularly important to conduct engineer-technical
and economic analysis of the various technological
solutions possible to implement wood use in the
cogeneration plants of the larger cities (including Riga
TEC 1 and TEC 2) [13]. Any possible choice and/or
scenario cannot be complete if it not references to a
Life cycle assessment (LCA) that it a good tool in order
to understand the environmental load of a certain
process strategy and in order to give a comparable
common base.
Based on the targeted indicators for Latvia the best
strategy can be identify in:
Enhance the diversity and variety of the energy
mix.
Improve maintenance of existing energy
infrastructure.
Eliminate constraints and investment in new
facilities.
Increase the efficiency of energy supply in
electricity generation.
Increase the share of electricity produced by
combined heat and power (CHP) plants.
Increase the share of renewable and domestic
energy sources in the energy mix.
September 5
th
to September 7
th

183

6. CONCLUSIONS
1. In the paper has been presented aspects and
problems of the Latvian energy-system connected to
the choice of the CHP and/or power stations for the
future national energy strategies in the light of the last
EU directive in the subjects of RES. The dependence
on imported energy sources, the growth of electricity
prices, and the need to support to local producers are
the main reasons for use of new renewable energy
technologies in the Latvian energy sector.
2. In this paper has been summarized the results from
the application of the Energy Indicators for Sustainable
Development (EISD), a good tool for analyzing trends,
setting energy policy goals and monitoring progress in
order to indentify good policy indicators. Also a testing
of landfill gas using MOO method has been reported
where only two of the independent parameters have
been chosen: quality of biogas (characterized by heat
value), and technological equipment (characterized by
electrical capacity). Model of power production in
landfill shows that feed-in tariff stated as financial
support today in Latvia allows to reach economically
feasible projects even in case if cogeneration unit is
operated in power station regime (generates only
electricity), but if is feasible from an economical point
view is not the same if reference to environmental
impact.
3. In the paper has been discussed how LCA can be a
good approach that enables the energy requirements,
GHG balance and other environmental impacts of
bioenergy production chains to be accounted and
accurately compared. Hence LCA is good tool in order
to give the possibility to compare different RES usage
strategies.
4. Due to high electricity feed in there is an economical
motivation for power plant operation with low efficiency.
For electricity produced in renewable energy power
plants with nominal capacity of up to 4MW high feed in
tariff has been transposed in Latvias legislative acts.
OF course this is not good from environmental point of
view.
5. The use of CHP instead of conventional plant will
always improve energy efficiency and will reduce CO2
emissions significantly, in Latvia there is potential to
replace some of the heat plants with co-generation
units.
6. Only crucial measures such as the reconstruction of
energy sources in the larger cities (including Riga
TEC 1 and Riga TEC 2) adjusting the use of fossil fuels
to biomass and conversion to non-natural gas sources,
will produce results. Biogas and landfill gas favorite the
environmental impact displacing usage of natural gas,
the possibility of the feasibility solution for connected
CHP in out-of-city region to heat consumer must be
evaluated.
7. REFERENCES
[1] Construction, Energy and Housing State Agency
Energy Department, Latvian energy in figures,
Riga, 2008.
[2] A. Volkova, E.Latev, A. Siirde, Small-scale CHP
potential in Latvia and Estonia, Scientific Journal of
RTU Environmental and climate technologies, Ser.
13, n. 2, Riga, 2009.
[3] Latvias district heating association , Heat supply in
Latvia,
http://www.lsua.lv/en/index.php?option=com_conte
nt&task=view&id=4&Itemid=5.
[4] D. Streimikiene, I. Roos, J. Rekis, External cost of
electricity generation in Baltic States, Renewable
and Sustainable Energy Reviews n. 13, 2009, pp.
863870.
[5] D. Streimikiene, I. Roos, J. Rekis, External cost of
electricity generation in Baltic States, Renewable
and Sustainable Energy Reviews n. 13, 2009, pp.
863870.
[6] L.H. Rasmussen, A sustainable energy-system in
Latvia, Applied Energy n. 76, 2003, pp. 18.
[7] G.P. Rangaiah, Multi-Objective Optimization:
Techniques and Applications in Chemical
Engineering, World Scientific, 2008, p. 454.
[8] D. Streimikiene, R. Ciegis, D. Grundey, Energy
indicators for sustainable development in Baltic
States, Renewable and Sustainable Energy
Reviews, 2007, Vol. 11, pp. 877893.
[9] G. Rebitzera et al., Life cycle assessment - Part 1:
Framework, goal and scope definition, inventory
analysis, and applications, Environment
International n. 30, 2004, pp. 701 720.
[10] D.W. Pennington et al., Life cycle assessment
Part 2: Current impact assessment practice,
Environment International n. 30, 2004, pp. 721
739.
[11] D.Blumberga, . Kuplais, I. Veidenbergs, E.Dace,
The benchmarking method for an evaluation of
biogas improvement methods, Scientific Journal of
RTU Environmental and climate technologies, Ser.
13, n. 2, Riga, 2009.
[12] G. Kuplais, D. Blumberga, E. Dace, F. Romagnoli,
Optimisation model of biogas use in landfills in
Latvia, 7th International conference ORBIT2010:
Organic resources in the carbon economy, June
29-July 3, 2010, Heraklion, Greece.
[13] A. Blumberga et al., Assessment on the use of
renewable energy resources in Latvia until 2020:
report, LVAF, December 2008, Riga.
September 5
th
to September 7
th

184

LCA OF COMBINED HEAT AND POWER PRODUCTION AT HELLISHEII
GEOTHERMAL POWER PLANT WITH FOCUS ON PRIMARY ENERGY EFFICIENCY
Marta Ros Karlsdottir, Olafur Petur Palsson, Halldor Palsson
University of Iceland, Faculty of Industrial Engineering, Mechanical Engineering and Computer Science
mrk1@hi.is

ABSTRACT
The aim of the study is to calculate primary energy
factors, f
p
, stating the primary energy efficiency as well
as factors for CO
2
emission, K, for geothermal
combined heat and power production at the Hellisheidi
CHP plant in South-West Iceland. These factors state
how much primary energy consumption and CO
2

emissions result from the production of 1 MWh of heat
and electricity due to geothermal utilization. Methods of
life cycle assessment (LCA) are used to calculate these
factors by taking into account all energy and material
streams to and from the CHP plant during construction
and operation. The results show that producing heat
and electricity in a combined heat and power plant
minimizes the primary energy factor for the electricity
generation and produces a relatively low primary
energy factor and CO2 production factor for the heat
generation process. From the results, it can also be
seen that life cycle assessment is a useful method to
evaluate the total impacts of the geothermal energy
conversion process, especially for the emission of
greenhouse gasses during the lifetime of the
production facilities. The experience in this study also
demonstrates that the method can equally be used for
processes as it is commonly used for the analysis of
total impact of products.
INTRODUCTION
The calculation of primary energy and CO
2
production
factors for geothermal power production has had little
attention while factors for some other types of energy
technologies such as hydropower, nuclear and coal
fired power plants have been developed during the
recent years. The importance of these factors is stated
mainly in the new recast of Directive 2002/91/EC of the
European Parliament and of the Council on the energy
performance of buildings [1]. There it is stated that
before the end of year 2010, all new building occupied
by public authority should be issued energy
performance certificates showing these factors, based
on the energy mix used by the building and the
buildings energy performance.
At present time, geothermal power plants are situated
in 24 countries [2] and a total of 78 countries have
reported direct use of geothermal energy [3]. With
increasing fossil fuel prices and focus on renewable
energy sources, these power plants producing green
energy become more viable in various locations
around the world. It is thus important to investigate
their primary energy efficiency and environmental
impact for comparison with other energy conversion
technologies. These energy performance indicators
can be used to help decision making of future
developments, policy making and energy rating of
buildings.
Countries that have access to geothermal areas and
produce power by geothermal utilization within the
European Union (EU) are: Austria, France, Germany,
Greece, Hungary, Italy, Netherlands, Portugal,
Romania, Slovakia and Spain. Other European
countries such as Iceland and Turkey, which are not
current member states of the EU, also utilize
geothermal energy extensively [2]. Also, 32 European
countries use geothermal energy directly for various
purposes such as district heating [3]. Thus, electricity
and heat based on geothermal energy are a part of
Europes energy mix. For countries using geothermal
based power and/or heat and complying to EU
legislation, it is therefore important to have easy access
to standardized factors accounting for the primary
energy efficiency and CO
2
emissions from geothermal
based heat and power.
The aim of this study is to produce standardized factors
for primary energy efficiency (f
p
) and CO
2
emission (K)
for geothermal heat and power production.
ENERGY PERFORMANCE INDICATORS FOR
PRIMARY ENERGY CONSUMPTION AND CO
2

EMISSIONS
The primary energy factor is defined as the ratio
between the total primary energy inputs involving
energy production to the actual energy delivered to the
consumer. According to [4], it should always account
for the extraction of the energy carrier and its transport
to the utilization site, as well as for processing, storage,
generation, transmission, distribution and delivery.
There are two primary energy factors defined:
- Total primary energy factor, accounting for
primary energy use of both renewable energy
sources and non-renewable sources.
- Non-renewable primary energy factor,
accounting only for the primary energy
consumption of non-renewable energy sources.
This factor is used when expressing only the use of
September 5
th
to September 7
th

185

fossil or other non-renewable or polluting energy
sources in the energy conversion system.
The CO
2
production coefficient, K, shall include all CO
2
-
emissions associated with the primary energy used.
Furthermore, equivalent emissions of other greenhouse
gases, e.g., methane, may be included [4].
According to Directive 2002/91/EC, indicators on the
energy performance of buildings shall include the
consumption of primary energy and the CO
2
emissions
resulting from the buildings energy usage. Factors for
primary energy consumption and CO
2
emissions have
been calculated for various energy chains producing
electricity, and values for these factors are given in
Annex E of the standard EN15603 on the energy
performance of buildings. An overview of these factors
is given in Table 1.
Table 1: Energy performance indicators for various
sources of electricity [4]
Source of
electricity
Primary energy factors fp
[MWh primary energy /
MWh delivered energy]
CO2
production
coeff. K
[Kg/MWh]
Non-
Renewable
Total
Hydraulic
power
0.50 1.10 7
Nuclear
power
2.80 2.80 16
Coal power 4.05 4.05 1340
Electricity
mix
UCPTE
3.14 3.31 617

As seen in the standard EN15603:2008 [4] and
Table 1, no indicators are given for geothermal power.
The directive is under reconstruction and a recast has
been released, as mentioned before. Also, the table
does not give factors for sources of thermal energy
used by buildings for space heating. Thus, there is
clearly a need to calculate these factors for energy
chains that involve geothermal energy, since they
produce both electricity and heat which is delivered to
buildings within the European Union and in countries
following EU legislation.
GEOTHERMAL HEAT AND POWER PRODUCTION
AT HELLISHEIDI CHP PLANT
Hellisheidi geothermal CHP plant is situated at the
Hengill geothermal area close to Reykjavik, the capital
of Iceland. A 90 MW electricity production started in
2006 after several years of construction and research.
In 2007, a low pressure turbine was added, increasing
the power generation to 120 MW. A year later, another
90 MW were added, resulting in a power generation
capacity of about 210 MW (213 MW in February 2009).
Further developments of the power plant include
adding heat production in 2010 for district heating and
also increasing the power production if possible.
Estimated production capacity for the completed
Hellisheidi Plant is 300 MW electricity and 400 MW
thermal energy [5].
The plant today is a double flash power plant with high-
and low-pressure turbines and separators as seen in
Figure 1. The heat production facilities are currently
under construction with a planned 133 MW thermal
capacity at the end of year 2010. The technical
complexity is moderate and the plant makes a good
basis for a LCA study to evaluate the primary energy
efficiency and CO
2
emission of this type of geothermal
power plant. Since it is fairly newly constructed, access
to detailed background data for the inventory modelling
is possible, making the study more reliable and
accurate. Environmental assessment for the
production is available as well as measurements of
various environmental impacts of the power plant,
providing data for the impact assessment of the LCA
study.
In this study, a steady production of 213,6 MW
electricity and 121 MW heat is used as a basis for the
LCA model. The reason for this choice is that the
newest inventory data on the construction phase and
mass extraction are built on these production
capacities, and that the base thermal load is estimated
to be 121 MW and not the full capacity of 133 MW.
PRIMARY ENERGY OF VARIOUS ENERGY
SOURCES
There is a matter of inconsistency in primary energy
calculations of various energy sources as many
different methods are in use and accepted by different
energy authorities [6]. As an example, the primary
energy factors for power produced from renewable
energy sources such as hydro power, wind energy and
solar energy are sometimes calculated by assuming
that the primary energy factor for the energy
conversion system is one, which is the same as
assuming that the energy conversion process is 100%
efficient. The reason for this assumption is that the
primary energy is defined as the first usable stage of
the energy flow, which in the case of wind, solar and
hydro is the electricity itself produced from these
primary sources [7]. For electricity production from
heat sources, the first usable stage of the energy
stream is defined as the steam input into the turbine,
according to an energy statistics manual from the
International Energy Agency (IEA) [8]. The methods
September 5
th
to September 7
th

186

HPT
LPT
HPS
LPS
LPC
TV
CT
G
G
W
CP
HPC
Cold water
tank
Hot water
tank
HX1 HX2
IW

Figure 1 A simple schematic of the Hellisheidi geothermal CHP plant

used to calculate primary energy demand of power
production from renewable energy sources tends to
underestimate the primary energy input from the
original energy sources into the energy conversion
system compared to the assumptions made for the
heat conversion processes such as coal, oil and also
geothermal.
PRIMARY ENERGY OF VARIOUS ENERGY
SOURCES
There is a matter of inconsistency in primary energy
calculations of various energy sources as many
different methods are in use and accepted by different
energy authorities [6]. As an example, the primary
energy factors for power produced from renewable
energy sources such as hydro power, wind energy and
solar energy are sometimes calculated by assuming
that the primary energy factor for the energy
conversion system is one, which is the same as
assuming that the energy conversion process is 100%
efficient. The reason for this assumption is that the
primary energy is defined as the first usable stage of
the energy flow, which in the case of wind, solar and
hydro is the electricity itself produced from these
primary sources [7]. For electricity production from
heat sources, the first usable stage of the energy
stream is defined as the steam input into the turbine,
according to an energy statistics manual from the
International Energy Agency (IEA) [8]. The methods
used to calculate primary energy demand of power
production from renewable energy sources tends to
underestimate the primary energy input from the
original energy sources into the energy conversion
system compared to the assumptions made for the
heat conversion processes such as coal, oil and also
geothermal.
Definition of Primary Energy of Geothermal Fluid
There is no clear definition of primary energy from
geothermal energy sources. Published methods of
determining the primary energy consumption in
geothermal power plants are the following [6]:
- Working Group III (WG III) of the
Intergovernmental Panel on Climate Change
(IPPC) records electricity from geothermal on a 1:1
basis. This results in a f
p
factor of 1.
- The Engineering Information Administration (EIA)
uses a factor of 6.16 units of primary geothermal
energy for each unit of geothermal electricity.
- International Energy Agency (IEA) records a f
p

value of 10 by assuming 10% conversion efficiency
of geothermal power plants.
In this LCA study, where the main goal is to calculate
an accurate f
p
factor for a specific conversion
technology, the main issue is the primary energy
content of the geothermal fluid extracted from the
production wells. The primary energy content of the
geothermal fluid can be based on different
assumptions. The first one is the energy content of the
geothermal fluid based on its enthalpy in kJ/kg.
Second, the exergy content of the fluid can be used as
a basis. However, in this study, the primary energy
content of the geothermal fluid taken from the
production wells (and utilized for both electricity and
heat production) is chosen to be the enthalpy above
W: Geothermal production well
G: Generator
HPS: High pressure steam separator
HPT: High pressure steam turbine
HPC: Condenser for high pressure
turbine
LPS: Low pressure steam separator
LPT: Low pressure steam turbine
LPC: Condenser for low pressure
turbine
CT: Cooling tower
CP: Cooling water pump
HX1: Heat exchanger 1 for DH system
HX2: Heat exchanger 2 for DH system
IW: Reinjection well

September 5
th
to September 7
th

187

15 C, an International Standard Atmosphere (ISA)
reference temperature [9], and calculated in the
following manner:

(1)

Where is the specific primary energy content in
kJ/kg, is the enthalpy of the fluid and is
the saturated liquid enthalpy of the fluid at standard
reference temperature of 15 C.
LIFE CYCLE ASSESSMENT
The Directive 2002/91/EC defines the concept of
primary energy as energy that has not undergone any
energy conversion process [1]. The primary energy
factor must thus represent all the primary energy
consumed in order to provide one unit of heat or power
to the consumer. Primary energy consumption of
energy chains is not only based on the consumption of
fuel (or other energy source) in the power or heat
generation process, but also all the primary energy
needed for the construction, operation and possibly
demolition of the production facilities. Also, some
primary energy is needed for the distribution of the
product. To calculate such accumulated primary
energy, the method of life cycle assessment is well
suited. LCA is a method that has been developing
since the earliest performance of such a study in 1969
and standards on the methodology where issued in the
late 1990s [10].
LCA has been considered a good tool to achieve a
holistic approach on evaluating the environmental
impact of products. Today, it is widely used to
investigate all kinds of production systems and has
given valuable insight on the total impact of products
and systems on the environment by not only focusing
on the operational aspect [11]. Many interesting results
have been achieved by using this methodology and
those results form a basis for evaluating and comparing
different solutions for production of various products,
such as vehicles for transport, soft drink containers and
power conversion technologies. On the other hand,
LCA in the process industry has had much less
attention than for manufacturing products, and
research is needed before complete methods for
processes are readily available [11]. The application of
LCA on geothermal energy utilization can be valuable
for LCA developers working on further improvements
and adjustments on the LCA methodology for the
process industry.
Using LCA to calculate the total primary energy
consumption and CO
2
emission for heat and power
production based on geothermal energy will help
identify how much effect the construction, collection of
geothermal fluid and even the demolition phase of the
power plant and the distribution system have on the
total primary energy consumption. It can identify the
impact of the drilling of wells, manufacturing of power
plant components and piping, construction of buildings
and roads associated with the power plant, operation of
the power plant itself and the primary energy extracted
from the geothermal reservoir and even the impacts of
constructing and operating the distribution facilities.
The different phases of performing LCA will be
described in the following sections. The main phases of
LCA include:
- Defining the goal and scope of the study
- Performing inventory analysis
- Performing impact assessment
Goal and Scope of the Study
The main goal of this LCA study is to analyze the two
energy performance indicators presenting the primary
energy efficiency and the CO
2
emissions for both the
electricity and heat production at Hellisheidi power
plant. The LCA calculations and impact assessment
where done by using the LCA software SimaPro 7 [12]
and using different databases such as the Ecoinvent
database [13] for the inventory information on various
raw materials and processes used in the geothermal
power plant.
There are numerous geothermal power plants
worldwide using similar technology as the Hellisheidi
power plant to produce electricity (double flash power
plants produced 23% of the electrical power from
geothermal resources in 2007 [14]), so the results for
the energy performance indicators for the power
production at Hellisheidi could be used to represent
these power plants. Other types of geothermal energy
conversion systems, such as single flash and binary
systems, should be treated individually when
calculating energy performance indicators for the
electricity production.
Geothermal combined heat and power plants are not
common worldwide, but regarding Europe they can be
found in Iceland as well as Austria and Germany. By
producing heat as well as electricity in geothermal
applications, the utilization of the heat taken from the
geothermal reservoir in the form of geothermal fluid is
maximized. The heat produced has a variety of useful
applications, such as for district heating, agriculture,
fisheries, swimming pools, snow melting and heating
up greenhouses [3]. The calculations of the primary
energy factor of the heat production at Hellisheidi
geothermal CHP plant will emphasize this increase in
thermal efficiency of the power plant.
September 5
th
to September 7
th

188

The scope of this LCA study includes making the
following choices [11]:
- Functional unit
- System boundaries
- Choice of impact categories
- Method for impact assessment
- Principles for allocation
- Data quality requirements

Electricity,
geothermal, at
Hellisheidi CHP
plant
Geothermal
power plant unit
Geothermal
fluid, at power
plant
Geothermal
fluid, in ground
Drilling of
geothermal
wells
Collection
pipelines
Power plant
structures
Power plant
equipment
1 kWh electricity
Geothermal
heat production
unit
Heat,
geothermal, at
Hellisheidi CHP
plant
1 kWh heat
Heating station
equipment
Heating station
Structure
Geothermal
fluid, from
electricity
production
Heat, from
condenser

Figure 2: Flow model for the life cycle assessment of the Hellisheidi CHP plant

Functional Unit
The primary energy and CO
2
factors are defined as
primary energy usage and CO
2
emission per MWh and
thus, the functional unit of the study is chosen to be
MWh of electricity or heat produced in the Hellisheidi
geothermal CHP plant. The functional unit is the
reference flow to which all other modelled flows of the
system are related.
System Boundaries
The processes included in this LCA study are mainly
the operation and construction of the power plant. The
demolition or end-of-life phase is disregarded due to
insufficient information at this time. Also, the energy
and material flows due to maintenance in the
operational phase of the power plant are disregarded
but both the demolition and the maintenance will be
included in further studies. The time horizon in this
study is chosen to be 30 years, which is the technical
lifetime of the power plant capital goods.
A flow model of the CHP plant as modelled in the LCA
study is shown in Figure 2. The two outputs of the
production system are 1 MWh of electricity and 1 MWh
of heat. The main material and energy inputs into the
energy conversion system are the geothermal power
plant unit, the geothermal heat production unit and the
geothermal fluid. The geothermal power plant is
constructed from the power plant structures and
equipment while the fluid is transported in collection
pipelines from geothermal wells that need to be drilled
for the production. The heat production unit consists of
the heating station structure and equipment. The
energy input into the heating process is waste heat
from the power production process in form of heat
taken from the steam in the condenser for preheating
of district heating water, and the waste geothermal fluid
from steam separators used for final heating of the
district heating water. Inventory data on all these
different components in the flow model was collected
and used for the LCA study of the Hellisheidi CHP
plant.
Impact Categories and Methods for Impact
Assessment
To calculate the two energy performance indicators,
the two main impact categories to be used are the
primary energy demand of the production process in
MWh and Global Warming Potential (GWP) given in
CO
2
equivalents.
September 5
th
to September 7
th

189

Two different methods of impact assessment had to be
used in the impact assessment calculations. For the
primary energy factor, the Cumulative Energy Demand
(CED) method [15] was used which is based on a
method published by Ecoinvent 1.01 and available in
SimaPro 7 impact assessment methods. For the
calculation of the CO
2
emission factor, the IPCC 2007
GWP 100a V1.01 [16] was used to get the CO
2

equivalent total global warming potential for the chosen
functional unit of 1 MWh electricity produced.
Principles for Allocation
To allocate the impacts of the different products,
electricity and heat, produced at Hellisheidi CHP plant,
several methods can be used. The method used
should reflect the physical relation between the two
products, such as how the different inputs and outputs
of the process are dependent on the two different
products. Simple methods of allocation for an energy
conversion process can be:
- Based on energy content of the products
- Based on exergy content of the products
- Based on the monetary value of the products
The abovementioned methods can be used when the
physical relation between the two products is unclear.
In the case of the Hellisheidi CHP plant, the physical
relation between the two outputs (electricity and heat)
is mainly the use of waste heat from condensers and
the geothermal fluid from the production wells, as
shown in Figure 2. The impacts of construction can
easily be divided between the electricity and heat
production with the detail of inventory data provided.
Also, the geothermal fluid used in the heat production
is taken from steam separators in the electricity
generation process and would otherwise be reinjected
back into the geothermal reservoir via reinjection wells.
The disposed heat in the condenser is utilized to
preheat the district heating water by using it as cooling
water. The condenser pressure determines the
temperature of the steam output from the turbines and
thus, also the final temperature of preheating of the
district heating water. If the heat demand is high, the
condenser pressure must be higher than the optimum
for power production in order to supply high enough
temperatures to the district heating water. This limits
the electrical power production and requires that more
geothermal wells have to be drilled in order to sustain
the electrical production under high thermal loads of
the district heating system. These limitations on the
electrical production imply that the allocation of impacts
from the drilling of wells should be related to the
number of wells that have to be drilled to sustain both
the electricity production and the highest thermal load
designed for the district heating system.
Data Quality
To calculate the energy performance indicators by
methods of LCA, reliable inventory information is
needed on material and energy flows to and from the
geothermal power production facilities during their
lifetime.. The inventory in this study is constructed
from data provided by Reykjavik Energy, the power
company in ownership of the Hellisheidi plant. The
data on the construction phase is retrieved from the
conditions and specifications in a tender for the
construction of the power plant, where quantitative
information is collected on all major material flows
required for the constructions and machinery. The
inventory information for the fluid collection and drilling
is retrieved from a report done by Reykjavik Energy,
including the power and performance of the geothermal
wells drilled for the power and heat production [17].
For a LCA study, the following data quality indicators
must be presented:
- Time period
- Region
- Type of technology and representativeness
- Allocation
- System boundaries
In this study, the time period of the data is from 2005 to
2009 and the region is Western Europe. The type of
technology is modern and the representativeness is
data from a specific company. The allocation, as
mentioned before, is by physical connections between
the two outputs. The system boundaries are described
by three different criteria. First, the cut-off criteria is in
general set to be less than 5% which means that all
inventory data that does not contribute more than 5%
to the overall impacts of the two products is
disregarded. Also, the system boundary is chosen to
be of the first order, only to account for the materials
used in the construction and operation of the CHP plant
but not the processing and transportation of these
materials. The third system boundary criterion is the
system boundary with nature, which in this study is
described as unspecified at this stage of the LCA
study.
RESULTS FOR THE ENERGY PERFORMANCE
INDICATORS
Energy Performance Indicators for Electricity
Production
The results for the impact assessment of the electricity
production alone, focusing on the two energy
performance indicators, is shown in Table 2. The
highest value of fp 6.33 MWh primary energy/MWh
produced energy, is obtained when no heat production
is present at the power plant and the effects of
reinjection of waste streams is not taken into account.
September 5
th
to September 7
th

190

The value of 5.33 for fp is obtained in the two latter
cases, where the waste heat is either reinjected back
into the reservoir or used for heating of DH water. In
those cases, the primary energy content of the waste
stream can be subtracted from the primary energy
content of the geothermal fluid used for the electricity
production, resulting in lower fp values. The share of
non-renewable primary energy sources such as oil and
gas used in the construction phase or in the
manufacturing of various power plant components, only
account for about 0.01 of the total fp value in all cases.
The factor K for the CO
2
emissions is the same for all
three cases of electricity production as reinjection and
utilization of waste stream does not have significant
effects on the total emissions due to the power for
production. The origins of the CO
2
emissions can be
seen in Figure 4. The largest contributor to the CO
2

emission from the electricity generation over 30 years
of production is the geothermal fluid, responsible 88%
of the CO
2
emissions per kWh of electricity production.

Table 2 Results for the primary energy factor and CO2 emission factor for electricity based on geothermal energy
Source of electricity
[MWh primary energy / MWh produced
energy]
CO2 production
coeff. K
[Kg/MWh]
Non-Renewable Total
Electricity from Hellisheidi geothermal power
plant
0.01 6.33 29
Electricity from Hellisheidi geothermal power
plant, with reinjection
0.01 5.33 29
Electricity from Hellisheidi CHP plant 0.01 5.33 29

A small share of 8% originates from the drilling of
geothermal wells while the construction of the power
plant, along with the manufacturing of its main
components, is responsible for 4% of the CO
2

emissions.

87.5%
4%
8%
0.5%
GWP 100a for Electricity Production
in kg CO2 eq
Geothermal fluid
(87.5%)
Power plant and
components (4%)
Geothermal well
drilling (8%)
Collection lines
(0.5%)

Figure 3 Origins of CO2 emissions from the different
processes of the power generation
Energy Performance Indicators for Thermal
Production
The energy performance indicators for the production
of heat for district heating are given in Table 3. Two
cases are presented for the heat production; heat
production process with or without the effects of
reinjection of waste geothermal fluid. The highest value
for f
p
is obtained in the case where reinjection is not
taken into account, with the value of 1.78 MWh primary
energy/MWh produced energy. With reinjection, the
value reduces to 0.69. In both cases, the share of
primary energy from non-renewable energy sources is
less than 0.01. In both cases, the CO
2
production
coefficient is 0.98 kg CO
2
equivalents per produced
MWh.
The origins of the CO
2
emission from the heat
generation process can be seen in Figure 4. The
largest contributor to the total emissions is the drilling
of the geothermal production wells that were needed to
sustain the electricity production while the heat
production is at maximum load of 133 MWth. The
manufacturing of the district heating pipeline from the
production area to the rural area of Reykjavk city
contributes to 15% of the total emission resulted from
the heat generation process.

Figure 4 Origins of CO2 emissions from the different
processes of the heat generation

September 5
th
to September 7
th

191

Table 3 Results for the primary energy factor and CO2
emission factor for heat from a geothermal CHP plant
Source of
heat
[MWh primary energy /
MWh produced energy]
CO2
production
coeff. K
[Kg/MWh]
Non-
Renewable
Total
Heat,
Hellisheidi
CHP plant
>0.01 1.78 0.98
Heat,
Hellisheidi
CHP plant,
reinjection
>0.01 0.69 0.98
DISCUSSION
The following discussion highlights the most significant
results from this study:
1) By comparing the energy performance indicators
calculated in this study and shown in Table 3 to the
indicators given in Table 3 it can be seen that electricity
from geothermal power plants has the highest total f
p

factor while the share of non-renewable energy
sources is the lowest. The main reason for the high f
p

factor is the low conversion efficiency of geothermal
power plants due to low working temperatures and
pressures. The CO
2
production coefficient is relatively
low compared to the other energy conversion
technologies and could be lowered even further if
measures are taken to control the emissions from the
power plant. The results for the Hellisheidi geothermal
CHP plant cannot be used to represent all geothermal
power plants producing either electricity alone or with a
combined production of electricity and heat. Further
studies are needed on different types of geothermal
power plants, such as single flash and organic Rankine
cycles, to be able to produce specific or average
factors representing geothermal utilization.
2) The results for the heat production at the
Hellisheidi geothermal CHP plant, given in Table 3,
show that the energy performance indicators are
relatively low and, in the case of reinjection, below
unity. This is because the primary energy needed to
preheat the DH water is not accounted for in the heat
production but rather assigned to the electricity
production. This is due to the fact that the preheating of
the DH water from 5 C to 41 C is done in the
condenser for the high pressure steam turbine as seen
in Figure 1 and is a necessary step in the electricity
production, but a beneficial step in the heat production
for the DH system.
3) Values for the indicators for both electricity and
heat are calculated with and without reinjection of the
cooled geothermal brine from the energy conversion
process. Reinjection of geothermal brine is recognized
to improve heat mining and stabilize the production
capacity of geothermal fields, if successfully carried
out. It can also counteract pressure draw-down in the
reservoir by providing an artificial water recharge [18].
In this study, reinjection of the waste stream is
modelled, which decreases the use of primary energy
in the energy conversion process, since a part of the
primary energy from the geothermal fluid it is returned
back to the reservoir. Reinjection is present at the
Hellisheidi geothermal CHP plant so the values of the
energy performance indicators with reinjection are valid
for the power plant.
4) Life cycle assessment is especially useful to
evaluate the total impact of geothermal power plants
with respect to their emission of greenhouse gasses.
Figure 3 and Figure 4 show how the different phases in
the life cycle of the power plant significantly contribute
to the overall emission in CO
2
equivalents. If LCA had
not been carried out for the process, 12% of the CO
2

emissions resulting from the electricity generation
would not have been accounted for and no emissions
would have been found for the heat production, since
the emissions from drilling, construction of buildings,
and manufacture of components had not been
accounted for.
ACKNOWLEDGMENTS
Special thanks are given to the following partners:
Nordic Energy Research (NER) for funding the study
and the Energy Research Fund of Landsvirkjun for their
support. To Orkuveita Reykjavkur for providing data for
Hellisheidi Power plant, to Mannvit engineering for
discussion and data provision and to Ragnar Gylfason
for his contribution in the data gathering phase.
REFERENCES
[1] EU. (2003, January 4). 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
Communities .
[2] Bertani, R. (2010). Geothermal Power Generation
in the World 2005 2010 Update Report.
Proceedings World Geothermal Congress 2010,
(April), 25-29.
[3] Lund, J. W., Freeston, D. H., & Boyd, T. L. (2010).
Direct Utilization of Geothermal Energy 2010
Worldwide Review. Proceedings World Geothermal
Congress 2010, (April), 25-29.
[4] EN 15603:2008. Energy performance of buildings.
Overall energy use and definition of energy ratings.
September 5
th
to September 7
th

192

Geneva: International Organisation for
Standardisation (ISO).
[5] Helisheidi Geothermal Plant. (2009). Retrieved
May 2009, from http://www.or.is/English/Projects/
HellisheidiGeothermalPlant/
[6] H. Douglas Lightfoot. (2007). Understand the three
different scales for measuring primary energy and
avoid errors. Energy, 32, 1478-1483.
[7] Segers, R. (2008). Three options to calculate the
percentage renewable energy: An example for a
EU policy debate. Energy Policy , 36 (9), 3243-
3248.
[8] IEA. (2004). Energy Statistics Manual.
International Energy Agency (IEA). Paris:
OECD/IEA.
[9] ISO 2533:1975. Standard atmosphere.
International Organization for Standardization,
Geneva, Switzerland.
[10] Russell, A., Ekvall, T., & Baumann, H. (2005). Life
cycle assessment - introduction and overview.
Journal of Cleaner Production , 13 (13-14), 1207-
1210.
[11] Baumann, H., & Tillman, A.-M. (2004). The Hitch
Hiker's Guide to LCA. Lund, Sweden:
Studentlitteratur AB.
[12] PRConsultants. (2009, September 6). SimaPro
LCA software. Retrieved October 14, 2009, from
SimaPro LCA software: http://www.ecoinvent.ch/
[13] Ecoinvent. (2009, August 13). Home. Retrieved
October 14, 2009, from Home:
http://www.ecoinvent.ch/
[14] DiPippo, R. (2008). Geothermal Power Plants
Principles, Applications, Case Studies and
Environmental Impact (2nd edition ed.). Oxford:
Butterworth-Heinemann.
[15] Klpffer, W. (1997). In defense of the cumulative
energy demand (editorial). International Journal of
Life Cycle Assessment , 2, 61.
[16] PRConsultants. (2009, September 6). Methods.
Retrieved October 14, 2009, from SimaPro LCA
software:
http://www.pre.nl/simapro/impact_assessment_met
hods.htm#CML2
[17] Gunnlaugsson, E., & Oddsdttir, A. L. (2009).
Helisheidi - Gufuborholur 2008 (Hellisheidi - Steam
wells 2008). Reykjavk: Orkuveita Reykjavkur.
[18] Stefansson, V. Geothermal reinjection experience.
Geothermics, 26, (1997), 99130.

September 5
th
to September 7
th

193

FLEXIBILITY FROM DISTRICT HEATING
TO DECREASE WIND POWER INTEGRATION COSTS
J. Kiviluoma
1
and P. Meibom
2

1
VTT Technical Research Centre of Finland
2
Ris DTU
ABSTRACT
Variable power sources (e.g. wind, photovoltaics)
increase the value of flexibility in the power system.
This paper investigates the benefits of combining
electric heat boilers, heat pumps, CHP plants and heat
storages in a district heating network when the share of
variable power increases considerably. The results are
based on scenarios made with a generation planning
model Balmorel [1]. Balmorel optimises investments
and operation of heat and power plants, including heat
storages. It uses hourly resolution and enforces
temporal continuity in the use of the heat storages.
Scenarios with high amount of wind power were
investigated and the paper describes how the increase
in variability changes the profitability and operation of
different district heating options in more detail than was
described in the article by Kiviluoma and Meibom [2].
Results show that district heating systems could offer
significant and cost-effective flexibility to facilitate the
integration of variable power. Furthermore, the
combination of different technologies offers the largest
advantage. The results imply that, if the share of
variable power becomes large, heat storages should
become an important part of district heating networks.
NOMENCLATURE
INTRODUCTION
Wind power is projected to be a large contributor to
fulfil electricity demand in several countries. This could
take place due to relatively low cost of wind power
electricity or policy mechanisms promoting renewable
energy. In any case, power systems with a large
fraction of power coming from a variable power source
will need to be flexible. Flexibility is used to cope with
the increased variation in residual load (electricity
demand minus variable power production) and with the
increased forecast uncertainty in the residual load. On
the other hand, lack of flexibility will cause larger costs
from increased variability and forecast errors.
Therefore, it is prudent to investigate the cost optimal
configurations for the combined power and heat
generation portfolios.
Heat generation could offer significant possibilities for
increasing the flexibility of the power system. Currently,
part of the inflexibility of the power system comes from
CHP plants that are operated to serve the heat load
while electricity is a side product. Installation of electric
resistance heaters next to the CHP units or elsewhere
in the heat network could break this forced connection.
During periods of low power prices, which will become
more common with high share of wind power, CHP
plants could be shut down and heat would be produced
with electricity. The dynamics can be made more
economic with the use of heat storages. Further option
is to have heat pumps in the DH network, but they will
require large amount of full load hours to be profitable
and will compete with CHP plants for the operating
space.
In most countries heat demand is in the same order of
magnitude as electricity demand. For example, in UK
the demand for primary energy due to heat is around
40% of total primary energy demand [3]. About 25% of
the primary energy demand is due to space and non-
industrial water heating. In the US all kind of heat use
accounts for about 30% of the primary energy
consumption [estimated from 4].
Heat is inexpensive to store compared to electricity.
Electricity storage has been seriously considered to
alleviate the variability of wind power [5-6]. Therefore, it
is apparent that the use of heat storages should also
receive serious consideration in the current context.
Some work has been done [7-9], but not considering
Indices
i, I Unit, set of units
I
HeatSto
Heat storage units
t, T Time steps, set of time steps
a, A Area, set of areas
Variables
C New capacity
P Power generation
Q Heat generation
Z Charging of heat storage
Parameters
c
Inv
Annualized investment cost
c
Fix
Fixed operation and maintenance costs
c
Operation
Operation cost function of the unit
w Weight of time period
h Heat demand
September 5
th
to September 7
th

194

optimal investments in new power plants and heat
storages.
The study has been restricted to residential and
industrial district heating systems. Buildings not
connected to district heating systems were not
considered, although these also require heat. Cooling
demand could also offer similar possibilities, but the
problem was not addressed here. Industrial heat
demand and water heating do not usually have strong
seasonal variation and can therefore be more valuable
towards the integration of variable power.
METHODS AND DATA
The model and assumptions used for the analysis are
described in more detail in [2]. For convenience, most
important sections are referenced below. The heat
sector of the model is described more thoroughly here.
The Balmorel model is a linear optimization model of a
power system including district heating systems. It
calculates investments in storage, production and
transmission capacity and the operation of the units in
the system while satisfying the demand for power and
district heating in every time period. Investments and
operation will be optimal under the input data
assumptions covering e.g. fuel prices, CO2 emission
permit prices, electricity and district heating demand,
technology costs and technical characteristics (eq. 1).
The model was developed by (Ravn et al. [1]) and has
been extended in several projects, e.g. (Jensen &
Meibom [10], Karlsson & Meibom [11], Kiviluoma &
Meibom [2]).
( ) ( )|
.
|
\
|
+ + +

e e e e T t I i
t i t i
Operation
i t
I i
i
Ex
i
Fix
i
I i
i
Inv
i
Q P c w C C c C c
, ,
, min
(1)
The optimization period in the model is one year
divided into time periods. This work uses 26 selected
weeks, each divided into 168 hours. The yearly
optimization period implies that an investment is carried
out if it reduces system costs including the annualized
investment cost of the unit.
The geographical resolution is countries divided into
regions that are in turn subdivided into areas. Each
country is divided into several regions to represent its
main transmission grid constraints. Each region has
time series of electricity demand and wind power
production. The transmission grid within a region is
only represented as an average transmission and
distribution loss. Areas are used to represent district
heating grids, with each area having a time series of
heat demand. There is no exchange of heat between
areas. In this article, Finland is used as the source for
most of the input data.
The hourly heat demand has to be fulfilled with the heat
generation units, including heat storages (eq. 2).
Loading of heat storage adds to the heat demand. Loss
during the heat storage process is not considered. The
dynamics of heat networks were not taken into
account.

e e
+ =
I i I i
t i t r t i
HeatSto
a
Z h Q
, , ,

A a T t e e ;
(2)
Analysis is done for the year 2035. By this time, large
portion of the existing power plants are retired. Three
district heating areas were considered. These have a
rather different existing heat generation portfolio by
2035. This helps to uncover some interesting dynamics
in the results section.
In this paper, scenarios without new nuclear power are
compared (scenarios Base NoNuc and OnlyHeat
NoNuc in article [2]). This meant that wind power had a
very high share of electricity production. Accordingly,
there was more demand for flexibility in the system.
Urban area presents the heat demand in the capital
region of Finland. The existing power plants in 2035
cover over half of the required heat capacity. Largest
share comes from natural gas, which is a relatively
expensive fuel in these model runs. The annual heat
demand is smallest of the considered areas: 6.2 TWh.
Industry area aggregates the known industrial district
heating demand from several different locations. This is
a necessary simplification, since Finland has over
hundred separate DH areas and the model would not
be able to optimise all of these simultaneously. The
industrial heat demand in Finland is driven by paper
and pulp industry, which produces waste that can be
used as energy input. This capacity is assumed to be
available in 2035 and as a consequence the model
does not need more industrial heat capacity. The
annual heat demand is 46.8 TWh.
Rural area aggregates non-industrial heat demand
excluding the capital region considered in Urban. This
is probably the most interesting example, as the
existing capacity covers only 20% of the heat capacity
demand. Therefore, the model has to optimise almost
the whole heat generation portfolio. There are wood
resources (limited amount of forest residues and more
expensive solid wood) available unlike in the urban
area. The annual heat demand is 21.0 TWh.
RESULTS
Figures 13 give an example how heat production
meets heat demand in the different areas during the
same 4.5 days in January. Negative production
indicates charging of heat storage. Electricity price is
on separate axis together with the cumulative content
of heat storage. When electricity price is low, storage is
loaded with electricity using heat boilers and heat
pumps. When electricity price is high, CHP units
September 5
th
to September 7
th

195

produce heat and electricity. Fluctuations in electricity
price are mainly driven by changes in wind power
production, since these are larger than changes in
electricity demand (Fig. 4).
Fig. 1. Example of operation in Urban heat area. Negative production indicates charging of heat storage.

-2000
-1000
0
1000
2000
3000
4000
5000
6000
H
e
a
t

p
r
o
d
u
c
t
i
o
n

(
M
W
)
0
75
150
225
300
375
450
E
l
e
c
t
r
i
c
i
t
y

p
r
i
c
e

(
/
M
W
h
)
H
e
a
t

s
t
o
r
a
g
e

c
o
n
t
e
n
t

(
%
)
PE_BP_IN
WR_BP_IN
WW_BP_IN
EL_HB
Storage use
Stor. content
Elec. price
-5000
-2500
0
2500
5000
7500
10000
12500
15000
17500
E
l
e
c
t
r
i
c
i
t
y

p
r
o
d
u
c
t
i
o
n

(
M
W
)
Wind
Natural gas (NG)
Hydro
Wood waste (WW)
Peat
Solid wood (WO)
Forest residues (WR)
Municipal waste (MW)
Nuclear
Electricity to heat
Fig. 4. Electricity production. Negative production indicates the use of electric heat boilers and/or heat pumps.
Fig. 3. Example of operation in Industrial heat area. Negative production indicates charging of heat storage.
-2000
-1000
0
1000
2000
3000
4000
H
e
a
t

p
r
o
d
u
c
t
i
o
n

(
M
W
)
0
75
150
225
300
375
450
E
l
e
c
t
r
i
c
i
t
y

p
r
i
c
e

(
/
M
W
h
)
H
e
a
t

s
t
o
r
a
g
e

c
o
n
t
e
n
t

(
%
)
MW_HB_RU
WR_EX
WW_EX
NG_BP_RU
NG_CC_EX
PE_BP_RU
WO_BP_RU
EL_HP
EL_HB
Storage use
Stor. content
Elec. price
-600
-400
-200
0
200
400
600
800
1000
1200
H
e
a
t

p
r
o
d
u
c
t
i
o
n

(
M
W
)
0
75
150
225
300
375
450
E
l
e
c
t
r
i
c
i
t
y

p
r
i
c
e

(
/
M
W
h
)
H
e
a
t

s
t
o
r
a
g
e

c
o
n
t
e
n
t

(
%
)
NG_EX_UR
NG_BP_UR
MW_HB_UR
MW_BP_UR
EL_HP
EL_HB
Storage use
Stor. content
Elec. price
Fig. 2. Example of operation in Rural heat area. Negative production indicates charging of heat storage.
Fig. 1. Example of operation in Urban heat area. Negative production indicates charging of heat storage.
September 5
th
to September 7
th

196

Effects of heat measures in the three heat areas
In the Industry heat area availability of heat measures
(electric heat boilers, heat pumps, and heat storages)
had relatively little effect (Fig. 2). The main reason is
that the existing heat production capacity from
industrial wood waste and the associated no-cost

0
5000
10000
15000
20000
25000
B
a
s
e
O
n
l
y
H
e
a
t
B
a
s
e
O
n
l
y
H
e
a
t
Cap. Prod.
H
e
a
t

c
a
p
a
c
i
t
y

(
M
W
)
H
e
a
t

p
r
o
d
u
c
t
i
o
n

(
G
W
h
)
HEATSTOR
EL_HB
WW_BP_IN
WR_BP_IN
PE_BP_IN
NG_BP_IN
FO_BP_IN
Fig. 5. Heat capacity and production
8
in the Industrial
heat area.

waste wood were not easily replaced. However, there
were some high wind situations with low power prices
where it was beneficial to use electric heat boilers to
produce heat and decrease heat production from wood
waste in the Industry area. There was an annual
resource limit on wood waste on the country level and
the wood waste use was transferred to the Rural heat
area. It was also profitable to install some heat storage
capacity. This enabled the full shut down of wood
waste back pressure power plants for the duration of
low electricity prices. This decreased electricity
production and gave more room for the upsurge in wind
power production.
In the Urban heat area heat measures enabled the
replacement of CHP coal units with production from
heat pumps and to smaller extent from electric heat
boilers (Fig. 6). Also wood based heat boilers were
replaced. Investment in heat storage was relatively
smaller. However, they were cycled more due to faster
charging rate.

8
Heat production is from the modelled 26 weeks and should be
multiplied by 2 to get an estimate on annual production.
0
500
1000
1500
2000
2500
3000
3500
B
a
s
e
O
n
l
y
H
e
a
t
B
a
s
e
O
n
l
y
H
e
a
t
Cap. Prod.
H
e
a
t

c
a
p
a
c
i
t
y

(
M
W
)
H
e
a
t

p
r
o
d
u
c
t
i
o
n

(
G
W
h
)
HEATSTOR
EL_HP
EL_HB
NG_HB
CO_EX
WO_HB
NG_EX_UR
NG_BP_UR
MW_HB_UR
MW_BP_UR
1
in the Urban heat
area.

The combined utilization of the heat measures was
used to shut down existing natural gas based CHP
power plants during hours of average or lower
electricity prices. During low electricity prices electric
heat boilers were used to charge heat storage.
Accordingly, during average electricity prices heat was
used from heat storage to prevent the use of electric
heat boilers. During the highest electricity prices
electric heat pumps were also shut down with the help
of heat from the heat storages.
The most important difference between Urban and
Rural heat areas is the availability of wood residues in
the Rural heat area (Fig. 7). For the most part this
resource was able to outcompete heat pumps as
means to produce heat. Heat measures still helped to
replace coal CHP. The combination of electric heat
boilers and heat storages was again a large source of
additional flexibility to the system.
0
2000
4000
6000
8000
10000
12000
B
a
s
e
O
n
l
y
H
e
a
t
B
a
s
e
O
n
l
y
H
e
a
t
Cap. Prod.
H
e
a
t

c
a
p
a
c
i
t
y

(
M
W
)
H
e
a
t

p
r
o
d
u
c
t
i
o
n

(
G
W
h
)
HEATSTOR
EL_HP
EL_HB
NG_HB
NG_CC_EX
CO_EX
WR_EX
WW_EX
WO_HB
WO_BP_RU
PE_BP_RU
NG_BP_RU
MW_HB_RU

1
in the Rural heat
area.
September 5
th
to September 7
th

197

Dynamics of heat storage
Most of the daily fluctuation in heat demand was
smoothed with heat storages and electric heat boilers
in all heat areas. If CHP units were operated, they were
usually operated at maximum heat output.
The investment cost for heat storage was assumed to
be 1840 /kWh. With the assumed ratio of 12 between
storage capacity and heat capacity this translates to
153 /kW. In comparison the capacity cost of electric
heat boilers was assumed to be 40 /kW and 50 /kW
for natural gas heat boiler. This means that investment
into heat storage capacity was not driven by need for
new capacity since heat boilers were cheaper. There
had to be operational benefits from the use of heat
storage to cover the additional investment costs.
Heat storages create operational benefits by moving
consumption from more expensive sources of heat to
less expensive by shifting demand in time. In all
heating areas whole operating ranges of heat storages
were extensively utilized. During most 168 hour periods
heat storage reached both the minimum and maximum
storage capacities. In the Rural area heat storage was
2.1% of the time either full or empty. With a larger
storage capacity this could have been reduced, but it
was not worth the investment.
The size of the heat storage in Industry area was
larger than in other areas in relation to daily heat
demand (Fig. 8). In Industry area charging of heat
storages took place over several days during higher
power prices, when wood waste CHP units were
producing extra electricity. Storing the extra heat
required larger heat storage capacity. On the contrary,
in Rural and Urban charging and discharging was
more balanced and smaller heat storage was enough.

0
20
40
60
80
100
120
140
160
180
Rural Urban Industry
H
e
a
t

(
G
W
h
)
Heat storage size
Max daily heat
Min daily heat
Average

Fig. 8. Heat storage size compared to maximum, minimum
and average daily heat demands.
In the Rural area during winter time, charging of heat
storages is mostly based on the use of electric heat
boilers. They create large amount of heat in relatively
short time during periods of low power prices. During
summer time, heat storages are charged by turning on
wood waste and forest residue CHP units. During
spring and fall CHP units operate more often, since the
heat load is larger, but still the heat storage helps to
shut them down for periods of some hours.
Urban area has similar dynamics, but during summer
time the adjustment is made by heat pumps instead of
CHP. In the winter during high power prices old natural
gas CHP units are less expensive to operate than the
heat pumps.
CONCLUSIONS
District heating systems offer good possibilities for
increasing the flexibility of the power system, if the
penetration of variable power like wind power increases
greatly in the future. According to the results, main
vessels to increase flexibility are the use of heat
storages, electric heat boilers and flexible operation of
CHP units.
Investment in electric heat boilers in district heating
systems is driven mainly by periods of very high wind
power production. The resulting cheap electricity is
converted to heat and to some extent stored in heat
storages for later use. Investments in heat storage in
turn are driven by the same mechanisms, but also to
create flexibility in the electricity production when prices
are higher. To enable this, the operation of CHP units
and heat pumps is altered with the help of heat
storages. Heat pumps mainly compete against CHP as
a source of heat. They succeed in replacing coal CHP,
but are not very competitive against wood residues.
This is naturally due to assumed costs where coal has
a considerably penalty due to CO
2
cost. Heat pumps
are not very important as a source of flexibility, since
they require lot of full load hours due to their
investment cost.
While the research has been conducted on district
heating, similar dynamics could be achieved in
household heating not connected to district heating
networks. However, the costs are likely to be larger
unless there is an existing hot water tank. Flexibility
could also be gained from district cooling or air-
conditioning units with the addition of a cold storage.
Further research should also address some of the
shortcomings of current study. Sensitivity analysis
would be important, especially concerning the cost
estimates of the analysed heat measures. Heat storage
model was very simple and this should be improved.
Heat grade, especially in the industrial environment,
can vary and the model should take this into account.
September 5
th
to September 7
th

198

Heat pumps were assumed to work at constant COP
and this is a crude approximation even if the heat
source is groundwater or sea water.
REFERENCES
[1] H. Ravn et al. Balmorel: A Model for Analyses of
the Electricity and CHP Markets in the Baltic Sea
Region. Balmorel Project 2001. See also:
http://www.balmorel.com/Doc/B-
MainReport0301.pdf
[2] J Kiviluoma and P. Meibom, Influence of wind
power, plug-in electric vehicles, and heat storages
on power system investments, Energy, Volume
35, Issue 3, March 2010, pp. 1244-1255. Elsevier.
doi:10.1016/j.energy.2009.11.004
[3] Energy consumption in the UK: overall data tables,
2009 update. Department of Energy and Climate
Change - secondary analysis of data from the
Digest of UK Energy Statistics, Office of National
Statistics and the Building Research
Establishment.
[4] Annual Energy Review 2008. U.S. Energy
Information Administration.
[5] H. Ibrahim, A. Ilinca and J. Perron, Energy storage
systemsCharacteristics and comparisons,
Renewable and Sustainable Energy Reviews,
Volume 12, Issue 5, June 2008, pp. 1221-1250.
Elsevier. doi:10.1016/j.rser.2007.01.023
[6] J.K. Kaldellis and D. Zafirakis, Optimum energy
storage techniques for the improvement of
renewable energy sources-based electricity
generation economic efficiency, Energy, Vol. 32,
pp. 22952305. Elsevier.
[7] H. Lund and E. Mnster, Modelling of energy
systems with a high percentage of CHP and wind
power, Renewable Energy, Vol. 28, 2003, pp.
2179-2193. Elsevier. doi:10.1016/S0960-
1481(03)00125-3
[8] H. Lund, Large-scale integration of wind power
into different energy systems, Energy, Volume 30,
Issue 13, October 2005, pp. 2402-2412. Elsevier.
doi:10.1016/j.energy.2004.11.001
[9] H. Lund, B. Mller, B.V. Mathiesen and A.
Dyrelund, The role of district heating in future
renewable energy systems, Energy, Vol. 35, 2010,
pp. 1381-1390. doi:10.1016/j.energy.2009.11.023
[10] K. Karlsson and P. Meibom, Optimal investment
paths for future renewable based energy systems
Using the optimisation model Balmorel,
International Journal of Hydrogen Energy Vol. 33,
2008, pp. 1777-1787.
[11] S.G. Jensen and P. Meibom, Investments in
liberalised power markets. Gas turbine investment
opportunities in the Nordic power system, Int. J.
Electr. Power Energy Syst. Vol. 30, 2008,
pp. 113124.

September 5
th
to September 7
th

199

DAILY HEAT LOAD VARIATION IN SWEDISH DISTRICT HEATING SYSTEMS
H. Gadd and S. Werner

School of Business and Engineering, Halmstad University
SE-301 18 Halmstad, Phone: +46 35 167757
henrik.gadd@hh.se, sven.werner@hh.se, www.hh.se

ABSTRACT
If daily heat load variations could be eliminated in
district heating-systems, it would make the operation of
the district heating system less costly and more
competitive . There would be several advantages in the
operation such as:
- Less use of expensive peak load power where
often expensive fuels are used.
- Less need for peak load power capacity.
- Easier to optimize the operation that leads to
higher conversion efficiencies.
- Less need for maintenance because of more
smooth operation of the plants
There are a number of ways to handle the daily
variations of the heat load. Two often used are large
heat storages or using the district heating network as
temporary storage. If it would be possible to centrally
control the customer substations, it would also be
possible to use heavy buildings connected to the
district heating system as heat storages.
To be able to find the best way to reduce or even
eliminate the daily heat load variations, you need to
understand the characteristics of the daily variations.
This paper will describe a way of characterizing daily
heat load variations in some Swedish district heating-
systems.
INTRODUCTION
For all heat generation/distribution systems, heat load
variations leads to inefficiencies. You need to design
your system for the peak load even though you only
need the top capacity for a very short period of time of
the year. This is of cause expensive. The solution to
this problem is heat storage. There are a number of
possibilities to store heat in DH systems:
- Large heat storages at the heat generation plants
- Heat storage in district heating networks
- Heat storage in heavy buildings in by allowing
small variation in indoor temperatures[1].
If it would be possible to extinguish daily variations it
would lead to several profitable advantages such as:
- Less use of expensive peak load power where
often expensive fuels are used.
- Less need for peak load power capacity.
- Easier to optimize the operation that leads to
higher conversion efficiencies.
- Less need for maintenance because of more
smooth operation of the plants
To do this some questions need to be answered:
- What input and output capacity to/from the heat
storage is needed?
- What size of the heat storage is needed?
- Are the daily heat variations in the specific system
large or small during a year?
METHOD
Nomenclature
P
h
= Present hour value [MWh/h]
P
d
= Mean hour value for the present day [MWh/h]
P
a
= Mean hour value for the whole year [MWh/h]
S
h
= Energy transfer capacity [MWh/h]
S
d
= Size of heat storage [MWh/day]
S
a
= Total annual daily heat load variation
h
t = Momentary daily variation [h/h]
d
t = Total daily variation [h/day]
a
t = Total annual relative daily variation [h/year]
Variables
Measured data has been collected from some district
heating systems in Sweden. The collected data is the
heat power that is generated and fed into the district
heating network. It is hour mean power that is used, i.e.
8 760 data points per year. Only whole years is used
from 1 of January to 31 of December. To describe the
daily variation three variables is defined.
1. Momentary daily variation (
h
t )
2. Total daily variation. (
d
t )
3. Total annual relative daily variation. (
a
t )
Three system examples are presented in this paper to
exemplify the method to characterize district heating
daily heat load variation:
System A: From a city in South of Sweden with an
annual heat generation of 200 GWh.
September 5
th
to September 7
th

200

System B: From a city in Southwest of Sweden with an
System C: From a city in the middle of Sweden with an
Momentary daily variation (
h
t )
The momentary daily variation is proportional to the
amount of heat that needs to be fed in or out to the DH
network to extinguish the daily variation. This variable
describe the heat power capacity needed for in and out
put from and to the heat storage. For each district
heating systems you will get 8 760 (8 784 during leap
years) values per system and year.
The momentary daily variation is defined as the
difference of each hourly measured value and the
mean value of heat per hour of the same day divided
by the mean heat per hour of the year.
a
d h
h
P
P P
= t
The momentary daily variation is presented in Fig. 1 for
the three example systems. The figure shows that the
variations are more pronounced in the two smaller
systems compared to the larger system.
Momentary daily variation
-0,7
-0,6
-0,5
-0,4
-0,3
-0,2
-0,1
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0 1000 2000 3000 4000 5000 6000 7000 8000
Hour of the year
M
o
m
e
n
t
a
r
y

d
a
i l y

v
a
r
i a
t
i o
n
,

h
[
h
/
h
]
Systen A
SystemB
SystemC
Fig. 1 Momentary daily variation sorted by size hour by
hour for the three different district heating systems.
Total daily variation ( d
t
)
Total daily variation is defined for each day and is a
variable that is proportional to the amount of heat that
divert from the daily mean heat load. If you want to
extinguish the daily variation in a system this variable
describe the size of the heat storage. For each DH
systems you will get 365 (366 during leap years) values
per system and year.
The total daily variation is defined as the sum over the
day of the difference of each hourly measuring value
and the mean value of energy per hour of the same
day divided by two times the mean energy per hour of
the year.
a
d
h
h
d
P
P P
=

=
24
1
2
1
t
The total daily variation is presented in Fig. 2 for the
three example systems. The figure verifies that the
variations are more pronounced in the two smaller
systems compared to the larger system. Another
implication is that the highest day values are very few,
giving an incentive to construct heat storages
somewhat smaller than the peaks in the figure. Hence,
the investment costs will be reduced more the lost
benefits from the storage, giving a more optimised heat
storage.
Total daily variation
0
0,5
1
1,5
2
2,5
3
3,5
4
4,5
5
- 50 100 150 200 250 300 350
Days of the year
T
o
t
a
l
d
a
i l y

v
a
r
i a
t
i o
n
,

d
[
h
/
d
a
y
]
Systen A
System B
System C

Fig. 2 Total daily variation sorted by size day by day for
the three different district heating systems.
Total annual relative daily variation (
a
t )
Total annual daily variation is a variable that is
proportional to the total amount of energy that at daily
basis divert from the mean value accumulated for a
period of one year. It is used to compare different
systems between themselves. For each DH systems
you will get 1 value per system and year.
Total annual daily variation is defined as the sum over
the year of the difference between each hourly
measuring value and the mean value of energy per
hour of the same day divided by two times the mean
energy per hour of the year.
a
d
d h
h
a
P
P P
=

= =
365 , 8760
1 , 1
2
1
t

The annual daily variation is presented in Fig. 3 for 10
different Swedish district heating systems. Since the
annual daily load variation has a magnitude of
250500 h, only 36% of the annual heat load is
generated above the daily average heat loads. Hence,
it is the seasonal variations that dominate the heat load
variations in the Swedish district heating systems.
September 5
th
to September 7
th

201

Total annual daily variation
0
50
100
150
200
250
300
350
400
450
500
1 10 100 1 000 10 000
Annual heat supply [GWh]
T
o
t
a
l
a
n
n
u
a
l
d
a
i l y

v
a
r
i a
t
i o
n
,

a
[
h
/
y
e
a
r
]
SYSTEM C
SYSTEM B
SYSTEM A

Fig. 3 Total annual daily variation for 10 different district
heating systems in Sweden.
RESULTS
To characterize daily heat load variations in district
heating systems three variables have been defined.
h
t = Momentary daily variation
d
t = Total daily variation
a
t = Total annual daily variation
Together with the mean annual heat per hour (P
a
) and
the energy transfer capacity in and out of the heat
storage, size of storage to extinguish the systems daily
variation and the total daily variation and can be
determined according to the expressions below.
Energy transfer capacity:
S
h
=
h
t P
a
[MWh/h]
Size of heat storage:
S
d
=
d
t P
a
[MWh/day]
Total annual daily heat load variation:
S
a
=
a
t P
a
[MWh/year]
CONCLUSIONS
An expected conclusion would be that large district
heating systems have smaller relative daily variations
(
a
t ) than small district heating systems. There are two
reasons for that:
1. In a large district heating system, the use of heat
power is spread on different distances from the heat
plant, i e the chilled water in the return pipe return back
to the heat generation at different time compared to
when the return water left ach substation (geographical
diversity)
2. In large district heating networks, you would expect
that the operators have more active operation of the
heat distribution network with respect to temporary heat
storage.
But as can be observed in the Fig. 3 there does not
seem to be such a trend. One explanation could be that
the heat users differ in different systems. e.g. in the
system in Fig 3 with an annual heat supply of 9 GWh,
mostly single and multi family houses are connected
and very few industry or office buildings are connected.
Since there is a large diversity among the annual daily
variation more data need to be collected to be able to
make any further conclusions.
REFERENCES
[1] Olsson L, Werner S: Building mass used as short
term heat storage, The 11th International
Symposium on District Heating and Cooling
Reykjavik 2008.
September 5
th
to September 7
th

202

DISTRICT HEATING AS PART OF THE ENERGY SYSTEM:
AN ENVIRONMENTAL PERSPECTIVE ON PASSIVE HOUSES
AND HEAT REPLACING ELECTRICITY USE
Morgan Frling
1,2
and Ingrid Nystrm
3

1
Engineering and Sustainable Development, Mid Sweden University, stersund, Sweden
2
Chemical Environmental Science, Chalmers University of Technology, Gteborg, Sweden
3
CIT Industriell Energianalys, Gteborg, Sweden

ABSTRACT
Energy use for space heating, hot tap water and other
heat use at comparatively low temperature levels
represent a substantial part of the total energy use in
Sweden and countries with similar climate. It is thus of
importance to meet this demand in a way generating as
small environmental impact as possible. However, it is
possible to create a system with higher environmental
impacts with energy efficient buildings compared to
less energy efficient buildings through choice of less
good energy carriers. It is not enough that the
individual parts of a system are good and efficient to
give a low environmental impact; the parts must be
connected into the system in a good way.
From environmental perspective energy efficient
buildings and district heating dont oppose each other
good parts connected in a good system will give an
optimal. The results from the study of the three items of
household equipment show possibilities for district
heating to be an alternative with good environmental
performance, but not under all heat generation
regimes.
INTRODUCTION
It is of importance to meet for space heating, hot tap
water and other heat use at comparatively low
temperature levels in a way generating as small
environmental impact as possible. This can be done by
increasing the efficiency in the use phase and in the
heating systems of buildings as well as through heat
generation systems with low environmental impact.
During recent years there has been a focus on houses
with low need of space heating, low energy houses or
passive houses. In such buildings the heat from the
incoming sun radiation together with body heat from
people living in the houses and different household
equipment will cover the whole or at least substantial
parts of the space heating need over a year (extra
heating might be needed during the coldest days of a
year). Hot tap water still need to be heated. For parts of
the year this can be achieved by solar panels, but there
is a need for extra heating during winter. This might
result in the extra heating demand being covered by
electricity, directly or indirectly.
Increased energy efficiency is in itself a desirable goal
for a society it increases the robustness of the energy
system and the possibilities for a resource efficient and
more sustainable energy system in the long run.
However, it is possible to create a system with higher
environmental impacts with energy efficient buildings
compared to less energy efficient buildings through
choice of less good energy carriers. It is not enough
that the individual parts of a system are good and
efficient to give a low environmental impact; the parts
must be connected into the system in a good way.
Thus it is important to identify system solutions that
avoids sub optimization and gives us energy efficient
buildings and an efficient energy system with good
environmental performance.
In a synthesis studies within the framework of
Chalmers Energy Center [1] the role of district heating
in a future society with more energy efficient buildings
have been investigated. Here we report on general
findings of this study with a special focus on the
environmental performance of the possibility to convert
some household electricity use into district heating - for
the use in dish washers, washing machines and tumble
driers [2]. The environmental performance is studied
using life cycle assessment methodology and different
assumptions regarding electricity and district heating
generation.
DISTRICT HEATING DEMAND SIDE
There are today several drivers in the direction of lower
total heat market for district heating in future [1].
Among possible such drivers in Sweden are:
- Warmer climate (due to climate change)
- Higher energy prices
- Increased environmental awareness
- Increased energy efficiency of existing building
stock
- Limited amounts of new housing
- New housing more energy efficient
However, there are also possible drivers for a larger
heat market in future, e.g.:
- Increased wealth giving larger living space per
person and higher demands on comfort
September 5
th
to September 7
th

203

- Electricity prices might increase faster than heat
prices might lead to interest in heat instead of
electricity for new applications (washer,
dishwasher, et.c.)
- Heat for comfort cooling
- Increased use of heat for other purposes e.g.
drying of biofuels et c.
With strategic planning the resulting effect for district
heating might be a lower total but at the same time
more even demand of heat (Fig. 1).
DISTRICT HEATING SUPPLY SIDE
A strategic role of district heating in the energy system
is the ability to utilize and deliver resources that
otherwise would have been lost. Among possible
system drivers on the supply side in Sweden are [1]:
- Increased utilization of industrial surplus heat
- Remaining large potential of waste incineration
- Increase of CHP power production

Fig. 1 Possible change for district heating demand in
future decreasing demand but more even over the year.

Fig. 2 Focus on the use of biomass e.g. for making optimal
amounts of high qualitative energy carriers with heat as a
residue (it could also e.g. be biomaterials production).

At the same time we can also expect:
- Increased competition for bio fuel resources
- Higher prices on high quality energy carriers
(electricity and fuels) might drive towards smaller
fraction as heat.
- Increased energy efficiency in industrial processes.
With strategic planning district heating might utilize
residual heat from processes producing combinations
of high quality energy carriers (or bio based material
production). The focus can probably not be on heat
production. Even combined heat and power production
from bio fuels might not be efficient enough for
competitive district heating (Fig 2).

Fig. 3 Illustration of the need for a systemic perspective in planning the details of the energy system; a): A CHP plant and a
potential energy customer (building); b): A CHP plant delivering district heat and electricity to a customer; c): A power plant
delivering only electricity to a customer with passive house standard using electricity for hot water and peak heat demands
excess heat is cooled away. The total primary energy demand increases; d): A CHP plant delivering both heat and
electricity to a customer with passive house standard (less total primary energy demand than in the b case).

a) b)
c)
d)
September 5
th
to September 7
th

204

THE OVERALL ENERGY SYSTEM
The energy system of a country is complex, and it is
important to understand how changes in sub systems
may affect the whole system. Sub optimizations might
easily occur. A simplified example of a situation where
a more energy efficient building through sub
optimization of the total system gives a larger overall
primary energy need is illustrated in Fig. 3. Obviously it
is possible to create a system with higher
environmental impacts with energy efficient buildings
compared to a system with less energy efficient
buildings. It is not enough that the individual parts of a
system are good and efficient to give a low
environmental impact; the parts must be connected into
the system in a good way.
Thus it is important to identify system solutions that
avoids sub optimization and gives us energy efficient
buildings and an efficient energy system with a good
environmental performance.
IMPLICATIONS OF NEW TYPES HEAT LOAD
To better understand implications of different new types
of heat load (as illustrated in the right hand side of
Figure 1) a life cycle assessment (LCA) has been
performed regarding the use of heat instead of
electricity for the three examples of house hold
appliances: dish washer, washing machine and tumble
drier. Basic data regarding the appliances are
exemplified with those in the district heating villa in
Gteborg, Sweden. The LCA model includes energy
production (electricity or/and heat) for an average use
of each machine and the materials needed to produce
it. Different types of energy mixes for electricity and
district heat generation were studied. Details of the
system boundaries and data can be found in the full
report of the study [2].
The results indicate that the total energy system
influences the results greatly. If we consider electricity
production with large environmental impacts, to utilize
district heating is a good alternative, even in cases
where the district heating generation in itself is not
optimally environmentally friendly. This is exemplified in
Fig. 4 where we consider Swedish average district
heating fuel mix (bio and residue heat, but also fossil
fuels and some peat [5]) and European average
electricity generation. If we for the long term
development consider electricity generation that is
much less fossil carbon intensive and compare it with
district heating based on forest bio fuels the results are
much more narrow, and it become important what
environmental impact category is considered. In Fig. 5
this is exemplified with climate impact and acidification
impact.
If district heating should continue to be seen in general
as an environmentally preferable option it is important
that district heating companies continue to develop
district heating production in a favourable direction.
Heat for district heating should originate from resources
that are otherwise wasted. In the long term that will
mean that bio fuelled district heating is not enough, but
heat from other primary production like bio energy or
biomaterial combines producing transport fuels and/or
bio based materials.
CONCLUSIONS
From environmental perspective energy efficient
buildings and district heating dont oppose each other
good parts connected in a good system will give an
optimal. It is not enough that the individual parts of a
system are good and efficient to give a low
environmental impact; the parts must be connected into
the system in a good way. The results from the study of
the three items of household equipment show
possibilities for district heating to be an alternative with
good environmental performance, but not under all
heat generation regimes. Heat generation must
continuously be considered.

Fig. 4 Environmental impact from using district heat for
dishwasher, drier and washer. Case: Swedish av. district
heating and European av. electricity.
September 5
th
to September 7
th

205

Fig. 5 Environmental impact from using district heat for
dishwasher, drier and washer. Case: bio based district
heating production and Swedish av. electricity.
ACKNOWLEDGEMENT
Financial support from the Knut and Alice Wallenberg
foundation and the Swedish District Heating
Association is gratefully acknowledged.
REFERENCES
[1] Ingrid Nystrm, Martin Eliasson, Torbjrn
Lindholm, Morgan Frling, Jan-Olof Dahlenbck,
Erik Ahlgren and Elsa Fahln (2009): Energieffektiv
bebyggelse och fjrrvrme i framtiden (in Swedish:
Energy efficient built environment and district
heating in future). Swedish District Heating
Association, Stockholm, Sweden. Available as pdf
from www.svenskfjarrvarme.se
[2] Morgan Frling and Ingrid Nystrm (2009):
Miljpverkan frn energieffektiva hus och
alternativ vrme- eller elanvndning (in Swedish:
Environmental impacts from energy efficient
buildings and alternative heat or electricity use).
Published in [2].
[3] Morgan Frling; Charlotte Reidhav; Jan-Olof
Dalenbck and Sven Werner (2008): Is there a role
for district heating in future cities with low energy
buildings? 11th International Symposium on District
Heating and Cooling, August 31 to September 2,
2008, Reykjavik, ICELAND
[4] Gteborg Energi. Fjrrvrmehuset (published in
Swedish; The district heating house). Brochure.
Gteborg Energi AB.
[5] Morgan Frling (2004): Environmental limitations
for the use of district heating when expanding
distribution into areas with low heat density. 9th
Cooling, August 30-31, 2004, Espoo, Finland.
September 5
th
to September 7
th

206

ADAPTIVE CONTROL OF RADIATOR SYSTEMS FOR A LOWEST POSSIBLE
RETURN TEMPERATURE
P. Lauenburg and J. Wollerstrand
Lund University, Faculty of Engineering, Department of Energy Science, Sweden,
patrick.lauenburg@energy.lth.se

ABSTRACT
The present paper describes how the control of a
radiator system connected to a district heating network
via a heat exchanger can be optimised to provide the
lowest possible district heating return temperature. This
can be achieved for each operating point by employing
an optimal combination of radiator circuit supply
temperature and circulation flow rate. The control
algorithm gradually creates a modified control curve for
the radiator circuit, enabling it to consistently provide
an optimal cooling of the district heating water. Since
the heat exchanger is dimensioned for very low outdoor
temperatures, it is oversized for all other heat loads. In
addition, radiator systems are often oversized due to
safety margins. Such facts render it possible to reduce
the district heating return temperature.
The objective of the present study was to develop a
control algorithm and to test it in practice. A description
is here given of the algorithm, as well as of field tests
that were carried out to practically verify it. The control
method could be implemented in any modern control
logics for adaptive control of a radiator circuit, and the
obtained results indicated that one can expect a
lowering of the return temperature in line with previous
theoretical calculations.
INTRODUCTION
The present paper demonstrates how the control of a
radiator system connected to a district heating (DH)
network via a heat exchanger (HEX) can be optimised
to provide the lowest possible DH return temperature.
This is done by always choosing the optimal radiator
supply temperature and flow rate.
Relevance of the topic
Low return temperatures are beneficial for the
production as well as the distribution of DH. A specific
advantage of the control method demonstrated in this
paper, as opposed to, for example, conventional low
flow balancing, is its robustness, enabling the lowest
possible return temperatures to be consistently
obtained. This is the case independently of the current
outdoor temperature and heat load, even if the DH
supply temperature changes, the HEX becomes fouled,
or the house heating requirements change. The idea is
also to utilise the fact that, since a HEX is dimensioned
for an extremely low outdoor temperature, it is in fact
oversized for all other heat loads. In addition, radiator
systems are generally also oversized for safety
reasons, as presented in both Swedish studies [3], [12]
and international ones [5], [8] and [10], thus providing
further potential to reduce the return temperature.
Objective
The objective of the study was to develop a control
algorithm for determining the optimal choice of supply
temperature and flow in an arbitrary radiator system for
every heat load in order to minimise the primary return
temperature.
Limitation
The present investigation has dealt with DH
substations that were indirectly connected to the
DH network, i.e., hydraulically separated by HEXs.
OPTIMISED HEATING SYSTEM TEMPERATURES
There exist various ways to control the heat output in a
heating system. Here, we have dealt with the prevailing
control method used in Sweden; an outdoor
temperature-compensated supply temperature,
ensuring that an adequate amount of heat is supplied
to the building at each outdoor temperature.
The benefits with regard to the primary return
temperature from adjusting the flow according to the
heat load are known. The idea of using an optimal
combination of flow and supply temperature was
conceived by Frederiksen and Wollerstrand [2], and
this theory has been further studied [13] [11]. The
guidelines from Euroheat & Power [1] state that the
lowest return temperature is obtained by varying the
flow according to the consumption. If such a variable
flow is used, it is controlled by thermostatic radiator
valves (TRV) either in combination with a constant
supply temperature or with an outdoor temperature-
compensated supply temperature. Langendries [4]
suggests a central control of the flow rate through the
pumps rotating speed, but claims that it appears to be
a rather difficult and expensive system. Petitjean [9]
proposes a lowering of the pump speed at low heat
loads, when the TRVs are almost fully open, but finds it
problematic to determine which parameter to use for
controlling the pump speed.
It should be possible to implement the control algorithm
presented in this paper in any modern, state-of-the-art
September 5
th
to September 7
th

207

control logics for building automation, which are today
often used for controlling DH substations. The control
method suggests how the flow can be determined for
each heat load. The flow is regulated by adjusting the
pumps rotating speed. Speed-controlled pumps are
commonly used nowadays and they provide a superior
controllability [1], [10].
Let us first study an example of an optimal control
curve for a 100 % oversized system. Such a curve is
presented in Fig. 1, which also shows the relative
magnitude of the varying radiator flow in relation to the
required flow. The blue dashed line in the diagram
corresponds to the primary return temperature. For the
sake of comparison, the primary return temperature for
a 55/45 C system is also shown (gray dashed line).

30
40
50
60
70
80
90
100
T
e
m
p
e
r
a
t
u
r
e

[
C
]
-15 -10 -5 0 5 10 15
0
25
50
75
Outdoor temperature [C]
R
e
l
.

f
l
o
w

[
%
]
m
s
T
p,s
T
p,r,opt
T
s,s,opt
T
s,r,opt
T
p,r,55/45
Primary return temperature reduction

Fig. 1 Temperatures with an optimised temperature curve
and a variable flow in a 100% oversized system. The
primary return temperature from a 55/45 C programme is
shown for comparison.

Flow-weighted, yearly mean primary return
temperatures from the radiator HEX have been
calculated with regard to the outdoor temperature
duration. Above the dashed line in Table 1, results are
shown for a correctly dimensioned system, with an
80/60C programme as well as with an optimised
programme. The gain is estimated to just under two
degrees C. The last column shows how the primary
return temperature is affected when the length of the
HEX is doubled. This comparison can be justified by
the fact that the primary return temperature is
significantly influenced by the lower secondary flow that
the optimisation entails, while the pressure drop and
heat transfer rate in the HEX can remain at a
magnitude close to the original ones.
Table 1: A summary of flow-weighted mean primary return
temperatures (bold) and resulting reduction for various
temperature programmes.

Under the dashed line, results are shown for a system
that is oversized by 100 %. The first three temperature
programmes are 55/45, 60/40 and 80/30 C, whereas
the last two are optimised ones with variable flow.
The following conclusions could be drawn from the
table:
- The oversizing of a radiator system leads, in itself,
to a significant reduction of the primary return
temperature, provided that some kind of
compensation has been made in order for the
system to work properly, i.e., that an accurate
indoor temperature has been provided.
- By optimising the system (through the use of a
variable secondary flow), the primary return
temperature can be further reduced, especially if
the system is oversized.
- By extending the radiator HEX, the return
temperature can be further reduced with the
temperature programmes that employ a relatively
low flow.
- Regardless of the degree of oversizing, a
combination of an optimised temperature
programme and an extended HEX provides a
substantially reduced primary return temperature.
The values presented in the table have been calculated
only for the radiator HEX. When considering the
substations total return temperature, it can be said to
be smoothed by the DHW consumption. Calculations
corresponding to those in Table 1 for a parallel and a 2-
stage substation for 20 flats (based on the Swedish
District Heating Associations recommendations for
sizing) result in reductions in the return temperature
that are approximately 20 % lower than the values
shown in the table. The difference between the parallel
and the 2-stage connection is negligible when the
return temperature from the radiator HEX is low or
moderate, a fact that has been previously
September 5
th
to September 7
th

208

demonstrated [6] [3]. Euroheat and Power recommend
that a 2-stage connection be used only in large multi-
residential buildings if the primary radiator return
temperature is high. However, it should not be
employed if a low-flow heating system providing low
return temperatures is used [1].
The advantage of extending the HEX when the
secondary flow is low actually demonstrates the
optimisation problem: When the secondary flow is
reduced, the secondary return temperature will
decrease. In the radiator HEX, the situation is different.
As the secondary flow decreases, the difference
between the primary and the secondary return
temperatures, increases as a result of the heat transfer
coefficient in the HEX being strongly flow dependent.
Fig. 2 shows how the secondary return temperature is
lowered with a decreasing secondary flow while the
difference between primary and secondary return
temperatures increases. This results in a primary return
temperature that, at first, decreases and then increases
when the secondary flow is further reduced. The values
in the figure have been taken from one of the test
objects. For this heat load, the lowest primary return
temperature was achieved for a secondary flow of
approximately 30 % of the original flow.

33
34
35
36
37
38
39
40
15 20 25 30 35 40 45 50
Flow, %
R
e
t
u
r
n

t
e
m
p
e
r
a
t
u
r
e
,

C
0
1
2
3
4
5
6
7
G
r
d
i
g
k
e
i
t
,

C
Tp,r,rad
Ts,r
Grdigkeit
Optimum, lowest T
p,r

Fig. 2 Primary and secondary return temperatures, as well
as the difference between them, as functions of the
radiator flow.

Another reason for including the impact of an extended
HEX in the comparison in Table 1 is the opportunity of
connecting to new installations. Large parts of the
housing stock in Sweden, built under strong political
incentives during the 1960s and 1970s, are facing
substantial renovation needs. The results of this project
can be considered consistent even if fewer radiator
systems be oversized in the future, whether
incorporated in older, renovated, or new buildings. The
smaller potential for return temperature reductions
resulting from less oversized radiator systems may be
compensated by the ability to install a HEX that is
dimensioned for of an optimised radiator programme,
i.e., a longer HEX. Furthermore, with optimised control,
there exists a preparedness for future changes in
system temperatures in the DH network. Should the
DH supply temperature be changed, an adaptive
control will ensure that the lowest possible return
temperature is always achieved.
In order to operate according to Fig. 1, the algorithm
must combine a control of the radiator supply
temperature with a control of the radiator flow as a
function of the heat load and the DH supply
temperature. In previous work [7], we have shown that
it is possible to manually determine the optimal radiator
supply temperature and flow. A natural continuation is
to develop a method for automatic adjustment of
parameter values for the optimal control algorithm.
THE TEST OBJECTS
The tests have been carried out in four multi-residential
buildings in the city of Karlshamn, Sweden. The houses
were built in 1967-1968: three of them had three stories
and a basement, and one had six stories and a
basement. The number of flats varied between 20 and
30 per house.
The radiators in all houses were fitted with TRVs, but
these were at least ten years old. It was thus uncertain
whether they functioned properly. The circulation flow
was found not to vary significantly in any of the radiator
circuits, which may have been an indication that many
of the TRVs were not working. However, it should be
noted that the presented control algorithm is
independent of the use of TRVs in a system. Whatever
combination of optimal supply temperature and flow
that is identified for a given outdoor temperature, the
heat supply will be the same. The main task for TRVs
is to limit the heat supply in a room where additional
heat supply (solar radiation, bodily warmth or electrical
equipment) would result in an overheating of the room.
The substations were of the 2-stage type and equipped
with control logics of the brand IQ Heat (Alfa Laval AB).
The equipment for the building automation was
manufactured by Siemens and furnished with a
separate communications module that could also be
used for executing minor computer programmes. There
was also an internet connection, rendering it possible
to communicate in a number of ways, such as via the
software Saphir ScopeMeter (Siemens), or FTP. After
a reconfiguration, the pump speed could be controlled,
since all pumps were equipped with communication
modules.
In order to monitor the circulation flow in the radiator
circuits during the tests, clamp-on ultrasonic flow-
meters were utilised. However, the objective was to
develop a control algorithm based on modern, state-of-
the-art equipment without using additional installations.
September 5
th
to September 7
th

209

To assure that the temperatures measured in the
substation corresponded to the average temperature
levels in the various risers in the radiator circuits,
temperature sensors were installed in two of the
houses. This enabled measurement errors or
disturbances in the radiator circuit to be identified. The
indoor temperature could be monitored thanks to six
wireless sensors installed in each house in the area.
Modifications in the substations
After some initial tests, the circulation pumps were
found to be generally oversized to such an extent that
the flow rate could not be decreased as much as
desired. There exists a predetermined minimum
rotational speed for this type of pump, implying that the
speed could be reduced by 6070%. Discussions with
the manufacturer revealed that the lowest pump speed
could not be changed in this model, for which reason
the decision was made to throttle the flow with an
existing shut-off valve located after the pump, which
shifted the pumps operating range. The throttling was
conducted in order for the pump to give half the flow
rate at 100% rotational speed. The control curve was
modified accordingly, leading to the temperature drop
in the radiator circuit becoming doubled and the heat
supply remaining unaltered.
We were unable to receive a comprehensive reply from
the pump manufacturer with respect to the possible
measures regarding the regulation of the pump. A
discussion with another manufacturer implied that there
were no technical limitations for how far down the
pump speed could be controlled. However, such an
extension of the manoeuvrable range has so far not
been requested. After a simple modification of the
pumps frequency converter, the working range could
be extended from todays 30100% to, in an extreme
case, 2100%.
Existing control of the radiator circuits
Although the radiator circuits within the area were
designed by the same consultant, there is today a large
spread in the choice of control curve and resultant
temperature drop (1030 C). It is likely that the curves
have been gradually adapted to the circuits hydraulic
properties and balancing, and one can assume that this
is a common situation.
When older houses are renovated and their radiator
circuits are modernised, there are no guarantees that
oversizing is taken into consideration. For example, the
radiator HEX in a substation that was installed in 2005
in one of the houses was dimensioned for 185 kW heat
output at DOT with temperatures corresponding to
80/60 C at a flow of 2.25 l/s. However, when
examining data for this substation, it turned out that the
substation delivered less than 40 kW at an outdoor
temperature around 0 C, which corresponded to a
load of approximately 50%. The actual flow rate was
about 1.1 l/s and the temperatures corresponded to
60/40 C, thus representing an oversizing around
100%.
ADAPTIVE OPTIMISATION - METHOD
In the theoretical example, the system was assumed to
be 100 % oversized, while in an arbitrary system one
cannot be sure of the degree of oversizing. It is also
desirable to have a robust and adaptive control
algorithm. The method found to function the best is
described below. This approach consists in gradually
modifying, by automatically performed tests, the control
curve and determining the associated flow rate.
Online testing
By locking the control valve (CV), one can assume to
have approximately the same primary flow through the
radiator HEX, and since the variations in the cooling of
primary water is relatively small, the heat supply is also
approximately constant. If the secondary flow is
reduced while the CV is maintained locked, the
temperature of the secondary flow leaving the HEX will
rise. When a new flow and its associated supply
temperature are tested, the current level of the primary
return temperature is compared to the level before the
experiment. In this way, the new combination of flow
and supply temperature can be either accepted or
rejected. This method renders it possible to implement
the adaptive algorithm in any arbitrary system, leading
to the control curve becoming gradually modified. This
method we suggested in [7].
One problem associated with this kind of optimisation is
that the method is sensitive to disturbances. If the
primary supply temperature, primary differential
pressure or the outdoor temperature changes during
the test, one cannot be sure that the heat supply is
constant. In that case, a reduced return temperature
could be the result of a heat supply that is too low.
Such tests have to be rejected.
In order to render the tests less sensitive to
disturbances, the CV is locked only briefly, in order for
the HEX to stabilise. Subsequently, we return to
automatic control, but instead of using the control
curve, the control aims at maintaining a constant
temperature drop in the radiator system. If this is
successful, the heat supply is also kept constant. One
can assume that the secondary flow is relatively
constant: as long as tests are conducted at night, no
solar radiation is present and internally generated heat
is likely to be at a relatively steady level. If, for instance,
the primary supply temperature or differential pressure
rises during the course of a test, the CV will close
somewhat causing the secondary supply temperature
September 5
th
to September 7
th

210

to decrease, and thereby also the temperature drop
and heat supply, to be detained at the same level.
A test is started by keeping the CV locked for ten
minutes. This leaves enough time for the HEX to
stabilise. The new level of the difference between the
primary and secondary return temperatures became
stable already after about two minutes in the tested
objects. The CV was maintained locked for ten
minutes, which should be sufficient even for very low
flows and most types of HEXs. Subsequently, the
control was resumed in order to ensure a constant
temperature drop on the secondary side.
The temperature drop was controlled by verifying the
current temperature drop, e.g., every five minutes, and
comparing it with the desired temperature drop, i.e., the
temperature that was observed when the CV was
locked. If the difference exceeded a certain value,
0.2 C has been used so far, the set-point for the
supply temperature was updated according to
T
setpoint
= T
s,r
+ AT
setpoint
.
Fig. 3 displays a performed test: At 1:00 a.m., the CV
was locked and the radiator flow rate was reduced from
0.59 to 0.36 l/s with the result that the secondary
supply temperature rose from 40 to 44 C. After ten
minutes, the temperature drop in the radiator circuit
was automatically controlled (in this case, the
temperature drop was stable and it took more than 15
minutes before the CV opening degree required
adjustment). After ninety minutes, the second flow
reduction was carried out, to 0.24 l/s, and the
secondary supply temperature increased to about
48 C.
The total primary return temperature varied to a
relatively large extent, partly because of tappings of
domestic hot water (DHW), but also due to the DHW
control in this substation being very unstable when no
tappings were made. However, the return temperature
from the radiator HEX was of interest for the tests. In
this object, the difference between the primary and
secondary return temperatures was very small, and
even for a low radiator flow, the grdigkeit was below
one degree. One can see from the figure that the return
temperature had fallen from just under 32 C to slightly
over 28 C during the test. This resulted in, for a
current outdoor temperature of 8 C, the set-point for
the secondary supply temperature being changed from
40 to 48 C while the flow should be reduced from 0.59
to 0.24 l/s.
00:30 01:00 01:30 02:00 02:30 03:00 03:30 04:00
0
10
20
30
40
50
60
70
80
Time
T
e
m
p
e
r
a
t
u
r
e

[
C
]
T
p,s
T
s,r
T
p,r,rad
T
p,r,tot
T
s,r
AT
s
T
o
T
o,damp
Gr

0
10
20
%
0
0.25
0.5
0.75
1
F
l
o
w

[
l
/
s
]
00:30 01:00 01:30 02:00 02:30 03:00 03:30 04:00
0
20
40
60
Time
H
e
a
t

s
u
p
p
l
y

[
k
W
]
CV,heat CV,DHW
m
p
m
s
Q
p
Q
s

Fig. 3 Results from a test. The flow was reduced at 1:00
and 2:30. The top graph shows temperatures in the
substation, the next graph presents the valve position for
heat and DHW, and the last two display the primary
(including DHW) and secondary flow and the primary
(including DHW) and secondary heat supply, respectively.

An interesting aspect of this test was that the primary
supply temperature fluctuated a lot. Since the
secondary temperature drop was kept constant, it had
no impact on the outcome of the test. One can see that
the CV generally demonstrated a lower opening degree
later in the night, as opposed to before 1:00, when the
primary supply temperature increased. Without the AT
control, the heat supply would have been too high
during the last part of the test.
The radiator flow was altered by changing the set-point
for the pump speed, expressed as a percentage of the
maximum speed. It has been found that two flow
alterations of ninety minutes each are suitable per test,
as this would allow the secondary return temperature to
stabilise even at very low flows. The first test for any
outdoor temperature, as was the case in Fig. 3, means
that starting conditions include the original control
curve and flow rate. It is then desirable to perform two
fairly large flow reductions since, according to the
theoretical calculations, one can expect to find an
optimum at a relatively low flow. If, however, the flow is
already on a low level, it is reasonable to attempt one
slightly higher and one slightly lower flow rate. The
algorithm for the adaptive control is illustrated by the
flow chart in Fig. 4.
September 5
th
to September 7
th

211

11:50 PM < time < 0:00 AM
Get T
o
, use
modified control
curve
Wait 10 min No
Wait 60 min with
constant pump speed
Save current
values
Set pump speed
for test
Set control
valve to Manual
Wait 10 min Set AT
Set control
valve to Auto
AT AT
set-p
> 0.2C
Yes T
set-p
= T
s,r
+ AT
set-p
No
Wait 5 min
< 80 min?
No,
> 80 min
Pump speed 1? Yes
Start timer
Yes No, pump speed 2
Test done
Determine T
p,r,rad,min
(pump(0), pump(1)
or pump(2))
Check maximum
deviation for Q
(e.g., 5%) and To
(e.g., 2C)
Not ok Reject test
result
Ok Update
curves
Test pump
speed 1
Test pump
speed 2
Yes

Fig. 4 Flow chart describing the adaptive control algorithm.

If a modified control curve is used before a test is about
to start, the control should be interrupted and the pump
speed kept constant for an hour prior to the test. This
way, one avoids the risk of the flow changing (due to
alterations in the outdoor temperature) too close to the
test, which could result in unstable radiator system
temperatures.
The supply and return temperatures were measured on
four of the most remote risers from the substation,
during the tests. A continuous matching against
measurements on risers gives a good indication that
the flow distribution in the system was not impaired by
the optimisation. The temperature profile was closely
matched to the profile at the substation. Both flow
reductions resulted in increased temperature drops.
Updating the control curves
After the completion of a test, the obtained information
needs to be evaluated. The influence of the variation of
the outdoor temperature is not entirely obvious; its
influence decreases with an increasing time constant
for the building. Variations on the primary side normally
have is compensated for since the heat supply is kept
constant. As a result, it is sufficient to verify that the
heat supply was maintained at a steady level during the
test, avoiding any disruptions.
If a test result is accepted, the primary return
temperatures for each tested flow are compared in
order to verify which flow resulted in the lowest return
temperature. This flow also gave rise to a secondary
supply temperature. It is however not obvious how to
read this temperature, given that it was regulated by
the controller and changed continuously. The most
logical choice is to read the mean value at the end of
the test period, before the pump speed changes. At this
point, one could expect a stable secondary return
temperature, e.g., during the last five minutes. In
addition to the secondary supply temperature, also the
primary supply temperature is recorded. However, the
dampened outdoor temperature, i.e., the input signal to
the controller, is recorded when the CV is locked for the
first time. The reason for this is that the heat supply is
subsequently kept constant at a level matching the
outdoor temperature (and heat load) at the time before
the test was started.
The next step consists in using the information attained
from the test to modify the control curves. Initially, the
original curve was used and the pump was, in our
case, controlled to give a constant differential pressure.
If the result of a test is that a lower primary return
temperature is obtained at a lower secondary flow rate,
the control curve is updated for that outdoor
temperature. A reasonable resolution is 1 C. The
original control curve, generally based on 58 points,
was therefore initially extended to comprise values for
each outdoor temperature.
If the experiment, as in Fig. 3 above, was performed at
8 C, this point on the curve would be updated. Along
with the new supply temperature there followed a new
radiator flow, which in our case was expressed as a
new set-point for the pump speed.
The adaptive control continues in this manner night
after night, and the control curves are continuously
updated. Outside the test periods of approximately
three hours each night, the modified control curves are
used for controlling the heating system.
Fig. 5 shows an example of the gradual development of
the modified control curve. The first graph shows a new
point at 0 C (used for 0 0.5 C). In the second
(upper) graph, a point for 3 C has been added, while
the range 0 to 3 C is complete in the third. The fourth
graph shows a much more complete control curve
(-5 to 10 C). Temperature curves corresponding to
constant flow systems with lower flows than the original
system have been included as thinner lines. The value
for 10 C coincides with the curves of a system with a
low flow, while the value of -5 C coincides with the
curves of a system with a moderately reduced flow
(normal flow). The last graph clearly demonstrates that
the modified curves are based on a variable flow, i.e.,
they coincide with various constant flow curves at
different points.
As shown in the second graph of Fig. 5, the modified
curve could emerge in sections that subsequently are
combined. One way to speed up the modification of the
control curves is to interpolate intermediate values
rather than wait for a flow optimisation at the missing
outdoor temperature. Even the return temperatures
could be interpolated, since it is possible to determine
September 5
th
to September 7
th

212

the required radiator flow for a known temperature drop
(and heat supply).
20
30
40
50
60
70
-10 -5 0 5 10 15
Outdoor temperature
T
e
m
p
e
r
a
t
u
r
e
20
30
40
50
60
70
-10 -5 0 5 10 15
Outdoor temperature
T
e
m
p
e
r
a
t
u
r
e
20
30
40
50
60
70
-10 -5 0 5 10 15
Outdoor temperature
T
e
m
p
e
r
a
t
u
r
e
Constant, further
decreased flow
Constant,
decreased flow
Original
control
curve
Modified
control curve
20
30
40
50
60
70
-10 -5 0 5 10 15
Outdoor temperature
T
e
m
p
e
r
a
t
u
r
e

Fig. 5 A stepwise modification of the control curve. The
supply temperatures are drawn in solid lines while the
returns are dashed.
For the first test to be carried out at a specific outdoor
temperature, it is logical to let the results of this test
fully replace the original points on the curve. As more
tests are performed for the same outdoor temperature,
one can proceed in several ways. Since the control
should be adaptive and thus able to take into account
changing circumstances both in the DH network and in
the building, the results of new tests should be
employed. However, one may expect that tests
performed close to one another in time, at equivalent
outdoor temperatures, still provide slightly differing
results for varying reasons. A solution would therefore
be to use a forgetting factor, i.e., to gradually forget
old values when the supply temperature curve is
updated with new data. A possible approach for doing
so consists in calculating the new supply temperature,
T
s,s,n
, as a mean value of the obtained, T
s,s,test
, and the
last used, T
s,s,n-1
, supply temperature according to:

2
1 , , , ,
, ,
+
=
n s s test s s
n s s
T T
T (1)

When a new test is performed at the same outdoor
temperature, a new mean value is calculated, which
means that older values will have less and less
influence. To determine the secondary flow associated
with the new supply temperature, i.e., the one providing
the correct heat supply at the current outdoor
temperature, the expected temperature drop is
calculated as:

n r s n s s n s
T T T
, , , , ,
= (2)

where T
s,r,n
is determined in analogy with T
s,s,n
,
according to:

2
1 , , , ,
, ,
+
=
n r s test r s
n r s
T T
T (3)

To ensure that the heat supply is kept constant, the
required flow for the new temperature drop is
calculated. Since the flow is inversely proportional to
the temperature drop, it can be determined from the
last used flow and temperature drop, together with the
new temperature drop, according to:

n s
n s
n s
T
T m
m
,
1
,
) (
A
A
=

(4)

As mentioned earlier, the flow rate is set by changing
the set-point for the pump speed. According to the
affinity laws for fluid machines, the flow is proportional
to the rotational speed. The process of letting the last
modified supply temperature and the result of a new
test form a new modified supply temperature is
illustrated in Fig. 6.
20
30
40
50
60
2 4 6 8
Outdoor temperature
T
e
m
p
e
r
a
t
u
r
e
Modified
curves
T
s,s,test
T
s,r,test
Original
curves
T
s,s,n
T
s,r,n

Fig. 6. An approach for modifying the control curve based
on new test results.
The proposed method for updating the control curves
indicates that if for instance the DH utility demonstrates
a long-term change in the supply temperature in the
network, the control system gradually adapts to the
new temperature. However, there are always variations
in the primary supply temperature. This may include
both unintended and intended variations which may be
the result of, for example, a charging of the network if
the outdoor temperature is expected to fall. Since the
primary supply temperature affects the primary return
September 5
th
to September 7
th

213

temperature, it is desirable for the adaptive control to
also compensate for such short-term variations. One
way of doing so is to develop a number of parallel
control curves for various intervals of the primary
supply temperature. If the temperature is greater than a
certain level, an alternative control curve is employed,
whereas if it is below a certain level, one utilises
another. This method has yet to be tested and there is
no basis for assessing how much impact one can
expect from normal variations in the supply
temperature or what would constitute reasonable
intervals for parallel control curves in this case. Another
variant could be to perform a linear adjustment for the
secondary supply temperature depending on the
primary supply temperature, according to:
)) ( 1 (
, 0 , , 0 , , , s p s p s s s s
T T a T T + = (5)

where a is a constant that can be determined from
tests.
Regarding the measurement of temperatures and flows
Regarding the temperature measurement in the
substation, supply and return temperatures on both the
primary and the secondary sides are required. One
should keep in mind that, on the primary side, the
return temperature from the radiator HEX is needed
since the total return temperature is affected by the
DHW system. This temperature is normally available in
modern substation control equipment.
It is desirable to avoid installation of a flow-meter in the
secondary circuit. On the primary side, where the
energy-meter is located, the total primary flow and the
total temperature drop in the substation are measured
and the energy required for DHW provision is thus
included. Since the tests are performed at night, DHW
tappings can be avoided to a large extent. By closing
the DHW CV for a short time, the primary flow passes
exclusively through the radiator HEX. By comparing the
average level of heat supply with a closed valve to the
level prior to closing the valve, the flow required for
DHW re-circulation can be estimated.
In the test objects, indoor temperature measurements
were used to verify that the adaptive control was able
to give the correct indoor temperature. However, one
can in fact be sure that the correct amount of energy is
transferred to the system for each operating point,
regardless of whether the original control curve or the
optimised curve is used. A possibility is that there is an
imbalance in the system. For example, the most distant
riser may not receive the required flow because of a
too low differential pressure when the pump speed is
decreased. It is, however, more likely that a better
balance in the system is achieved when the differential
pressure is lowered this since the pressure losses in
the system decreases and all risers receive a more
similar differential pressure. However, one must be on
the look-out for errors (e.g., short circuits) in the
systems, a problem that is often emphasised in
connection with low-flow systems, as these tend to be
more sensitive to hydraulic imperfections [12].
Reduction of the primary return temperature
To estimate a yearly mean return temperature
reduction (as presented in Table 1) achieved by the
adaptive control, an entire, or a major part of the,
heating season needs to be evaluated. The control
method presented in this paper was developed during
the winter and spring of 2009, and only a limited
number of tests were performed during the spring.
However, Fig. 7 shows the obtained primary return
temperature that was attained for the tests that were
performed in one of the houses. Note that these results
were first runs for each outdoor temperature (i.e., the
flow was reduced to approximately 40%), signifying
that no further optimisations were undertaken. The
curve displaying the original return temperatures was
based on the average return temperatures from the
radiator system prior to any of the modifications (i.e.,
for the tests or the constant flow rate change, as
described in section 3.1).
20
25
30
35
40
45
50
-10 -5 0 5 10 15
Outdoor temperature
P
r
i
m
a
r
y

r
e
t
u
r
n

t
e
m
p
e
r
a
t
u
r
e
Tp,r,rad,orig
Tp,r,rad,opt

Fig. 7. Primary return temperatures in the radiator system
when the flow is reduced (dots), compared to the original
return temperatures (curve).
CONCLUSIONS AND DISCUSSION
An adaptive control algorithm was developed in order
to minimise the DH return temperature. The control
algorithm can be implemented in any modern control
logics for building automation. Some refinement may
be done by compensating for short-term temperature
variations in the DH network. During the field studies,
limitations in the speed control of the circulation pumps
have presented a complication. A modification of the
frequency converter could increase the working range.
September 5
th
to September 7
th

214

There was not enough time to develop completely
modified control curves for the test objects during the
present heating season. On the other hand, a control
curve with an adaptive controller is never definitive;
rather it increases as more operational points (different
outdoor temperatures) are added and is then gradually
modified if outer conditions change. In order to receive
values for the primary return temperature on a yearly
basis using the adaptive control algorithm, the new
control curve needs to be modified for the entire
temperature range. During the performed field studies,
the reduction of the primary return temperature was
about 3 C. Even though the test period limited the
number of tests, the temperature range was still rather
wide, including temperatures from -2 to 14 C.
It is plausible that certain circuits are more suitable for
a variable flow rate, e.g., depending on hydraulic
balancing. It would also be possible to map out under
which circumstances other heat emitters than radiators,
such as fan coil heaters, can be included in a radiator
circuit where the flow varies.
ACKNOWLEDGEMENT
The Swedish District Heating Association, the Swedish
Energy Agency and Nordic Energy Research are
gratefully acknowledged for financing this work.
REFERENCES
[1] Euroheat & Power, Guidelines for District Heating
Substations, Downloaded from:
http://www.euroheat.org/documents/Guidelines%2
0District%20Heating%20Substations.pdf,
20081117.
[2] Frederiksen, S., Wollerstrand, J., Performance of
district heating house station in altered operational
modes, 23rd UNICHAL-Congress, Berlin, 1987.
[3] Gummrus, P., Petersson, S., Robust Fjrrvrme-
central (Robust District Heating Substation), Report
A 99-223, Dept. of Energy and Environment,
Chalmers Univ. of Technology, Gothenburg, 1999.
[4] Langendries, R., Low Return Temperature (LRT) in
District Heating, Energy and Buildings, 12 (1988)
191-200.
[5] Liao, Z., Swainson, M., Dexter, A.L., On the control
of heating systems in the UK, Building and
Environment 40 (2005) 343-351.
[6] Lindkvist, H., Walletun, H., Teknisk utvrdering av
gamla och nya fjrrvrmecentraler i Slagsta
(Technical evaluation of old and new district
heating substations in Slagsta), Report 2005:120,
Swedish District Heating Association, 2005.
[7] Ljunggren, P., Johansson, P.-O., Wollerstrand, J.,
Optimised space heating system operation with the
aim of lowering the primary return temperature,
Proceedings from 11th International Symposium on
District Heating and Cooling, Reykjavik, 2008.
[8] Peeters, L., Van der Veken, J., Hens, H., Helsen,
L., Dhaeseleer, W., Control of heating systems in
residential buildings: Current practice, Energy and
Buildings 40 (2008) 1446-1455.
[9] Petitjean, R., Total hydronic balancing, Tour &
Andersson Hydronics AB, Ljung, Sweden, 1995.
[10] Skagestad, B., Mildenstein, P., District Heating and
Cooling Connection Handbook, published by the
International Energy Agency (R & D Programme on
District Heating and Cooling), 2002.
[11] Snoek, C., Yang, L., Frederiksen, S., Korsman, H.,
Optimization of District Heating Systems by
Maximizing Building Heating System Temperature
Differences, Report 2002:S2, International Energy
Agency (R & D Programme on District Heating and
Cooling) & NOVEM, Sittard, 2002.
[12] Trschel, A., Hydronic Heating Systems The
Effect Of Design On System Sensitivity, Doctoral
Thesis, Chalmers University of Technology,
Gothenburg, Sweden, 2002.
[13] Volla, R., Ulseth, R., Stang, J., Frederiksen, S.,
Johnson, A., Besant, R., Efficient substations and
installations, Report 1996:N5, International Energy
Agency (R & D Programme on DHC) & NOVEM,
Sittard, The Netherlands, 1996.
September 5
th
to September 7
th

215

POLICIES AND BARRIERS FOR DISTRICT HEATING AND COOLING
OUTSIDE EU COUNTRIES
A. Nuorkivi
1
and B. Kalkum
2

1
Energy-AN Consulting
2
Energy & Utility Consulting

ABSTRACT
The policies and barriers faced by DHC in the countries
outside the EU will be investigated during 20102011
as a part of the Annex IX of the IEA Implementing
Agreement on District Heating and Cooling (DHC),
including the integration of CHP.
The countries to be covered are China, USA, Canada,
South Korea, Russia and some other selected
European countries outside the EU. The work is based
on both interviews of the key officers and specialists
and the existing laws, regulations and policies of each
selected country. The project will also provide
examples of best practices useful for sustainable
development of DHC as well as offer recommendations
to the countries to improve the institutional set up of the
DHC.
Regarding each country, the project will review, for
instance, the tariff setting, DHC related legislation,
taxation rules, price regulation, customer definition and
points of delivery; ownership of fixed assets; allocation
of CHP costs and environmental fees; social
considerations; municipal heat planning; and, heat
metering and control.
The project here is a twin project to EcoHeat4EU that is
a thorough analysis of the barriers and opportunities of
DHC as well but in the selected EU member countries.
INTRODUCTION
There is no reliable statistics of DHC in most of the
subject countries. The countries are in different stages
of DHC development, as can be read out in the paper.
The market drivers and barriers are different as well.
The aim of the study is to identify lessons learned from
all countries, including the EU that might be useful to
boost DHC development in the particular subject
country. Nevertheless, the lessons learned and
recommendations will be developed in fall 2010, after
the Symposium, and the final and complete study will
be available in May 2011. Therefore, all information
presented in the paper regarding four countries,
Canada, China, Ukraine and USA is based on the
preliminary survey that will be finalized by
October 2010.
PRELIMINARY COUNTRY SPECIFIC SURVEYS
1. Canada
1.1. Status of DHC
The old DH systems before 1985 are predominantly
with steam, whereas water systems have been built
since 1985. Both domestic hot water (DHW) and space
heating (SH) have been included. Based on
water/steam carrier, various combinations of heating
and cooling are available in Canada.
Historically, Canada has had the highest per capita
energy use of the developed countries, as a result of
the harsh climate and relatively low-cost, abundant
energy. So the benefits of DHC would be particularly
welcome to save energy. In Canada, there are records
of some 120-160 DHC systems in the country, and
almost a half of them located in Ontario Province alone.
About 27 Mm
2
of residential, industrial and institutional
floor area are connected to the DHC systems. This
represents about 1,3% of all floor space in Canada.
The largest DHC system is in Toronto with 522 MW
thermal capacity.[1]
Natural gas distribution has spread everywhere, which
is a challenge for DHC expansion. Moreover, at
relatively low electricity prices, there is a little market
for CHP. No economic market for CHP exists in
Canada unless the feed-in tariff is in place or the
electricity is used in-house of producer. Power and gas
utilities have not been co-operating so far, because
there has not been any incentive to such co-operation.
Because of the structure of the provincial utilities and
low electricity prices, only a few CHP based DHC
systems are in operation.
The utilities are empowered to provide the people with
gas and electricity at the lowest costs possible.
Economic drivers support the selection of the proper
technologies, and the provincial regulators ensure that
the system availability and safety are maintained at all
times. Provincial governments provide some directions
to the energy industry, but limit themselves to setting
overall goals only. The selection of the technologies is
left to the utilities. Natural gas is widely available
throughout the country, which is a challenge for other
heating modes to enter the market. Serious lack of gas
reserves is expected in the future, which means
alternative energy sources to become increasingly
September 5
th
to September 7
th

216

realistic. To substitute natural gas, DHC based on
biomass and possibly with CHP is a superior option.
For DH, two-tier tariffs are used in which energy fee is
pass-through of energy costs, and the fixed fee covers
the profit, the connection costs and all other cost
except energy. The fixed fee can be adjusted
annually/biannually with CPI (Consumer Price
Index).The customer contracts are made for a long
period, say 10-20 years, during which the capital cost
have been discounted to the fixed fee. Municipal
companies operate as non-profit but private companies
with reasonable profit.
1.2. Market Drivers
In Canada, the federal government is committed to
reducing GHG emissions by 17% below 2005 levels
by 2020, being the main driver of DHC. The DHC
market is expanding smoothly to start creating a
different infrastructure to substitute depleting resources
of natural gas.
As mental drivers, there is strong interest in
municipalities to consider DHC introduction and further
expansion very much based on European practise.
Many municipalities have set voluntarily targets to the
reduced GHG emissions. DHC systems are widely
recognized as a potential measure to achieve the
targets. The DHC is considered a tool for the urban
planners but not an energy issue per se.
As an example of investment support, Ontario Power
Authority (OPA) subsidizes investments in electricity
savings by paying up to $800/kW of the saved electric
capacity. The subsidy used to be 400/kW, but was
doubled at the end of 2009. Customers can use that
money as the partial payment of the connection costs
of DHC, thus DHC companies indirectly benefitting
from the subsidy system as well.
1.3. Main Barriers
There is no formal DHC strategy or policy supporting
DHC and CHP development in Canada. The
Government does neither have the tradition nor the
willingness to take strong position in DHC
development. The private sector that could bring
investments and entrepreneurship cannot be much
interested, because starting the DHC is risky: long pay-
back times ranging beyond 10 years, limited access to
municipal property, challenging contracting of
residential, municipal and federal buildings, overall
billing and collection of different types of customers.
Nevertheless, the municipalities are rather weak,
because the municipal taxation only covers property
and tourism taxes but no corporate or income taxes.
Moreover, municipalities have no mandate on energy.
The federal government hesitates to take a strong role
while fearing of intervening the private sector driven
heating market.
1.4. Current Activities
The Integrated Community Energy Solutions (ICES)
Roundtables have been established to accelerate
progress toward reducing GHG emissions by bringing
together senior-level stakeholders to exchange views
on the best way forward from here. The Roundtables
build upon ICES. The Roadmap for Action, which was
released by the Canadian Council of Energy Ministers
at its annual meeting in September 2009, describes the
role that Canada's federal, provincial and territorial
governments can play in advancing ICES and it sets
out a broad strategy for action. It also includes a variety
of options from which the governments can choose,
according to their priorities, to advance community
energy performance and complement existing energy
efficiency activities in different sectors.
The ongoing collaboration of key energy actors and
enablers across Canada from the private and public
sectors through the Quality Urban Energy Systems of
Tomorrow (QUEST) collaborative also informed the
Roundtable discussion. In particular, preliminary results
from a QUEST-led study suggest that ICES could
reduce GHG emissions at the community level by as
much as 40% to 50%, resulting in reduction of 65 Mt
by 2020, which is about 20% of Canada's official
2020 target reductions. These results are very
promising and highlight how ICES could contribute
significantly to improving Canadas energy and GHG
performance.
2. P.R. China
2.1. Status of DHC
In China, the DH development has been very strong,
more than 10% annually during the past decade on
average. By the end of 2005, DH supply (including
steam and hot water) was over 2 100 PJ; of which CHP
accounted for 47% and boilers accounted for 51%.In
the supply of steam and hot water, steam supply is 715
PJ, of which CHP accounts for 81% and boilers
account for 17%; the total hot water heating supply is
1395 PJ, of which CHP accounts for 29% and boilers
account for 69%. The heating supplied by CHP units
and boilers are respectively 992 PJ and 1086 PJ.
Apart from Europe, only SH is supplied by the DH
systems, and the DHW by individual systems: solar
collectors, propane, electricity, etc. [2,3]
During the few years to come, China will become the
largest DH country in the world.
September 5
th
to September 7
th

217

2.2. Market Drivers
The rational of strong DH development in China is
based on eliminating the small and polluting coal fired
boilers in the northern, western and central provinces
and to provide feasible living conditions to the
population massively moving in to the cities.
DH has been encouraged by the Chinese government
for several decades. China's DH heating area has
increased from over 276 Mm
2
in 1991 to over
1100 Mm
2
in 2000, and exceeded 2500 Mm
2
in 2005,
with an annual growth rate of 17%. The growth in DH
mainly came from the northern and the northeast
regions. In China, residential buildings account for
about 70% of the total DH area and commercial
buildings the balance of about 30%.
The urban communities are very densely built, which
effectively supports centralized heating and cooling
solutions. The new buildings comprise about half of the
DH connections, whereas the balance for existing
buildings, the latter previously having had been heated
by small coal boilers.
2.3. Main Barriers
The DHC sector is expanding fast but there are still
some barriers regarding economy, policy, financing and
technology as summarized below.
ECONOMIC AND PRICING BARRIERS
In order to become cost-effective and an attractive
investment, power and heating reform policies will need
to be undertaken. Some of the key issues include:
- Energy price policy reform is a priority. At present,
in China, the coal price is based on the market,
which has grown rapidly in recent years. However,
electricity and heating prices are still controlled by
the government, and have only slightly increased.
While the government has provided limited
subsidies to DH companies, most CHP enterprises
and DH companies are currently not making a
profit as a result of the lack of energy price reform.
- In addition, heating reform needs to be further
developed. Currently, in most cases, heat tariffs
are based on the building area, rather than on the
actual heat consumption, which has a negative
influence on improving the energy efficiency in
district heat facilities and buildings.
- Power sector reform is also needed. At present,
the electricity produced by most DHC (and some
CHP) projects cannot interconnect with the power
grid, which has strongly reduced development. The
technical issues of grid connection can likely be
addressed. However, there are also administrative
interconnection issues, such as added-capacity
charges and power grid balancing that need to be
addressed. At present, the State Power Grid Group
is responsible for the power grid operation. As
such, more communication and coordination
activities could be conducted between the DHC
industries and the State Power Grid Group.
- Centralized DHW would benefit CHP. Missing
DHW load hampers economic development of
CHP schemes. Without DHW, the CHP plants can
operate all year round only if there is industrial
steam load existing nearby.
POLICY BARRIERS
There also exist barriers in the area of economic
support and administrative policies related to
CHP/DHC, including:
- There is a lack of monitoring and enforcement of
the governments policies related to the efficient
operation of CHP projects. Currently, it appears
that some newly- built CHP projects are operating
only in thermal generation mode after they have
been approved, thereby reducing their energy
efficiency.
- There is a lack of targeted policy for smaller CHP
units. In order to fulfill the energy conservation
target, China is attempting to increase the number
of more efficient large power generation plants and
to close down smaller, older units. While it is
important that the smaller, more inefficient units be
closed down, some small CHP units with high
efficiency are also being targeted for phase-out.
Based on the goal of increasing energy supply
efficiency, a different policy should be adopted. For
example, in regions with low heating loads, small
CHP units could provide most of their energy
needs at a fraction of the cost of larger units.
FINANCING BARRIERS
There are promising energy conservation projects
particularly in the DH sector that could be realized if
there were sufficient funds or other means available to
address the gap in investment capital. In particular:
- Some planned CHP/DHC projects are not operated
efficiently because they lack sufficient resources to
invest in expanded heat pipeline infrastructure.
Further, at many existing DHC projects, the heat
loss in pipelines is high, reducing the overall
efficiency of the heating system. Additional
financing is needed to invest in cost-effective heat
pipeline retrofit projects, which will generate
sizeable energy efficiency benefits and GHG
reductions.
September 5
th
to September 7
th

218

- While energy service companies are expanding in
the commercial building energy conservation
arena, they have not yet entered the CHP/DHC
area. There is some room for these types of third-
party players to come up with innovative means to
finance projects.
TECHNICAL BARRIERS
While CHP/DHC are proven, existing technologies that
do not require major research and development, there
are some advanced technologies that could be
introduced from IEA Member Countries to improve
efficiency and operational benefits. In addition, there is
currently some debate about the relative merits of DC
technology. China-specific research studies could be
conducted to confirm the primary energy conservation
performance of these technologies.
ORGANIZATIONAL BARRIERS
There are some organizational barriers for optimal
development as well.
- Scattered organizations with several heat suppliers
and distributors prevail in one city. In the same DH
system, the heat supplier is responsible for
operation and maintenance until the group
substations that serve several buildings through
the secondary network, and the distributors being
responsible from the substations to the indoor
heating elements. Therefore, the holistic
optimization can be often compromised by partial
optimizations.
- The DHC companies are operation and
maintenance companies only, whereas investment
decisions and financing depends on the municipal
and provincial budgets. This is one more reason for
that there is little business minded atmosphere in
the extensively staffed DHC companies.
The DH systems are expanding fast in China,
simultaneously restricting coal consumption and
reducing overall GHG emissions of the heating
services.
The Ministry of Construction has issued the Housing
and Building Reform on Energy Efficiency (HRBEE),
which requires more efficient buildings to be built as
well as introduction of heat metering and consumption
based billing. The first consumption based billing pilot
was initiated in Tianjin a few years ago with a two-tier
heating tariff. Such billing systems are slowly
expanding to other regions.
3. Ukraine
3.1. Status of DHC
Ukraine is one of the largest DH countries in Europe.
Currently almost 80% of urban housing is supplied with
DH through extensive grids of hot water pipes.
The DH sector is rather saturated, but in some eastern
(Donbas) cities the DH systems are deteriorating fast,
and customers are either adopting apartment level gas
boilers or even remain without heating, thus enjoying
on the heat losses penetrating to them through walls
free of charge from their heated neighbours. Even the
municipalities are offering investment subsidies to the
apartment owners to purchase apartment level gas
boilers while disconnecting the DH services.
Such practices have led to extremely poor quality of
DH services: low water and room temperatures a well
as periodical heating are used to minimize fuel costs.
There are coal (and anthracite) mines in Ukraine, but
little used for providing fuel for DH: Most DH is based
on natural gas imported from Russia. The costs of gas
comprise 5070% of the DH, which explains why the
DH is vulnerable to gas price changes.
Ukrainian heat generating facilities are ineffective for
many reasons. The most important reasons are as
follows:
- technology used for heat generation is outdated
and inefficient;
- key assets are heavily deteriorated;
- equipment is being used in a switching mode on
unspecified fuel;
- delays and failures to carry out regular repairs.
According to the Ministry of Fuel and Energy, more
than 90% of energy units have worked out their
projected service life (100 000 hours), more than 60%
have been in service longer than 200 000 hours.
Heat tariff for final consumers is defined as a sum of
tariffs for production, transportation and supply.
Tariffs for heat that is produced by CHPs, co-
generation or alternative/renewable energy sources are
set by the National Energy Regulatory Commission
(NERC) but they should not be higher than heat
produced by other sources.
Tariffs for heat production, transportation and supply
other than CHPs, co-generation or alternative
/renewable energy sources are approved by local
governments. Due to that the tariffs differ much across
the territory of Ukraine.
According to the Law of Ukraine On Heat Supply,
heat tariffs should cover all the economically sound
expenses for heat production, transportation and
supply. Tariffs should include full costs of heat
September 5
th
to September 7
th

219

production and provide for marginal profitability level
that is not lower than the level defined by the Cabinet
of Ministers on the base of calculations by the central
body of executive power in heat supply.
If heat tariffs do not cover the cost of heat and marginal
profitability level, the body that has set the tariff should
provide for the compensation according to effective
legislation. That is, if the tariffs for heat from thermal
power station and boilers that are approved by the local
government on the basis of heat producer calculation,
and they are lower than economically sound cost
including marginal profitability level, the local
governments must compensate the losses from the
local budgets.
Meanwhile, the Ministry of Economy elaborated the
draft that specifies binding of the household services
tariffs to energy prices. First of all, it means heat, hot
water and gas supply to households. The current
system of tariff setting reduces the competitiveness of
Ukrainian industry, since industry is forced to
compensate for low households tariffs.
The procedure to raise the heat tariffs is rather
complicated and time consuming, as follows:
1) The district heat supply company receives official
notification from NERC on gas price increase. Only
after that the company may start developing the
proposal on the heat tariffs increase.
2) The new heat tariffs have to be approved by the
following authorities: Commissions of the Municipal
Council (miskrada) and regional council (oblrada). The
tariff proposal has to be reviewed by several instances
as listed below:
- Trade unions
- Antimonopoly Committee
- Department for Price Administration
- Department for the Protection of Consumer Rights
- Public hearings
3) Municipal Executive Committee (miskvykonkom)
has to approve the new heat tariffs as well.
4) The tariff changes shall be publicized via official
mass media of Municipal or Regional Council. If during
a month there are no official protests from the Office of
Public Prosecutor, the company is entitled to apply the
new tariffs.
The above steps clearly show how cumbersome any
tariff increase can be in practise.
3.2. Market Drivers
Article 54 of the state budget of Ukraine for 2006 and
the Cabinet of Ministers Decree No.207 of 9 March,
2006 stipulate for subsidies from the state budget to
local budgets. No less than 75% of the subsidy must be
directed to energy saving in heat, water supply and
sewerage. But according to monitoring results, the
funds are allocated to other purposes. Only four
regions used the funds for energy saving. Other
regions used from 7% to 40% instead of the required
75% to energy saving.
Other measures of energy saving that would be
appropriate include:
- replacement or reconstruction of steam and gas
boilers with efficiency that is lower than 89%;
- improvement of heat pipes insulation to decrease
losses in transmission pipelines;
- installation of heat meters; and,
- Installation of co-generation equipment.
Another stimulus for companies to introduce energy
saving technologies is outlined in the Law of Ukraine
On Heat Supply. According to the Article 8, in case
heat supply or heat transportation companies introduce
energy saving measures that result in saving of heat
losses, the body of executive power, that is entitled to
regulate heat tariffs according to the Law, may leave
the tariffs unchanged for the three consecutive
years.[4]
3.3. Main Barriers
In general, there are a number of decent laws and
regulations that would support DHC development, but
they are not implemented properly, as mentioned
above already.
Therefore, there is little if any incentives to business
oriented development of the heating services, but the
systems are run at minimum investments and reduced
technical performance. The DH companies are solely
operation organizations, mainly departments of the
municipality. The municipalities take care of billing and
collecting based on subsidized lump sum tariffs, and on
investment decisions.
There are many privileged customer categories that
enjoy reduced costs of DH services. In Odessa, for
instance, 25% of the customers in year 2006 enjoyed
such privileged heating prices. Their billings were
decreased by 20, 30, 50, 75 or even 100%, which
effectively destroys the business opportunities of DH.
Individual and autonomous heating in every apartment
seems the most favourable option for consumers. In
such a case they do not pay for heat and hot water but
only for gas and cold water. In addition, they can
regulate temperature in their apartments and do not
suffer from overheating in spring and insufficient
heating in winter. But sometimes it is impossible to
install autonomous boilers in every apartment, because
there is not enough space for heating equipment and
the vertical ventilation ducts are not designed for flue
gases. Therefore, it would be appropriate to install one
September 5
th
to September 7
th

220

boiler for the whole building (several apartments) or
several buildings. Another problem for individual and
autonomous heating is that in case of gas supply
interruption there is no reserve fuel resources to
continue heating. Reserve fuel can be provided only for
centralized DH.
Frequent failures in the heating systems as a result of
outdated equipment and poor funding are still common
throughout the country. Some service breaks in coldest
winter times have caused serious impacts on human
life already.
Legally, local authorities that establish tariffs for
population lower than the cost coverage level have to
compensate the difference to energygenerating
companies. In practice the compensation is not always
paid in full which leads to arrears accumulation and
aggravates financial state of heatgenerators. The
procedure of heat tariffs increase is rather complicated,
as well as time consuming.
According to the Law of Ukraine adopted in April 2006,
heat producers such as CHPs and renewable sources
power plants are not allowed to crosssubsidy heat
production to cover losses from heat production at the
cost of electricity production or other activity.
Nevertheless, official sources say that due to low heat
tariffs for CHPs heat production is subsidized by the
cost of electricity production. But the unofficial sources
assert that CHPs may charge heat tariffs that are even
higher than heat production cost to cover losses from
electricity production, because electricity tariffs are set
only by NERC while heat tariffs are set by heat
production companies with the approval of local bodies
of power.
The DH strategy is under preparation in Ukraine as a
multi-ministerial approach and it should be ready in fall
2010. CHP development is in the focus of the strategy.
There has also been comprehensive framework
support initiated by USAID, EBRD and EU to
reformulate the national energy policy, including DHC
and CHP. It is uncertain now how much the political
election of April 2010 will influence availability of such
foreign technical assistance in the years to come.
4. U.S.A.
4.1. Status of DHC
The total DHC industry base comprises approximately
2 500 systems, in which the number of customer
buildings served by a typical DHC system may range
from as few as 3 or 4 in the early stages of new system
development to the largest system served by
Consolidated Edison in Manhattan. The downtown
DHC/CHP system in New York City is the Worlds
largest steam system with 1850+ customers.
DHC (primarily DH currently) delivers about 3,5 % of
the total final energy demand in the industrial,
residential, public, and commercial sectors. In the past
two decades, some 47 Mm
2
has been connected to the
DHC systems, but the total customer base volume
number is not available.
The DHC systems are predominantly (80%) with
steam, the consumption being a mixture of steam
heating, cooling and DHW depending on the particular
case. There is little residential heat load but the
majority is public: offices, malls, universities and
military bases.
Countrywide, the DH and DC markets are expanding at
3-4%/a and up to 10%/a, respectively, but almost solely
on campuses, hospitals, military bases and in the
downtown commercial and public buildings.[5]
In general, however, DHC together with CHP has been
tragically underutilized as a tool to combat climate
change, to reduce life-cycle costs of energy supply and
to defend energy independence in U.S.A.
4.2. Market Drivers
The U.S. Congress has acknowledged the benefits
DHC/CHP by stating that:
- approximately 30% of the total quantity of energy
consumed in the United States is used to provide
thermal energy heating and cooling building
space, DHW and industrial processes;
- thermal energy is an essential, but often
overlooked segment of the national energy mix;
- DHC systems provide sustainable thermal energy
infrastructure by producing and distributing thermal
energy from CHP, sources of industrial or
municipal surplus heat and from renewable
sources such as biomass, geothermal, and solar;
- DHC systems provide advantages that support
secure, affordable, renewable, and sustainable
energy for the U.S., including use of local fuels or
waste heat sources that keep jobs and energy
dollars in local economies, stable, predictable
energy costs for businesses and industry,
reduction in reliance on fossil fuels, reduction in
emissions of GHG, and flexibility to modify fuel
sources in response to future changes in fuel
availabilities and prices and development of new
technologies;
- DHC helps cut peak power demand and reduce
power transmission and distribution system
constraints; and,
CHP systems increase energy efficiency of power
plants by capturing thermal energy and using the
thermal energy to provide heating and cooling, more
September 5
th
to September 7
th

221

than doubling the efficiency of conventional power
plants.
The Department of Energy has estimated that
increasing CHP from its current 9% share of U.S.
electric power to 20% by 2030 would avoid 60% of the
projected increase in U.S. carbon dioxide emissions
(equivalent to taking half of all U.S. passenger vehicles
off the road); and, generate $234 billion in new
investments.
DHC would be a critical component of this CHP growth.
The local electric distribution companies (LDCs) are
interested in DHC as a means to reduce the summer
peak and to release transmission and distribution
capacity to other electric applications that have more
even consumption during the year.
The developers of the building sector are interested in
DHC as well, because it would leave more room space
in the building for sale.
At the municipal level, the market driver for DHC is the
reduction of the GHG emissions. Many municipalities
have set voluntarily targets to the reduced GHG
emissions.
4.3. Main Barriers
In general, the barriers are very much the same as
already discussed in Canada. Private sector as
investor cannot be much interested, because starting
the DHC is risky: long pay-back times ranging beyond
10 years, limited access to municipal property,
challenging contracting of residential, municipal and
federal buildings, overall billing and collection of
different types of customers.
Only little expansion on residential sector is
recognized, and that is because there is voting needed
among the condominium owners. The centralized
energy systems, that the condo owners are not familiar
with and perhaps difficult for them to understand the
benefits, have not been adopted on the residential
sector in a considerable scale so far.
There are several laws and regulations that are
expected to support DHC development in the
U.S.A.[6,7]
Rising interest on development and extension of
renewable energy sources as well as improving overall
energy efficiency is to be converted to legislation at the
moment. Unfortunately, DHC has not been successful
in the legislation process so far, but both the
Department of Energy as well as the DHC and CHP
associations such as IDEA and USCHPA are working
on it.
The definition of CHP is rather complicated. The
Internal Revenue Code 26 USC and its 48 define
CHP as producer of:
- at least 20 % of its total useful energy in the form
of thermal energy which is not used to produce
electrical or mechanical power (or combination
thereof), and
- at least 20% of its total useful energy in the form of
electrical or mechanical power (or combination
thereof), and
- the energy efficiency percentage of which exceeds
60%.
The Thermal Energy Efficiency Act of 2009 establishes
the Thermal Energy Efficiency Fund that would award
grants for DHC, CHP, and recoverable waste energy
projects. It includes biomass facilities. Under a federal
GHG emissions regulation program, 2% of emission
allowances established for each calendar year from
20122050 would be allocated to the Fund.
This legislation would dedicate 2% of revenues from
climate change legislation to fund CHP, waste energy
recovery, and DHC projects. Based on various
estimates, this could mean roughly between $1 billion
and $1,5 billion per year for clean energy infrastructure.
The Thermal Energy Efficiency Act would provide 40%
of its funding for institutional entities (defined as public
or non-profit hospitals, local and state governments,
school districts and higher education facilities, tribal
governments, municipal utilities, or their designees),
40% for commercial and industrial entities, and 20% to
be used in the discretion of the Secretary of Energy to
fund institutional entity projects, commercial and
industrial projects, or federal facility projects. A match
is required of all non-federal applicants, starting at 25%
from 2012-2017, and rising to 50% from 2018 to 2050.
The breakdown of how the money would be used is
75% for construction of infrastructure, 15% for
planning, engineering, and feasibility studies, and the
remaining 10% to be used at the discretion of the
Secretary for either infrastructure or planning,
depending on the need.
In competition with grid power plants receiving
generous allowances in ACES, CHP systems could be
shut down. Unless allowances are allocated to the
DHC CHP system, it will have to purchase allowances
for all gas consumed in the facility, resulting in an
additional cost equal to 15% of the average 2007
wholesale power price ($57 per MWh) at the $16 per
metric ton allowance price projected by Environmental
Protection Agency (EPA) for the year 2020. In contrast,
the merchant coal plant will only have a GHG
allowance cost of only 5% of the average 2007
wholesale power price, because allowances will be
allocated for nearly all (83%) of its emissions.
September 5
th
to September 7
th

222

Faced with this huge competitive disadvantage in the
marginal cost of power generation, some existing CHP
facilities will shut down and construction of new CHP
plants will be choked off.
In the ACES, DHC systems are not directly covered
entities unless they qualify as electricity sources.
District systems would be covered indirectly through
the costs of allowances built into the prices of
purchased fuel oil, or natural gas if purchased from the
gas LDC. However, gas purchased on the wholesale
market or coal users not qualifying as an electricity
source would not be required to submit allowances.
This is a fundamentally good framework with the
exception of the concerns about CHP systems to be
covered or not. However, if the upcoming legislative
process results in modifications that make many DHC
systems covered entities, it is critical that changes in
allowance allocations be made as discussed below.
For example, if the final climate change bill regulates all
sources with emissions greater than 25 000 metric tons
CO2e (the threshold generally used in the ACES as
well as a number of past bills), over 70% of DHC
systems and over 95% of DHC output would be
capped. In such a way, more efficient systems will have
competitive advantage, because the quantity of
allowances needed per unit of energy will be lower.
American Clean Energy Leadership Act (ACELA) in
June 2009 and Federal Renewable/Energy Efficiency
Standard establishes a Renewable Electricity Standard
which includes provision for energy efficiency credits as
well as renewable energy credits that can benefit DHC
as well.
Renewable Electricity and Energy Efficiency Standard
established by ACELA is applicable with the electric
utilities selling >4 TWh a year. The utilities are required
to supply 20% of demand from combination of
renewable sources and increased energy efficiency
meaning 15% renewable together with 5% efficiency
increase. If the state determines that it cannot meet the
renewable requirement, then the portion of renewable
sources may fall lower to 12% but with efficiency
increase equal or higher than 8%. These requirements
provide important leverage for DHC/CHP development.
Energy and Water Development Appropriations Act of
2010 will provide $15 M for DHC feasibility studies.
CONCLUSION
The survey work is still underway, and therefore, the
lessons learned and recommendations will be issued in
the final report in spring 2011.
ACKNOWLEDGEMENT
The authors express their gratitude to the interviewed
specialists, Mr. B. Gilmour (Canadian Urban Institute),
Mr. K. Church (Natural Resources Canada), Mr. M.
Wiggin (Public Works and Government Services
Canada), Mr. R. Thornton (International District Energy
Association IDEA), Mr. D. Kaempf and Ms. P.
Garland (U.S. Department of Energy), Mr. B. Hedman
(ICF International) and Messrs. G. Draugelis and P.
Salminen in the World Bank.
REERENCES
[1] National DHC Survey, Canadian DHC Association
(CDEA), 2009.
[2] Ministry of Construction, China City Construction
Statistic Annual. The DH data does not include
industrial steam and hot water.
[3] T. Kerr, IEA Collateral, Sustainable Energy in
China: The Role of CHP and District
Heating/Cooling, 2008.
[4] A. Tsarenko, Overview of Heating Sector, CASE
Ukraine, 2007.
[5] IDEA Report, The DHC Industry, 2005.
[6] DHC Services, Commercial Data Analysis for EIAs
National Energy Modeling System, Energy and
Environmental Analysis, Inc. and International DHC
Association, 2007.
[7] M. Spurr, Climate Change Legislation in Dollars
and Cents, presentation in IEA DHC in
Gustavelund, Finland, in Aug 2009.
September 5
th
to September 7
th

223

BARRIERS TO DISTRICT HEATING DEVELOPMENT
IN SOME EUROPEAN COUNTRIES
Dag Henning
1
and Olle Mrdsj
2
1
Optensys Energianalys, rng 8c, SE-582 39 Linkping, Sweden, phone +46 70 536 59 22,
e-mail dag.henning@optensys.se, www.optensys.se
2
Manergy, P.O. Box 271, SE-581 02 Linkping, www.manergy.se

ABSTRACT
District heating (DH) offers low primary energy
demand, high security of supply and small CO
2

emissions. Barriers to DH in the UK, Ireland, France,
Romania and the Czech Republic have been compiled
through publications and interviews.
DH systems require large investments, have negative
initial cash flow and long payback time, which obstructs
financing. One actor should control DH from source to
consumption. If the value chain is fragmented,
contracts are required between the links. It increases
risks and financing costs, like in the UK and Ireland,
where DH is not established. There are few multi-family
houses with central heating and it is expensive to build
DH networks in built areas.
Most French DH systems are operated according to
long-term concessions by companies that sell electricity
and gas. No strong actor provides unbiased DH
support. In the Czech Republic, gas offers DH severe
competition. Much DH is produced at the expense of
electricity that is considered more valuable, and waste
incineration is not popular. In Romania, DH
consumption was reduced by one-half. Distribution
losses are enormous. New less polluting plants are
needed.
Consortia from established DH countries could offer DH
systems from fuel to customer if local policies facilitate
DH development.
INTRODUCTION
This paper describes barriers to district heating (DH) in
various parts of Europe and to Swedish involvement in
district-heating business abroad. The paper is based
on a report called District Heating in Europe: Barriers
to overcome for Swedish export [1], which was
prepared for The Swedish District Heating Association.
The losses by energy conversion in Europe are of the
same magnitude as the European heat demand and
consist mainly of heat that is wasted by electricity
generation [2]. District heating is a means to utilise
such surplus heat to cover heat demand.
District heating can utilise the heat from electricity
generation in combined heat and power (CHP) plants.
District heating can also use other heat sources that
are difficult to use for individual buildings, such as
unrefined biomass fuels, heat from waste incineration
and industrial surplus heat. The latter may, for
example, be a by-product from production of
automotive biofuel. District heating can provide cheap
energy to consumers by using low-cost energy
sources, such as wood, waste and surplus heat. Many
of these resources can be of local origin and promote
local business and industry.
The main advantages with district heating are high
security of supply through utilisation of domestic
renewable energy resources, if available, low primary
energy demand due to high conversion efficiency, as
well as small CO
2
emissions thanks to low fossil fuel
use and the high energy efficiency. Incineration of
waste with heat recovery to district heating may be
used at very low cost. District heating also gives
opportunity for cogeneration of power and heat with
very high efficiency. District heating enables profitable
heat supply with outstanding environmental
performance but there are in many places various
barriers to a prosperous DH development.
Barriers to district heating in the United Kingdom (UK),
Ireland, France, Romania and the Czech Republic, as
well as barriers to export of Swedish district heating
knowledge and products to these countries have been
compiled from publications and through personal
communication with people in public and private energy
bodies and companies in Sweden and abroad [1].
In the studied countries, there are large potentials for
district-heating development and for Swedish sales of
DH related goods and services. But for district heating
and export to succeed, there are several barriers to
overcome in Sweden as well as in the other countries.
It should be emphasised that this paper focuses
barriers and does not give the full picture of the
conditions for district heating, which also includes many
possibilities.
BARRIERS IN WELL-DEVELOPED DH COUNTRIES
In many countries with well-developed district-heating
industry, such as Sweden, much DH competence
resides in municipally owned energy companies. They
have system knowledge, which could be applicable in
other countries. District-heating companies owned by
September 5
th
to September 7
th

224

Swedish municipalities must, for judicial reasons, limit
their business abroad to sales of services, and to a
very limited extent goods. For municipal district heating
companies, domestic judicial restrictions are the first
barriers to overcome before operations in other
countries can commence.
Only certain components for production and distribution
of district heating are manufactured in a single country,
which calls for international cooperation. The Swedish
Government provides certain but limited support to
promotion of district heating business abroad. For
example, Swedish district heating consultants work
abroad but it is seldom followed by goods export.
FINANCING THE DH VALUE CHAIN
Financing is a large barrier to district heating
development. DH systems require large investments
and may have long payback times. The cash flow is
negative for a long time during the establishment of a
new DH system. Time horizons are distant, which
stresses financers in our present situation of rapidly
changing conditions. Private companies often focus on
short-term profit and public involvement may be
necessary for the deployment, modernisation and long-
term development of district heating systems.
District heating is a comprehensive concept for heat
from source to consumption. Its strength lies in
maintaining the value chain (Fig. 1). This may fit badly
in an exaggerated market context where every little link
of the value chain is organised separately with an
interface of costs and revenues to other links. A
fragmented value chain increases interface costs and
total risk. EU regulations have a tendency to promote
such fragmentation. Between the links of a fragmented
supply value chain, many complicated agreements are
required, which all include risks. It means a larger total
financing risk, which raises interest rates and shortens
amortisation periods for loans. This implies a mismatch
with the depreciation in the balance sheet due to the
long economical lifetime of district heating versus the
short amortisation time.
TWO GENERAL DH BARRIERS
Two general district-heating barriers are related to CO
2

emissions and the attempts to reduce these through,
for example, reduced energy use. Global warming and
better insulated houses reduce heating demand and,
hence, the advantages of district heating because
investment costs must be carried by less supplied heat.
Another general barrier to district heating is the EU
emission trading scheme, which favours individual
heating because individual CO
2
emissions do not need
allowances.
TYPES OF DH BARRIERS
In the countries analysed in this project, the barriers
are of very diverse nature. The obstacles are
dominated by difficulties for district heating itself rather
than for foreign companies operations in the countries.
In the British Isles, it is largely a question of
establishing district heating as a natural element in
society. In France, it is about large domestic companies
that may offer superior competition to foreign firms. In
the Czech Republic, French and other companies from
abroad dominate the DH business but the technical
design of district-heating production may hamper DH
development. In Romania, there are several problems
with facilities in bad shape and public bodies that have
not addressed the issues properly.
Table I is an attempt to assess how large the various
barriers are in the studied countries. The table starts
with some general conditions. Ownership and
organisation considers if district-heating companies are
owned, or DH operations are organised, in ways that
make it more difficult for Swedish companies to do
business. Corruption may be a problem through, for
example, indirect bribes by procurement. National and
local control encompasses national laws and policy
instruments that are disadvantageous for district
heating, DH price regulations, as well as municipalities
not facilitating district heating by planning of new
developments. But rules complicating combined heat
and power production are included in the CHP line in
Table I.
Financing is one of the largest barriers to district
heating, primarily because DH schemes give a low rate
of return. A fragmented value chain cause contract
risks at several instances. Entrance barriers for foreign
companies in Table I consider additional difficulties for
foreign firms besides the other parameters and the
general disadvantage of not being familiar with the
domestic business culture.
Some parameters in Table I are related to district-
heating sales. DH competitiveness includes the
availability and price of other forms of heating, primarily
natural gas. Customer relations concern customer
attitudes toward district heating, customers and
suppliers perceived insecurity whether they can

Fig. 1. District heating value chain with heat production, distribution and sales in focus [1]
September 5
th
to September 7
th

225

establish and maintain relations, as well as if
disconnections have occurred or may occur. Built
environment relates to how common multi-family
buildings are and if these have a central heating
system for the whole house. Table I ends with district-
heating production and distribution issues. Biomass
considers domestic biomass supplies and
infrastructure for biomass fuel supply. Waste includes
current waste management and attitudes toward
waste incineration. CHP concerns regulations
hampering CHP production as well as problems in
existing plants. Finally, district heating distribution in
Table I encompasses difficulties with building networks
and deficiencies in existing distribution.
The assessments in Table I were primarily made within
each country and secondly countries were compared but
mostly the ranking of countries for a parameter is
appropriate. However, every grade has a certain width
and two countries with the same digit may differ. As an
example, district heating is assessed to be somewhat
less competitive in Romania than in the Czech Republic.
It follows a description of barriers in the individual
countries emphasising the largest barriers.
Table I. Height of DH barriers in analysed countries [1]
BARRIER UK IRELAND FRANCE CZECH REPUBLIC ROMANIA
Ownership and organisation 1 0 4 2 3
Corruption 0 0 0 2 3
National and local control 3 2 1 2
Financing 4 3 2 3 3
Fragmented value chain 4 3 1 2 1
Entrance barrier for foreign companies 1 1 4 2 2
DH competitiveness 2 1 3 4 4
Customer relations 2 2 1 4
Built environment 3 4 2 0 0
Biomass 3 3 1 3 1
Waste 1 1 3 4 2
CHP 3 3 2 4 4
DH distribution 4 4 1 4

THE BRITISH ISLES
In the United Kingdom (UK), and even more in Ireland,
district heating is not really an established
phenomenon. Figure 2 shows that residences mostly
are heated with gas in the UK, often through a gas
boiler for the individual household. Oil is the most
common fuel in Irish homes but gas is expanding.
The largest problem is district heating distribution
(Table I). It is expensive and complicated to build DH
networks in already built areas and, at least in the UK,
it is not straightforward to obtain a licence for putting
district heating pipes into streets. The financing
difficulties in the British Isles are primarily due to a
fragmented value chain with many contract issues that
need to be solved before a larger district heating
scheme can be deployed. British thinking is based on
competition and individual choices. A collective large
scale solution, such as district heating, may conflict
with principles and tradition. Another large barrier is the
built environment. Few people live in multi-family
houses in the UK and even fewer in Ireland [3], and
even these buildings often lack central heating, but
individual heating of apartments is common. Biomass
is rated as a rather large barrier in Table I because
supplies are limited in the British Isles and fuel supply
systems are less developed.
UK Government and municipalities have hitherto not
facilitated district-heating development sufficiently and
strong incentives for deploying district heating systems
are lacking. Heating is generally not regarded as a
public concern, but as a concern for each individual.
National and local control is therefore indicated as a
rather large barrier in Table I. In Ireland, the situation
seems to be slightly better but in both countries certain
regulations, designed with electricity and gas in mind,
are disadvantageous for district heating. CHP suffers
especially from rules on how produced heat and power
may be supplied.
Customer relations are complicated because district
heating is a rather unknown energy form and there is a
certain resistance against collective solutions [3]. There
is a lack of standardised terms of contract for
connection to and delivery of district heating. Potential
heat suppliers and customers feel insecure concerning
how many users that will connect to a DH grid, for how
long they will stay and if heat supply may be
September 5
th
to September 7
th

226

interrupted. The competitiveness of district heating
compared to gas concerning availability and price is
considered as a medium severe barrier in the British
Isles (Table I).
Legend: Grade 4: Large barrier, Grade 3: , Grade 2: , Grade 1: Small barrier, Grade 0: Assessed not to be a barrier,
No grade: No assessment.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
UK Ireland France Czech Republic Romania
District heating Gas Biomass Peat Electricity Oil Coal
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
UK Ireland France Czech Republic Romania
District heating Gas Biomass Peat Electricity Oil Coal

Fig. 2. Heating of residences [1], [4][6]

FRANCE
Table I shows that one of the largest barriers in France
concerns the organisation of district-heating operations.
Most DH systems are managed by private French
companies according to long-term concessions [7]. The
companies have successfully applied this DH
management model in several other countries. By such
arrangements, it is important that operators have
incentives to make investments even if these have
payback times longer than the concession period [8]. It
is unclear if the French DH management model is
disadvantageous for district heating development but it
should anyway be a large barrier for foreign companies
wanting to enter the French market. In general,
domestic solutions are preferred. There is no strong
actor who provides unbiased support for district
heating. The dominating DH operators also sell
electricity and gas, which both cover a large fraction of
the heat demand (Fig. 2) and offer district heating
severe competition. Only ten percent of the apartments
and four percent of all residences have district heating
today, and DH expansion is slow [6].
Fig. 3 shows that one-half of the district heating in
France is produced with natural gas, mostly in CHP
plants. The main part of the renewable energy used for
district heating production is waste, which is used to a
slowly growing extent [7]. But French waste
incineration plants are mostly built far away from towns,
which makes it difficult to utilise the heat [6].
Natural
gas CHP
Natural
gas heat
Oil
Coal
Renew-
ables
Miscell-
aneous

Fig. 3. District heating production in France [9]
Financing is considered to be a smaller problem in
France. The market domination by a few actors may
present an indirect financial barrier. Quite a few people
live in apartments but most multi-family houses lack
central heating. The large French nuclear power
production is one reason for worse CHP conditions,
which is assessed as a medium-grade barrier (Table I).
THE CZECH REPUBLIC
Fig. 2 shows that district heating covers a substantial
part of residential heating in the Czech Republic, but
electricity is used to the same extent and gas is the
most common heat source. District heating covers one-
half of the apartments and 60% of urban heating [7].
A large barrier in the Czech Republic is, according to
Table I, the competitiveness of district heating.
September 5
th
to September 7
th

227

Gas prices make it difficult for gas-based district
heating to compete with individual gas heating [7].
There are some disconnections from DH systems.

Natural
gas
Hard
coal
Lignite
Miscell-
aneous

Fig. 4. District heating production in the Czech Republic
[10]

Domestic coal dominates Czech district heating
production (Fig. 4). Most of the district heating is
produced in CHP plants. The problem concerning CHP
(Table I) is that a large share of Czech district heating
comes from coal-fired power plants with extraction
turbines where the heat is produced at the expense of
electricity [7], which is considered more valuable. The
benefit of this CHP production is not allocated to the
heat [8]. Some biomass is used to produce district
heating, but biomass use is complicated due to
deficient fuel supply systems [7] and government
scepticism toward renewable energy. There is also
much resistance to waste incineration from the public
as well as from politicians.
Financing may be a rather large barrier, partly due to a
certain district heating disconnection tendency. The
many private foreign district-heating companies in the
Czech Republic [7] may be a difficult target for Swedish
and other district heating companies from abroad that
are not established in the country. There may also be
some reluctance toward foreign enterprises. A certain
barrier is the common corruption by public procurement
(Table I). The value chain is sometimes fragmented
into production and distribution run by different actors.
ROMANIA
In Romania, biomass covers the largest fraction of
residential heat demand among the countries under
study (Fig. 2). Individual boilers and stoves for wood
and gas cover more than one-half of the heat use in
households. Gas is the most widely used heating
source for residences and it is expanding at the
expense of district heating [7].
Table I shows that district heating has large problems
with competitiveness and customer relations. Today,
the DH consumption is just one-half of the previous
use. Many district heating users switched to gas due to
low gas prices and heavy, government-regulated DH
price increases [7], whereas households and district
heating plants had the same gas price.

Natural
gas
Oil
Coal

Fig. 5. District heating production in Romania in 2005 [7]

Fig. 5 shows that Romanian district heating production
is completely based on fossil fuels. One-half of the heat
is produced in, normally coal-fired, CHP plants. Large
investments are required in the Romanian district
heating systems. CHP plants and heat-only boilers
must be replaced for environmental reasons.
Distribution losses are enormous [7].
Organisation is a rather large obstacle for district
heating in Romania (Table I). The municipalities are
now mostly in charge of the district heating systems [7]
but much lobbying is required to achieve improvements
and it takes time to reach an investment decision.
Corruption is common. Some politicians and
employees try to make their own profit on DH business.
Financing difficulties largely concern insecurity whether
customers will remain because many have
disconnected from district heating. National and local
control is a certain barrier because DH companies
partly get heat production costs covered by central
Government and City Councils [7]. Besides the
mentioned problems, the entrance barrier for foreign
companies should be rather low. Waste collection and
sorting are now deficient but, on the other hand, new
possibilities should emerge when Romania wants to
introduce waste incineration, and waste is therefore
considered to be a medium-size barrier in Table I.
HOW TO OVERCOME BARRIERS
This paper focuses barriers and omits more positive
circumstances for district heating. It may be depressing
but the message is not that district heating has no
prospects. The report should rather be understood as a
realistic guide to DH development in the studied
countries.
To have a chance to overcome the outlined barriers to
any significant extent, powerful initiatives are required
from countries with established district-heating
September 5
th
to September 7
th

228

industries, such as Sweden. Initiatives should comprise
many different players, for example, district-heating
companies, equipment manufacturers, consultants and
governmental bodies. Such consortia could offer
district-heating systems from fuel supply, via heat
production plants and DH networks to customer
contracts. Now, many foreign groups visit municipal
district-heating systems in Sweden but these
opportunities are seldom utilised to sell a
comprehensive DH solution.
Municipally owned district heating companies have
system knowledge that can be applicable in other
countries. A competence transfer may be realised
through deeper involvement that might include
ownership of plants in other countries. Business
models should be developed, which allow utilisation of
municipal knowledge abroad and give municipalities
reasonable returns.
For a successful transfer of district-heating solutions
from established to emerging markets, private and
public companies must focus marketing on the
countries, places, projects and forms of involvement
that have the greatest expectations to succeed. At the
same time, national and local policies should reduce
and remove described barriers and facilitate district
heating development as a means for increased
efficiency of energy utilisation, higher security of supply
and decreased environmental impact.
CONCLUSION
There are several barriers to district heating
development in the countries under study. In the UK,
there are not many district heating systems. There are
few multi-family buildings with central heating in
Ireland. The long-term operating concessions of French
district heating systems might hamper their
development. In the Czech Republic, much district
heating is produced in extraction turbines at the
expense of more valuable electricity. Romanian district
heating use was reduced by one-half by cheap gas.
In general, it should be advantageous that one actor
controls the whole district-heating value chain from
source to consumption in order to utilise synergies and
to avoid economic risks with contracts between the
separate entities of a fragmented value chain. Like for
other long-term large-scale infrastructure investments,
public involvement may be necessary for district
heating development.
Through cooperation among various well-established
players in the district heating industry, knowledge,
products and services can be transferred to evolving
district heating markets, which promotes industrial
prosperity for all parties and helps building sustainable
energy systems in Europe.
ACKNOWLEDGEMENT
The Swedish District Heating Association and The
Swedish Energy Agency are gratefully acknowledged
for financing this study through the Fjrrsyn
programme. We would also like to thank everybody
who has contributed to the study with facts and
viewpoints.
REFERENCES
[1] D. Henning and O. Mrdsj, Fjrrvrme i Europa:
Hinder att vervinna fr svensk export, Rapport 2009:3,
Fjrrsyn, Svensk Fjrrvrme, Stockholm (2009)
http://www.svenskfjarrvarme.se/index.php3?use=biblo&
cmd=detailed&id=1440
[2] S. Werner, Ecoheatcool work package 4:
Possibilities with more district heating in Europe,
Euroheat, Brussels (2006)
www.euroheat.org/ecoheatcool
[3] WS Atkins Consultants Ltd, Assessment of the
Barriers and Opportunities Facing the Deployment of
District Heating in Ireland, Sustainable Energy Ireland,
Dublin (2002)
www.sei.ie/uploadedfiles/InfoCentre/DistrictHeatingRep
ortatk.pdf
[4] S. Werner, Ecoheatcool work package 1: The
European heat market, Euroheat, Brussels (2006)
www.euroheat.org/ecoheatcool
[5] SEI, Energy in Ireland: Key Statistics, Sustainable
Energy Ireland, Dublin (2008)
www.sei.ie/Publications/Statistics_Publications/EPSSU
_Publications/Energy_in_Ireland_Key_Statistics/Energ
y_in_Ireland_Key_Statistics_Final.pdf
[6] P. Cousinat, District Heating: A Tool for Rational
Heat Management, Master thesis 2006:21, Department
of Civil and Environmental Engineering, Chalmers,
Gothenburg (2006).
[7] Euroheat, District heating and Cooling country by
country 2007 survey, Euroheat, Brussels (2007).
[8] J. Zeman and S. Werner, District Heating System
Ownership Guide, DHCAN project, BRE, Watford
(2004) http://projects.bre.co.uk/DHCAN/guides.html
[9] SNCU, Les rseaux de chaleur et de froid: lnergie
citoyenne, SNCU, Paris (2004).
www.fg3e.fr/public/federation/syndicats/plaquettes.php
?root_page=6
[10] T. Zenaty, CHP/DH sector in the Czech Republic:
situation / problems / wishes, Energy Policy EHP
meeting, Budapest, 11 September 2008,
www.lsta.lt/files/seminarai/080911_Budapestas
/CZ.pdf

September 5
th
to September 7
th

229

IMPACT OF THE PRICE OF CO2 CERTIFICATES ON CHP AND
DISTRICT HEAT IN THE EU27
Markus Blesl
1

1
Institute of Energy Economics and the Rational Use of Energy (IER)
University of Stuttgart

INTRODUCTION
In the current energy and climate policy debate, one of
the key points is the discussion about emission
reduction targets and how they are spread among
different world regions or countries and also among
different sectors. To find a cost optimal burden sharing
of an emission reduction target, the different reduction
potentials of the particular sectors or technologies have
to be known. To reach a reduction target, emission
certificates in a country or region (like EU-27) are
allocated among the different sectors or between
different types of heat and power generation
technologies. This allocation (for example, auctioning)
of emission certificates is an important issue to
negotiate since the costs of buying certificates could be
an important factor in technology choices for
investment.
The significant advantage of this approach is that the
analysis of the different competing pathways to achieve
emission reductions also assesses how they influence
each other. In the context of efficiency improvement in
industrial CHP and district heating and cooling, the use
of waste heat becomes an interest field. Efficiency
improvements in the residential or commercial sector is
examined in the topic of energy saving. Without
analysing the entire energy system the possible
advantages of CHP and district heating and cooling
couldnt be taken into account. This shows the
difference to a standard cost potential curve approach,
which has a fixed order of measures depending on
their avoidance cost.
This analysis will evaluate the reduction potential of
CHP plants or in general the production of district
heating and cooling in the EU-27 using the energy
system model, TIMES PanEU /Blesl et al 2008; Blesl
2008; Blesl et al 2008b, Kuder Blesl 2009; Blesl 2009/.
TIMES PAN-EU MODEL
The energy system model, TIMES (The Integrated
Markal Efom System), is a further development of the
two model generators, MARKAL and EFOM-ENV,
written in GAMS. TIMES was developed in recent
years within the Energy Technology Systems Analysis
Programme(ETSAP) from the IEA with contribution
from the IER. It is classified in one category with the
models MARKAL, EFOM or MESSAGE. The model
generator, TIMES, was developed in the general
modelling language of GAMS due to reasons of being
better transferable. TIMES is a multi-periodic linear
optimization model based on a technical approach at
which single plants are aggregated. The purpose is the
evaluation of the economically optimal energy supply
structure at a given need of end use energy and energy
services and also at given energy and climate policy
requirements. For this, the discounted system costs are
minimized, whereas the single players (industry,
supply, households) could have different economic
considerations. The main objective of the model
development of TIMES is the flexible structure to
ensure a simple mathematic adjustment to the
respective problem.
The pan European TIMES energy system model
(abbreviated as TIMES PanEU) is a model of 30
regions which contains all the countries of EU-27 as
well as Switzerland, Norway and Iceland. The objective
function of the model is a minimization of the total
discounted system costs over the time horizon from
2000 to 2050. A perfect competition among different
technologies and paths of energy conversion is
assumed in the model. The TIMES PanEU model
covers on a country level all sectors connected to
energy supply and demand such as the supply of
resources, the public and industrial generation of
electricity and heat and the industrial, commercial,
household and transport sectors. Both greenhouse gas
emissions (CO2, CH4, N2O) and pollutant emissions
(CO, NOx, SO2, NMVOC, PM10, PM2.5) are covered
by TIMES PanEU.
The transport sector is disaggregated into four areas:
road transport, rail traffic, inland shipping and.aviation.
The road traffic includes five demand categories for
passenger transportation (car short distance, car long
distance, bus, coach, motor bikes) and one for freight
service (trucks). The rail traffic includes three
categories: rail passenger transportation (divided into
short and long distance) and rail freight transportation.
The transport modes of inland shipping and aviation
are represented by a non-specified general process
where the development of the transport demand is
embodied by the final energy demand.
The household sector contains eleven demand
categories (space heating, cooling, hot water, cooking,
refrigeration, lighting, washing machines, laundry dryer,
dishwasher, other electrics, other energy use), whereof
September 5
th
to September 7
th

230

the first three correlate to specific building types (single
family houses in urban and rural areas and multi-family
houses each described as existing stock and new
build). The commercial sector is represented by a
similar reference energy system (RES) and consists of
nine demand categories (space heating, cooling, hot
water, cooking, refrigeration, lighting, public street
lighting, other electrics, other energy use). The first
three of them are subdivided according to different
building types (large/small).
The agricultural sector is described by a general
process with a mix of several energy carriers as input
and an aggregated demand of end use energy as
output.
The industrial sector is subdivided into several
branches (for example, iron and steel, cement, lime)
and into energy intensive and non-intensive branches.
While the intensive ones are modelled by a process
orientated approach, the other industries have a similar
structure but with five energy services (process heat,
steam, machinery drive, electrochemical, others)..
The generation of electricity and heat in power plants,
CHPs and heating plants is differentiated into public
and industrial production. The model contains three
different voltage levels of electricity (high voltage,
medium voltage, low voltage) and two independent
heat grids (district heat, local heat).
In the supply sector, all primary energy resources
(crude oil, natural gas, hard coal, lignite) are modelled
by supply curves with several cost steps. Three
categories can be differentiated: discovered reserves
(or developed sources), growth of reserves (or
secondary and tertiary extraction) and new discoveries.
Additionally, seven bio energy carriers are
differentiated: matured forest, bio gas, household
waste, industrial waste, as well as energy plants
containing sugary, starchy and lignocelluloses..
Due to its regional resolution, TIMES PanEU allows the
consideration of country specific features, for example
different structures of the stock of power plants,
different extension potentials for renewables as well as
potentials for storing CO
2
. An interregional electricity
trade is implemented in the model, so that exports and
imports of electricity according to the existing border
capacities could be calculated endogenously in the
model.
The role of CHP and district heating will be influenced
in the future by the heating demand for the heat, space
heating and cooling processes. The following chapters
describe the status and the assumed development for
Europe.
Industrial heat demand by temperature and sub-
sector in the EU27
The particular sub-sectors of the industrial sector use
different chemical and physical conversion processes.
Therefore, they need heat on different temperature
levels (Figure 1). Processes with a need for very high
temperatures (> 1400 C) are e.g. blast furnaces
(iron/steel industry) or kilns (cement or lime industry).
Processes with lower temperature levels occur in the
food/tobacco (sugar production, dairy) industry, other
industries or in general for the supply of space heating
and hot water. Also, the pulp/paper industry has a high
need for heat at a lower temperature level (< 100 C).
Most of the heat is produced by the combustion of
fuels. Other heat is generated by the use of electricity.
Key processes using electricity for high temperature
heat are chlorine electrolysis, aluminium electrolysis,
electric arc processes (iron/steel) and copper
electrolysis.
0
500
1,000
1,500
2,000
2,500
0
-
6
0
6
0
-
1
0
0
1
0
0
-
1
2
0
1
2
0
-
1
8
0
1
8
0
-
2
4
0
2
4
0
-
3
0
0
3
0
0
-
3
6
0
3
6
0
-
4
2
0
4
2
0
-
4
8
0
4
8
0
-
5
4
0
5
4
0
-
6
0
0
6
0
0
-
7
0
0
7
0
0
-
8
0
0
8
0
0
-
9
0
0
9
0
0
-
1
0
0
0
1
0
0
0
-
1
1
0
0
1
1
0
0
-
1
2
0
0
1
2
0
0
-
1
3
0
0
1
3
0
0
-
1
4
0
0
1
4
0
0
-
1
5
0
0
>

1
5
0
0
F
i
n
a
l

e
n
e
r
g
y

c
o
n
s
u
m
p
t
i
o
n

[
P
J
]
Others
Food/Tabacco
Pulp/Paper
Oth. non-metallic
minerals
Glass flat
Glass hollow
Lime
Cement
Oth. chemicals
Chlorine
Ammonia
Oth. non-ferrous
metals
Copper
Aluminium
Iron/Steel
Figure 1: Final energy consumption for industrial heat
production by temperature and sub-sector in the EU-27
in 2005

On the country level, the role of the different member
states concerning a particular temperature level
depends on the structure of the industrial sector in that
country. In general, the final energy consumption for
heat production at a specific temperature level is
dominated by the bigger member states and members
of the EU-15 like Germany, Italy, UK, France and
Spain. However, new member states like Poland,
Czech Republic or Romania also play an significant
role. Some countries only play a key role at single sub-
sectors and thus only for some temperature levels.
The lower temperature levels are dominated by the
industrial sub-sectors pulp/paper, food/tobacco and
others. Due to high activities in those areas, the
heating demand is clearly influenced by France (strong
for food/tobacco), Sweden and Finland (strong for
pulp/paper) next to other big countries like Germany,
Italy and UK. Italy and Spain play a large role,
especially at very high temperatures, due to their high
amount of cement production. In the Netherlands, the
September 5
th
to September 7
th

231

chemical and food & tobacco industries are the most
important ones. Each country is clearly specialised in
differing industrial sub-sectors.
Space heating and cooling demand in Europe
today and in future
The demand for space heating and cooling differs
among the countries in Europe due to the differences in
climatic conditions and in living standards (e.g. square
meters per capita) and building standards. This is
especially applicable to the assessment of current and
near future energy demand.
0
2000
4000
6000
8000
10000
12000
14000
2005 2010 2015 2020 2025 2030 2035 2040 2045 2050
H
e
a
t

d
e
m
a
n
d

i
n

[
P
J
]
Commercial Urban MFH Urban SFH Rual New Houses
0
200
400
600
800
1000
1200
2000 2005 2010 2015 2020 2025 2030 2040 2050
year
C
o
o
i
l
i
n
g

d
e
m
a
n
d

i
n

P
J
Residential SFH
Residential MFH
Commercial

Figure 2: Demand for space heating/hot water and for
cooling in the EU27

In the 2000, the useful demand for cooling was less
than 5% lower than the useful demand for space
heating and hot water. In the long term, the cooling
demand will be dominated by the commercial sector.
The increase of cooling demand in the EU27 up to
2050 will reach approx. 1120 PJ in the residential and
commercial sectors.
SCENARIO DEFINITION
A parameter variation is used to evaluate the reduction
potential and the role of CHP and district heat in the
energy system of the EU27. By varying the CO
2
price,
the reduction potential curves are constructed.
Therefore, different scenarios with different CO
2
prices
(one common price for ETS and Non-ETS sectors) are
calculated with TIMES PanEU and the reductions
occuring at this price level are analysed according to
the role of the different reduction possibilities.
The foundation for the CO
2
price variation is set based
on the CO
2
price outcomes from two scenario runs with
a reduction target of 15% [scenario: 15% reduction
(2020)] and 40% [scenario: 40% reduction (2020)] in
2020 compared to the Kyoto base year (Table 1). In the
long run (2050), both of these restricting scenarios
have the same target which equals a 450ppm goal
(-71% in 2050 compared to 1990).
2020 2025 2030 2035 2040 2045 2050
40%reduction (2020) -40% -45% -50% -55% -61% -66% -71%
15%reduction (2020) -15% -20% -25% -37% -48% -60% -71%

Table 1: CO2 reduction pathways for the two restricting
scenarios

The resulting CO
2
prices of these two restriction
scenarios build the framework for the price variations.
Within the range of the resulting CO
2
prices, the carbon
price varies between 10 /tCO
2
and 110 /tCO
2
in 2020
in increments of 10 . In 2030, the price varies between
27 / tCO
2
and 123 / tCO
2
. The price increases until it
reaches the level of a 450 ppm scenario in 2050
(Figure 3). The emission reductions are evaluated
using the results from the different scenarios in
comparison to the case of the lowest CO
2
prices (10 /t
in 2020, 27 /t in 2030). First, the total reductions over
all sectors are presented and afterwards the focus will
be on the industrial sector. The drivers of the reduction
are shown separately. Looking at the industrial sector,
the reasons for the emission reductions could be split
up into more efficient production processes, more
efficient heat supply, fuel switch in heat generating
units or CCS technologies in production processes and
energy supply.
0
50
100
150
200
250
300
350
400
450
2020 2025 2030 2035 2040 2045 2050
C
O
2
p
r
i c
e
[
2
0
0
0
]
15%reduction
(2020)
40%reduction
(2020)
CO2_10
CO2_20
CO2_30
CO2_40
CO2_50
CO2_60
CO2_70
CO2_80
CO2_90
CO2_100
CO2_110

Figure 3: CO2 prices of the different scenarios
OVERVIEW OF THE DYNAMIC DEVELOPMENT OF
THE ENERGY SYSTEM OVER TIME
In the following analysis, the two scenarios with the
lowest (10 /t CO
2
in 2020, scenario CO2_010) and the
highest (110 /t CO
2
in 2020, scenario CO2_110)
prices are displayed to show the range in which the
results of the price variation occur. Therefore, the
development over the whole modelling horizon
(20002050) is presented to rank the more detailed
September 5
th
to September 7
th

232

results from a point of time within these more general
results over a period of time. Since the CO
2
reduction
target of the two bounding scenarios [scenario 15%
reduction (2020) and scenario 40% reduction (2020)]
clearly differ in the mid-term periods of 2020 and 2030
(Table 1) and the corresponding prices are more
different in these periods (Figure 3), the energy system
shows the most variations during this time.
To show the development over the modelled time
period, first of all the net electricity generation of EU-27
is displayed (Figure 4). The overall electricity
generation remains almost constant at 2010 levels
(about 3 200 TWh) until 2030. In later periods, there is
a clear increase in electricity generation up to
4 255 TWh (2050, scenario CO2_110). The increase in
the later periods is driven by stronger emission
reduction targets. To fulfil the restrictions, more
electricity with low specific emissions and high end use
efficiency in the demand sectors is used.
According to the given CO
2
prices of the two scenarios
(CO2_010 and CO2_110), the main differences occur
in the mid term periods. While the total electricity
demand in 2020 is lower in the scenario with higher
emission certificate prices (-22 TWh in 2020 between
CO2_110 and CO2_010), the demand is higher by
86 TWh in 2030. The increase is due to the use of
more efficient technologies in the end use sectors
resulting in lower electricity demand in 2020, while by
2030 the switch to electricity based technologies to fulfil
the emission restrictions has already taken place.
0
500
1000
1500
2000
2500
3000
3500
4000
4500
S
t
a
t
i
s
t
i
c
C
O
2
_
0
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
2000 2010 2015 2020 2025 2030 2035 2040 2045 2050
N
e
t

e
l
e
c
t
r
i
c
i
t
y

g
e
n
e
r
a
t
i
o
n

[
T
W
h
]
Others / Waste
non-ren.
Other
Renewables
Biomass /
Waste ren.
Solar
Wind
Hydro
Nuclear
Natural gas
Oil
Lignite
Coal

Figure 4: Net electricity generation in the EU-27
Aside from the changes in the total electricity demand,
there is also a change in the structure of the electricity
generation. At higher CO
2
prices, less coal (-120 TWh
from coal fired power plants in 2030) and more gas
(+44 TWh) and nuclear (+30 TWh) are used and more
electricity from renewable energy sources (+35 TWh
from wind, +56 TWh from biomass and renewable
waste) is generated. Furthermore, CCS is used more
widely under the conditions of the CO2_110 scenario in
2030 compared to CO2_010.
The electricity generation from CHP plants in the EU27
increases by 79% from about 380 TWh in the year
2000 to 640 TWh by the year 2020 (see Figure 4). The
extension of the electricity generation from CHP plants
is essentially supported by gas-fired and biomass
based CHP plants. Additionally, existing public CHP
plants with an extraction condensing turbine are
substituted by CHP plants with a higher power-to-heat
ratio and there is also an extension of industrial CHP
plants, which are often used in cooperation with
communal facilities. The intermediate growth of CHP
plants in the commercial sector between the years
2015 and 2035 are based on efficiency advantages of
CHP plants with a medium sized internal combustion
gas engine. In the long term, the limited possibilities of
using CO
2
free fuels in commercial CHPs will result in
these phasing out in the commercial sector. Until the
year 2050 the electricity production by CHP plants in
the scenarios further increases up to a level of 1055 to
1100 TWh. CHP plants based on biomass as well as
CCS CHP are an important option in the year 2050.
0.00
200.00
400.00
600.00
800.00
1000.00
1200.00
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
2000 2005.0 2010 2015 2020 2025 2030 2035 2040 2045 2050
N
e
t

e
l
e
c
t
r
i
c
i
t
y

g
e
n
e
r
a
t
i
o
n

C
H
P

i
n

[
T
W
h
]
Public
Comercial
Industry

Figure 5: Net electricity generation CHP by sector in the
EU-27

In addition to the net electricity generation, the primary
(Figure 6) and final energy (Figure 7) consumption of
the EU-27 are also analysed over the whole time
period. Overall, the primary energy consumption (PEC)
does not show clear changes and remains at a level of
about 75 000 PJ. The lowest total PEC occurs in the
mid-term periods. The total consumption is influenced
by an increasing efficiency till 2030 and later on by a
higher share of renewables and also CCS which both
lead to a higher consumption due to the lower thermal
efficiency in the combustion processes.
Looking at the impact of the single energy carriers,
there is a distinct change between the two scenarios
than in the total sum of the PEC. In 2030 at a higher
CO
2
price, less coal (-1 675 PJ) and petroleum
products (-881 PJ) and more Hydro, wind, solar
(+338 PJ) and other renewables +4856 PJ) (mainly
biomass) are used.
September 5
th
to September 7
th

233

0
10000
20000
30000
40000
50000
60000
70000
80000
90000
S
t
a
t
i
s
t
i
c
C
O
2
_
0
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
2000 2010 2015 2020 2025 2030 2035 2040 2045 2050
P
r
i
m
a
r
y

e
n
e
r
g
y

c
o
n
s
u
m
p
t
i
o
n

[
P
J
]

Electricity
import
Waste (non
renewable)
Other
renewables
Hydro, wind,
solar
Nuclear
Natural gas
Oil
Lignite
Coal

Figure 6: Primary energy consumption in the EU-27

The final energy consumption (FEC) shows
comparable results (Figure 7). The use of petroleum
products declines over time in both scenarios
(-9 052 PJ in scenario CO2_010 between 2000 and
2050). The use of gas increases at lower CO
2
prices in
the mid-term periods (up to more than 13 500 PJ in
2020 at scenario CO2_010), but declines in both
scenarios at the very end. This shows that one early
and cost-effective measure for emission reduction is
the fossil fuel switch from petroleum products and coal
to gas in the end use sectors.
As already shown with electricity generation, the use of
electricity also increases in the end use sectors.
Especially in the long run at higher carbon prices, there
is a clear rise. The use of renewable energy sources
also increases constantly in both scenarios. In 2020
and 2030, clearly more renewables are used in the
CO2_110 scenario due to the higher CO
2
prices
(+3900 PJ in 2030).
In contrast to the PEC, the total FEC decreases slightly
in the long run. The reason for this different
development is that the higher conversion losses
arising from a higher electricity demand and the
extended use of renewables and CCS at the public
electricity generation are balanced at PEC and do not
influence the FEC.
Even though more renewables (mainly biomass) are
used, due to the higher use of electricity with its high
end use efficiency and other efficiency improvements,
the total FEC declines to 49 482 PJ (in 2050 at
scenario CO2_110). This efficiency improvement
occurs in the industrial sector mainly at industrial
production processes, but is also clearly driven by
efficiency improvements ain the residential and
transport sectors.9

9
For a detailed discussion of the effects in the different end
use sectors and its impact on the total final energy
consumption see /Blesl et al. (2010)/
0
10000
20000
30000
40000
50000
60000
S
t
a
t
i
s
t
i
c
C
O
2
_
0
1
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
2000 2010 2015 2020 2025 2030 2035 2040 2045 2050
T
o
t
a
l

f
i
n
a
l

e
n
e
r
g
y

c
o
n
s
u
m
p
t
i
o
n

[
P
J
]

Others (Methanol,
Hydrogen)
Waste
Renewables
Heat
Electricity
Gas
Petroleum
products
Coal

Figure 7: Final energy consumption in the EU-27

In contrast to the year 2000, the distribution of local
and district heat to the household, commercial and
industrial sectors changes by the year 2050 with an
additional approx. 1000 PJ district consumed in the
year 2050 (see Figure 8).
0
500
1000
1500
2000
2500
3000
3500
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
C
O
2
_
0
1
0
C
O
2
_
1
0
0
2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050
F
i
n
a
l

e
n
e
r
g
y

c
o
n
s
u
m
p
t
i
o
n

h
e
a
t

i
n

[
P
J
]
Supply
Residential
Industry
Comercial
Agriculture

Figure 8: Final energy consumption district heat in the
EU27

In the long term, the CO
2
contents of the heat supply for
the end use sectors will be reduced from 130 kg
CO2/MWh to 122 kg CO2/MWh in 2020 and from 113
kg CO2/MWh to 36 kg CO2/MWh in the year 2050,
which is one explanation for achieving the CO2
reduction targets in this area. On the other hand, the
possibility to use renewable energy or to install CCS,
increasingly influences the penetration of CHP. By
2050, fossil heat plants will also be substituted with
large heat pumps and solar thermal heat plants in
combination with storages, biomass heat plants fuelled
with wood or woody crops and biogas.
The overall emissions decrease is based on the
emission reductions of the single sectors leading to
different CO
2
abatement costs (Figure 9). The total
emissions correspond to the emission pathway of the
two restricting scenarios (scenario 15% reduction
(2020) and scenario 40% reduction (2020), see
Table 1. The earliest and strongest reductions take
place in the conversion/production sector. The
industrial sector and the residential/commercial sector
also show clear reductions. The transport sector tends
September 5
th
to September 7
th

234

only to reduce its emissions with very strict reduction
targets connected to high carbon prices.

0
500
1000
1500
2000
2500
3000
3500
4000
4500
S
t
a
t
i
s
t
i
c
C
O
2
_
0
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
C
O
2
_
0
1
0
C
O
2
_
1
1
0
2000 2010 2015 2020 2025 2030 2035 2040 2045 2050
E
m
i
s
s
i
o
n
s

o
f

C
O
2

[
M
t
]
Transport
Households,
commercial,
AGR
Industry
Conversion,
production

Figure 9: CO2 emissions in the EU-27
ANALYSIS AT A SPECIFIC POINT OF TIME WITH
FOCUS ON 2030
After the general effects are described and the
scenarios with the lowest and highest CO2 prices are
analysed over the whole period of time, a more detailed
analysis shows the effects in the industrial sector
during the mid-term periods with a particular focus on
2030.
Firstly, the reduction potential of the different sectors
should be analysed (Figure 10). Both conversion
/production and the other end use sectors are taken
into account. As in the results of the emission reduction
from 2000 to 2050 (Figure 9), the industrial reduction
potential plays the key role next to the
conversion/production sector. Looking at the year 2030
and comparing the additional CO
2
reductions when the
CO
2
price is increased from 27 /t to 123 /t, the
strongest additional reduction occurs at the conversion
sector (+351 Mt at a price of 123 /t compared to
27 /t). An additional 301 Mt of CO
2
are reduced by the
industrial sector.
Especially at higher prices above 94 /t, the reduction
potential of the industrial sector becomes more and
more important. Its share of the total additional
reduction increases from 33% (36 /t compared to
27 /t) to 37% (123 /t to 27 /t). The lowest reduction
occurs in the transport sector. Till a price of 85 /t, only
an additional 6.3 Mt are reduced, while at a price of
123 /t an additional 18.9 Mt are reduced. In the
residential and commercial sector, some reduction
possibilities are cost-effective even without a price on
CO
2
. The energy savings outweigh the additional
investment costs. Those reduction measures are
especially connected to the building/heating sector.
0
100
200
300
400
500
600
700
800
900
36 46 56 65 75 85 94 104 114 123
Carbon price [2000/tCO2]
A
d
d
i
t
i
o
n
a
l
e
m
i s
s
i
o
n

r
e
d
u
c
t
i o
n

[
M
t
]
Transport
Commercial
Residential
Industry
Conversion/
Production

Figure 10: Additional CO2 reduction in the EU-27 in 2030
by sector compared to the scenario with the lowest CO2
price of 27 /t

One reason for the CO2 reduction in the residential,
commercial and industrial sectors is the increase in
final energy demand from district heat (see figure 11).
The overall increase of the district heat demand
influenced by the different CO2 prices is 14%. The
biggest growth can be seen in the commercial sector,
where the total district heat demand for district heating
grows by over 30% between the min the minimum and
maximum CO2 certificate price.

0
200
400
600
800
1000
1200
1400
1600
1800
2000
27 36 46 56 65 75 85 94 104 114 123
Carbon price [/T CO2]
F
i
n
a
l

e
n
e
r
g
y

d
e
m
a
n
d

h
e
a
t

i
n

[
P
J
]
Residential
Industry
Comercial
Agriculture

Figure 11: Final energy demand of district heat in the
EU-27 in 2030 by sector

However, the generation of district heat from the use of
renewable sources and CCS will be one reason for the
growth of the reduction potential in the conversion
sector (Figure 12). The share of the use of renewables,
especially biomass, will rise from 29% to 60%. More
than 1300 PJ of additional biomass will be needed.
Due to this increase, the average heat to power ratio of
all CHPs will fall from 0.9 to 0.66. In the cases where
CO
2
prices exceed 56 /tCO
2
, the district heat
generation in CCS CHP plants grow more rapidly. The
specific emissions of the district heat generation
decrease from approx. 380 kg / MWh to 84 kg / MWh.
September 5
th
to September 7
th

235

0
200
400
600
800
1000
1200
1400
1600
1800
2000
27 36 46 56 65 75 85 94 104 114 123
carbon price in [/t CO2]
D
i
s
t
r
i
c
t

h
e
a
t

g
e
n
e
r
a
t
i
o
n

i
n

[
P
J
]
Heat Plant RES
Heat Plant
CHP RES
CHP FC
CHP CCS
CHP

Figure 12: District heat generation in the EU-27 in 2030 by
technology group

In the industrial sector, the share of CHP will grow. The
additional emission reductions by the industrial sector
of 301 Mt in the year 2030 could be split into industrial
supply and industrial production processes. The supply
side covers the industrial generation of energy
commodities or energy services. These are electricity
from industrial condensing power plants and CHPs,
heat and steam from CHPs and boilers, space heating
and heat for hot water as well as cooling. The supply
activities play an important role in the industrial sub-
sectors with a high share of space heating (such as
food & tobacco or other industries) or low temperature
process heat (such as pulp & paper or food & tobacco).
In total, from the additional reduced emissions, 147 Mt
are reduced by industrial supply processes and 154 Mt
by production processes in 2030. While at lower a CO
2

price more emissions are reduced on the supply side
(66% of the additional reduction based on supply
processes at 46 /t), at higher prices more and more
reductions take place on the production side (49 %
based on supply processes at 123 /t).
The additional electricity needed at high CO
2
prices is
mainly generated by industrial autoproducers. Within
this industrial production, the additional electricity
mainly comes from CHP power plants. The use of
electricity in the industrial sector from public generation
remains relatively constant even when the CO
2
price
increases. Accordingly, one key way to reduce the
emissions on the supply side is through the extended
use of CHP plants for industrial power generation. This
higher amount of electricity from industrial
autoproducers (Figure 13) leads to higher conversion
losses in total when the fuel use is considered. As
described above, that is one reason for the difference
between final energy consumption and fuel
consumption. Another reason is the lower efficiency of
electricity generation due to the higher use of CCS.
4000
4020
4040
4060
4080
4100
4120
4140
4160
4180
4200
27 36 46 56 65 75 85 94 104 114 123
U
s
e

o
f

E
l
e
c
t
r
i c
i
t
y

[
P
J
]
-10%
-5%
0%
5%
10%
15%
20%
25%
30%
C
h
a
n
g
e

o
f

E
l e
c
t
r
i c
i t
y

s
u
p
p
l
y

b
y

t
e
c
h
n
o
l
o
g
y

c
o
m
p
a
r
e
d

t
o

C
O
2

p
r
i
c
e

o
f

2
7

/
t
SUM
public
generation
Condensing
industrial
CHP
industrial

Figure 13: Use of electricity in the EU-27 in 2030 by
technology

The other part of the industrial supply processes is the
industrial heat generation. The drivers for the emission
reduction in industrial heat production are a switch to
biomass (from coal and clearly from gas) and the use
of CCS in industrial CHPs (Figure 14). Between a CO
2

price of 36 and 56 /t of CO
2
in 2030, there is a clear
increase in the use of renewables in boilers. The share
of renewables in the total fuel use in industrial boilers
increases from 33% to 51%. As a result, the thermal
efficiency of boilers has an overall decrease.
In industrial CHPs, there is also a slight increase in the
use of s. This switch takes place between CO
2
prices
of 27 /t to 65 /t. However, the main change
concerning CHPs is the increasing use of CCS. At a
CO
2
price above 94 /t, there is a clear rise in the use
of this technology. These CCS CHPs are mainly gas
fired
10
. This is why the share of renewables used in
industrial CHPs declines at a price over 75 /t again.
Like biomass, the extended CCS use also leads to
lower efficiencies resulting in both the efficiency of
boilers and CHPs to decline over time. Accordingly, the
key driver is not efficiency improvements, but the use
of renewables and CCS. The effects of renewables and
CCS compensate the trend to lower energy intensity
within one technology. Gas boilers become more
efficient and as do biomass boilers. However, the more
efficient biomass boilers still use more fuel than the gas
boilers.
Looking at the heat output by technology, there is also
a shift (Figure 14). At lower emission prices, the heat
output from industrial boilers stays almost constant.
Within this range, the share of renewables used
increases (as illustrated in Figure 13). Afterwards, at a
price above 65 /t, boilers are substituted with heat
from CHPs and district heat. Both heat commodities

10
For a detailed analysis of the CCS potentials, costs
and the modelling of CCS in TIMES PanEU see
/Kober, Blesl (2010a), Kober, Blesl (2010b), Kober,
Blesl (2009)/
September 5
th
to September 7
th

236

are generated in combination with an increasing share
of renewables, a higher CO
2
price and from CCS.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
27 36 46 56 65 75 85 94 104 114 123
E
f
f
i c
i
e
n
c
y

a
n
d

s
h
a
r
e

[
%
]
CHP industrial
(total)
CHP industrial
share RES
Boiler industrial
Boiler industrial
share RES
CCS CHP
industrial (share
Fuel input)
Figure 14: Efficiency of heat supply technologies and
share of CCS at industrial CHP in the EU-27 in 2030

-150
-100
-50
0
50
100
150
36 46 56 65 75 85 94 104 114 123
c
h
a
n
g
e

i n

h
e
a
t

o
u
t
p
u
t

[
P
J
]
District Heat
CHP industrial
Boiler
Figure 15: Heat supply by technology in the industrial
sector in the EU-27 in 2030 compared to the scenario with
the lowest CO
2
price of 27 /t

In total, all these described effects concerning the
industrial supply processes lead to the additional
emission reduction in 2030 of 147 Mt at a price of
between 123 /t and 27 /t. In general, more emissions
are reduced in boilers than in CHPs. The reasons are
the fuel switch from coal and mainly gas to renewables
at lower CO
2
prices and later on the substitution of
boilers with CHPs (less boilers are used and therewith
produce less emissions).
Due to a higher use of CHPs, there is no clear increase
in emissions during the mid-term ranges. When the
output of heat stays constant and a higher share of
CCS is used, then clear emission reductions from
CHPs (additional 48.7 Mt in 2030 at 123 /t compared
to 27 /t) occur.
CONCLUSION AND OUTLOOK
District heating generation offers an economic potential
for expansion in the future. Depending on the regions
or countries, the development will be different because
the starting point is economic growth and the existing
national laws or cross-subsidies for competitors energy
carriers. In addition, it is necessary to take into account
that the climate conditions within Europe differ
substantially.
Within the energy system of the EU-27, there are
different emission reduction pathways. The emissions
could be reduced by a fuel switch in more efficient (or
better, less carbon intensive) energy supply or by a
change in production processes. Key drivers
concerning the emission reduction in production
processes and in heat demand side are efficiency
improvements due to new technologies and
technological improvements. The key driver concerning
the supply side of electricity and heat generation is the
increased use of renewables, mainly biomass, for heat
generation. The CCS technology also plays an
important role in the reduction of emissions. Due to the
increased use of renewables in CHP and heat plants
and the use of CCS, the efficiency in the supply
processes decreases at higher CO
2
prices.
In the long run to a CO
2
-free world, the possibility to
generate district heat with renewable energy and the
use of CCS make the decarbonisation of the energy
consumption in the end use sectors possible.
In general, the progression of district heat depends
crucially on the possibility of generating CO
2
emission
free district heat and electricity.
REFERENCES
[1] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Effects
of climate and energy policy related measures and
targets on the future structure of the European
energy system in 2020 and beyond, Energy Policy,
2010 (forthcoming)
[2] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Beitrag
von technologischen und strukturellen
Vernderungen im Energiesystem der EU 27 zur
Erreichung ambitionierter Klimaschutzziele,
Zeitschrift fr Energiewirtschaft 04/2008
[3] Blesl, M.: CHP and district heat in the Europe
under an emission reduction regime, in: 11th
Cooling in Reykjavik, Island
[4] Blesl, M., Cosmi, C. ,Kypreos, S. , Salvia, M.:
Technical paper n Technical Report n T3.18
RS 2a Summary report of Pan European model
results BAU scenario EU Integrated Project
NEEDS New Energy Externalities Developments
for Sustainability October, 2008
[5] DEHSt (2010): Deutsche
Emissionshandelsstelle, Kohlendioxidemissionen
der emissionshandelspflichtigen Anlagen im Jahr
2009 in Deutschland, Mai 2010
September 5
th
to September 7
th

237

[6] EEA (2010): European Environment Agency,
European Union emission trading scheme (ETS)
data viewer, 2010
[7] Kober, Blesl (2010a): Analysis of potentials
and costs of storage of CO2 in the Utsira aquifer in
the North Sea; report work package 4: Regional
analysis at North Sea level, 2010, www.fenco-
era.net
[8] Kober, Blesl (2010b): Perspectives of CCS
in Europe considering technical and economic
power plant uncertainties; in PLANETS work
package 6 deliverable No. 15 Report on
Probabilistic Scenarios, 2010, www.feem-
project.net/planets
[9] Kuder, Blesl (2009): Kuder, R.; Blesl, M.: Effects of
a white certificate trading scheme on the energy
system of the EU-27, Fullpaper 10th IAEE
European Conference in Vienna, Austria, 2009
[10] UNFCCC (2009): GHG inventory reports for the
single member states of the EU-27, submission
2009 situation / problems / wishes, Energy Policy
EHP meeting, Budapest, 11 September 2008,
www.lsta.lt/files/seminarai/080911_Budapestas/CZ.pdf

September 5
th
to September 7
th

238

Waste coll.
Logging
Extraction
Processing
Storage
Generation
Transformation
Air and room
Hot tap water
Electricity
Delivered Energy (DE)dh+el
Net energy demand
Primary energy use = DEdh PEFdh(f (x,y,z)) +DEel PEFel (f (x,y,z)) = (Weighted delivered energy indicator (kWh/m
2
)) A
C
Delivered energy (el)
Delivered energy (dh)
Energy carrier (y)
Energy carrier (x)
Transportation
Distribution and
Transmission(dh)
Heating
systems
Energy carrier (z)
Primary energy use calculated by PEF(x,y,z)
End use
demand!
( A
C
= conditioned floor Area )
Distribution and
Transmission(el)
DH substation
Calculation of Primary Energy use according to EPBD and mandated EN-standards
2010/ 04/ RU
CHP-plant
Heat boilers
Storage
(PEF might be PEF
R
or PEF
T
depending on purpose)
Calculating direction
Calculating enduse and losses byENstandards worked out according to mandate from the EUCommission
Heating systems efficiency
heating system
+
System border for the energy performance indicators is the whole energy chain from the source to the end use
Figure 1.- Sketch of the calculation of Primary Energy use according to EPBD and mandate EN-standards
Figure 1 Sketch of the calculation of Primary Energy Use according to EPBD and mandate EN-standards
CONSIDERATIONS AND CALCULATIONS ON SYSTEM EFFICIENCIES OF HEATING
SYSTEMS IN BUILDINGS CONNECTED TO DISTRICT HEATING
Maria Justo Alonso
1
, Rolf Ulseth
2
and Jacob Stang
1

1
SINTEF Energy Research, Department of Energy Processes
2
NTNU, Faculty of Engineering Science and Technology,
Department of Energy and Process Engineering

ABSTRACT
In order to harmonize the implementation of the EC
Directive on the energy performance of buildings
(EPBD) [1], and to provide guidelines and common
calculation tools, several technical standards have
been worked out by CEN in accordance with a
mandate from the EC. This paper focuses on
calculating system efficiencies of hydronic heating
systems by using the standards EN 15316-x-x [4], [5],
[6].
The paper has been written in order to ease and
diminish the time consuming process of interpreting
details in the standards such as the numbered EN
15316-x-x, and with the goal to enlighten main parts of
these standards.
To exemplify some results, an apartment building of
1000 m
2
floor area located in a climate like Oslo is
chosen. In the base case, the design distribution
temperatures in the building are 80/60. The different
efficiency figures applying for this case are calculated
efficiency values for the production of the heat, for its
distribution through the building and its emission in the
room. The room efficiency is the one that has the
bigger influence on the total system efficiency.
INTRODUCTION
The Directive on the energy performance of buildings is
carried out in order to be used together with a number
of EN-standards. The main goal of the Directive is to
promote the improvement of the energy performance of
buildings within the Community, taken into account
outdoor climatic and local conditions as well as indoor
climate requirements and cost-effectiveness. The main
focus is on reducing the primary energy use and the
associated CO
2
emission of buildings.
Figure 1 shows how the Primary energy use is
calculated based on all the steps where the energy is
changing its nature from the source to the end use. In
the current case, the energy calculations are performed
for the systems within the building to be able to
calculate the delivered energy to the building. This
means that the building substation with the heat
exchangers and tap water storage are included.
In the current scenario, all the losses before the heat is
delivered to the building are included in the primary
energy factor (PEF) for the delivered heat. In case of
considering the complete scenario, the boundaries for
the energy performance indicators are the whole
energy chain from the source to the end use. In this
case, if a CHP plant is represented, the power bonus
method (EN 15316-4-5) should be used. This method
is giving the produced district heat a bonus for the
electricity produced assuming that this electricity
replaces electricity production with a high PEF-value.
According to the implementation of EPBD, it is crucial
that the system borders are clearly defined so that the
delivered energy is doubtlessly defined.
September 5
th
to September 7
th

239

METHOD
Definition of building and system build up
In this exemplified case, the main chosen building is an
apartment building. This building category seems to be
the most representative concerning heat use among
the building categories defined in the EPBD [1].
In the shown example, the size of the building is
chosen to be 1000 m floor area since this size should
be rather representative and be a good compromise
between the previous and the proposed new recast of
the EPBD. [2]
A building of these features corresponds to a three
storeys squared building with four flats of about 80 m
per storey.
In this setting, the total heating system efficiency in the
building is built up based on the differentiation between
the three main parts of the system. It must be defined
where the substation is located in the building, i.e.
where the heat is exchanged from the distribution
network DH stage in Figure 2. The heat supply to the
heating system within the building from the district
heating system is assumed to be provided by two heat
exchangers and hot water storage defined as the
building substation part of the system.
Figure 2 Sketch of the system elements for production,
distribution and conditioning of the rooms

From the substation, the hot water is distributed either
for air and space heating or as domestic hot water.
Both uses are provided by their own heat exchanger
and the necessary pipelines will now be referred to as
distribution pipelines.
Finally; once distributed, the heat is emitted according
to the demand. For the present case, the heat is
delivered either by radiators (80/60 C) in the base
case, floor heating (35/28 C), or domestic hot water at
60 C. Figure 2 gives a further visual explanation. For
the present paper, the supply of heat is just done by a
hydronic heating system. The possible heat loss from
the distributed air is neglected since the temperature of
the air is assumed to be slightly lower than the
temperatures in the rooms.
Categories of building
The presented analysis shows results for five kinds of
buildings described in the EPBD which are: single
family house and apartment block, office buildings,
hotel and restaurants, educational buildings and
hospital buildings. When it comes to heat consumption
for these buildings, the measurements performed in
Linda Pedersens PhD thesis [3] show that the
consumption of the apartment buildings is about
116 kWh/m, while hospitals use 150 and office
buildings use 100 kWh/m. These measured values
include the domestic hot water (DHW) and the space
heating (SH) consumption.
The calculated efficiency for the system will depend on
the size of the building as well. The present apartment
building shows a higher efficiency value than a single
family house with the same consumption. This is due to
higher relative losses in the substation.
Climate influence
The calculations in the present paper are based on a
climate like in Oslo, Norway. This climate is defined to
have approximately 5100 degree days with 20 C as
the internal reference temperature and an external
design temperature of -20 C [8].
In practice, the outdoor climate can vary widely from
place to place. Owing to this, the outdoor climate
affects not only the heat consumption but also the
relative losses. In general the relative losses are
increasing with an increased ratio between the degree
days and T between the dimensioning internal and
external temperature.
The average outside temperature affects the heat
consumption and the temperature variations affect the
regulation of the heat emitters. This means that during
cold periods, the temperature of the supply water tends
to be increased imposing an increase in the losses
related to the transport of water with higher
temperatures. The design temperature for the radiators
in the base case in this paper is 80/60, and in warmer
periods, this temperature is decreased in order to
reduce losses and adapt the supply temperature to the
outside temperature. This affects the efficiencies in a
positive way.
+
Hot tap
water
distrib.
Room
heating
distrib.
Substation
system
border
+
+
Ventil-
ation
air
CW
DH
2010/05/RU
+
Hot tap
water
distrib.
Room
heating
distrib.
Substation
system
border
+
+
Ventil-
ation
air
CW
DH
2010/05/RU
September 5
th
to September 7
th

240

Finally, the outdoor climate affects the length of the
heating season. Usually, the lower the average outside
temperature, the longer the heating season. This
however, does not affect the DHW since this is more or
less steady all the year along.
Positioning of substation
Figure 2 shows the positioning of the substation. The
heat is delivered from the district heating pipelines
through two separate heat exchangers, one for heating
the water in the storage tank for DHW by a circulating
loop, and the other for the air heating and space
heating system. The main reason for having two heat
exchangers is due to the different needs of temperature
levels. In the calculations dealing with the production,
the used heat demand used is the total demand, while
for the distribution the heat is divided into heat
distribution for SH and for DHW.
In order to calculate losses related to the storage tank,
it is assumed that the tank is of a common type with a
common value for the stand by heat loss. The system
design consists of the coupling in series of 289 litres
storage tanks. There are considered two tanks for the
apartment building of 1000 m
2
but one more tank if the
case is dealing with hospitals, educational and hotel
buildings. For other sizes the number of tanks is
adjusted according to the demand.
Efficiencies to be studied
For the present paper, as written previously, the
system is divided in three smaller system parts which
are independent. For every component, the efficiency
is calculated following different standards:
- Production; according to EN 15316-4-5:2007[4]
- Distribution; according to EN 15316-2-3:2007 and
EN 15316-3-2:2007[5]
- Room emission; according to EN 15316-2-1:2007[6]
The efficiency of the production includes the losses
depending on the thickness of the insulation material,
the insulation material itself, the storage tank, the
complete local piping system of the substation system
and the temperature difference between the two media
and the ambient. It takes into account the thermal loss
of the total substation. For this case the substation is
considered to be in an unheated part and therefore the
losses are considered as unrecoverable.
In case of the distribution, the efficiency depends on
the use of the heated water. In case of being a part of a
DHW system; the energy used for heating the water
which is not drawn-off and which slowly gets cold in the
pipelines, has to be considered as loss. Moreover, heat
is used to heat up the pipes and fittings. Since the
building is large enough to need a circulation loop this
loop is considered to be a source of loss? The water
temperature is here constantly at the designed point of
60
o
C.
When dealing with distribution of SH, the losses are
considered dependent on the kind of insulation material
and the ambient and the mean water temperature in
the supply and return pipes.
The heat emission to the room from the DHW draw-off
tap discharge cocks is considered to be negligible in
comparison to the total heat consumption.
Dealing with space heating a distinction is done with
respect to the kind of emission. Two major groups are
considered: the emission by floor heating and by
radiators. The first has a low temperature distribution of
35/28 C. As for the radiator system, the analyzed base
case is 80/60 C for supply/return design values.
Besides the temperature level, the placing in the room
affects the stratification efficiency and the loss through
the outside wall. Furthermore, another point related to
the temperature is the regulation of the room
temperature, which in our case, is assumed to be a
PI-regulator, even if in a lot of apartments this
regulation is quite often done by on/off regulation.
RESULTS
Production efficiency
The production efficiency is shown in Figure 3. By using
the losses on the heat demand and the temperature
difference as basis for calculation, the values in
Table 1 are obtained.
0,950
0,960
0,970
0,980
0,990
0 250 500 750 1000 1250 1500 1750 2000 2250
S
y
s
t
e
m

e
f
f
i
c
i
e
n
c
y
Building floor area [m
2
]
Production efficiencies for the different types of buildings according to the
EN 15316-4-5 :2007 (Oslo climate) with a distribution temperature 80/60
Apartment block
Of f ice building
Hotel and restaurant
building
Educational building
Hospital building
Single f amily

Figure 3 Production distribution of the 80/60 C district
heating for different buildings

As Figure 3 shows, the bigger the building, the higher
the efficiency. This effect is due to the reduction of the
relative losses when the size of the substation (kW)
increases. The curve profile is decreased slightly from
2000 m
2
and downwards, and then decreasing rapidly
from about 1000 m
2
down to 500 m
2
.
September 5
th
to September 7
th

241

In addition, it can also be observed that among all
types of buildings, apartment houses represent
somehow the highest efficiencies which justify the main
focus in this study. The displayed case applies for the
values where the design temperature level is 80/60 C.
It can be concluded from other calculations that the
higher the design distribution temperature level, the
lower the production efficiency. This conclusion is what
could be expected considering the difference between
the average temperature and the ambient temperature;
the larger this difference, the larger the losses.
As shown in Table 1, the efficiency varies only between
0.9784 and 0.9673. It can be concluded, compared to
the distribution loss values that the production
efficiency is not changing significantly even if the
temperature level is changed. As a conclusion it can be
said that the losses in the production are relatively low
for bigger houses but increasing quite rapidly for
smaller buildings.
Table 1 Efficiency of DH production system for different
design temperature levels.
Kind of
building 80/60 70/55 55/45 35/28
Apartment
blocks 0,9776 0,9778 0,9780 0,9784
Office building 0,9729 0,9732 0,9735 0,9740
Hotel and
restaurant
building 0,9701 0,9703 0,9706 0,9709
Educational
building 0,9676 0,9678 0,9681 0,9685
Hospital
building 0,9773 0,9775 0,9777 0,9780

The quality of the insulation of the storage tank will also
influence the production efficiency. Manufactures
should follow the standard pr EN5044:2005 [7] in order
to calculate these losses. Losses from storage tanks
should be considered closely in practice, and tanks
with relative high losses should be considered for
replacement or to be replaced by direct heat
exchangers for DHW.
Distribution efficiency
First the system for the distribution of tap water is
analyzed. In this case, the building includes a
circulation loop (in small dots Figure 2) which goes
from the storage tank, to the third floor and the
distribution branches (in bigger dots in Figure 2) which
deliver DHW from the central loop to the consumer.
The water temperature in the circulation loop is
assumed to be at 60 C throughout the whole year. The
biggest share of the losses come from the circulation
loop ranging up to 35 % of the total losses from the
total distribution system. This loss is related to the fact
that the water that remains in the distribution pipelines
represents 5% of the total losses per flat.
For this calculation it is assumed that the pipelines
have insulation which a loss of 0.3 W/mK (the
pipelines are considered to be according to the
category installed after 1995 in [5]). In this calculation
the losses due refilling the pipes with hot water are
included. This heat could be considered as recoverable
loss for space heating during the heating season but in
lack of a special national annex all the losses related to
the distribution of DHW should be considered as non
recoverable. These losses are not related to the
demand for heat and will consequently be lost or result
in increased room temperatures.
When it comes to SH, the losses are related to the
temperature difference in the non-heated areas where
the water goes through. These losses are relatively low
compared with tap water since most of these losses
are considered to be recoverable. The values used are
tabulated in the EN standard [5].
The percentage of recoverable losses is the cause of
the higher efficiency for distribution of space heating
which ranges 0.99, whilst the efficiency for distribution
of DHW is in thee range of 0.60.
Emission efficiency in the rooms
In this case, domestic hot water is not considered to
contribute to the room heating since the losses from
the discharge cocks are considered negligible.
In case of space heating a difference has to be made
between floor heating and radiator heating when it
comes to the efficiency calculations.
Floor heating is by its nature emitted at lower
temperature, which has an effect on the stratification
efficiency since the lower the temperature level, the
higher this
str.
By definition in [6] the stratification
efficiency of floor heating is 1 whilst this parameter for
radiators goes down to 0.91 on the 80/60 distribution
system. This value is combined with the efficiency
value of 95 % due to the positioning of the radiators on
a normal external wall. Together these values make a
total room efficiency of 0.93.
However, a regulation with PI controllers for the
radiators delivers an efficiency of 0.97 while the same
controller remains at 0.95 for floor heating.
In case of the embedded floor heating efficiency the
efficiency is 0.93. Since it is considered to be normal
insulation layer according to EN 1264, it results in a
emb
of 0.95, the combination results in
emb
of 0.94.
Due to these three parameters, floor heating all in all
has a room efficiency of 0.90 and radiators of 0.88.
September 5
th
to September 7
th

242

Total system efficiency
The total system efficiency is in this paper calculated
as the product of the efficiencies of the different pasts
of the heating system in the building.
Figure 4 shows the room efficiency and the total
system efficiency related to the temperature level
supply/return. The delivering water temperature to the
radiators is the parameter which affects the efficiency
the most.
The total system efficiency follows the pattern of the
room efficiency since this parameter has far the largest
influence.
Another conclusion from Figure 4 is that the efficiency
of the complete system varies significantly with the
design temperature level of the heating system. This is
due to the temperature difference between the heating
system components and the ambient. With a lower
distribution temperature the losses will be smaller.

System efficiency for an hydronic heating system in
an apartment building of 1000m
2
floor area
0.82
0.84
0.86
0.88
0.9
0.92
0.94
=80/60 =70/50 =55/45
Temperature distribution
E
f
f
i
c
i
e
n
c
y
Room
Efficiency
Total system
efficiency

Figure 4 Room efficiency and total system efficiency for
the different design distribution temperatures according to
EN 15136-2-1:2007 (Oslo climate). Radiators with
thermostatic valves mounted on normal external walls and
with heat supply from district heating

Calculated efficiencies for a floor heating system and
the hot tap water system are presented in Table 2.
Space heating with floor heating has a slightly lower
efficiency than radiators due to the lower efficiency for
the emission of the heat in the room.
Tap water systems have a lower efficiency since the
system is by its nature losing a considerable amount of
heat when leaving the hot water in the pipes between
the tapping cycles. This water is cooled down inside
the pipelines and is then being tapped without being
useful. In the present case the distance from the
substation to the furthermost apartment forces an
installation of a circulation loop in order to reduce the
waiting time for hot tap water at the tapping cocks. This
is a stand by source of loss. These two factors cause
the rather low efficiency of the hot tap water system.
Table 2 Total efficiencies of the systems for space
heating with floor heating and hot tap water system
CONCLUSIONS
The design temperature level for the system is the
most important factor when referring to the efficiency of
a hydronic heating system in buildings supplied by
district heating. . Therefore the possibility of lowering
the design temperature level of the heating system
should be considered closely. This increases the
emission efficiency in the room and reduces the losses
from the distribution pipelines. It saves energy and
increase the cooling of the district heating water
through the substation. Changing the positioning of the
radiator from the external wall to the internal wall
actually decreases the room emission efficiency.
The introduction of an energy performance certificate
for buildings according to EPBD requires a transparent
calculation model according to the standards in the EN
15316 series. This paper gives a picture of system
efficiencies for hydronic heating systems and also an
idea of the time consuming process that has to be
performed in order to calculate the efficiency of a
system in detail. Therefore, it is concluded that some
user-friendly guiding material should be desirable in
order to enlighten and facilitate the calculation process.
In the present paper the potential heat losses from the
ventilation system are neglected due to the fact that it
is assumed that the air temperature is distributed at
temperatures slightly below the room temperatures.
ACKNOWLEDGEMENT
This work has been supported by SINTEF, NTNU and
has been related to the project Systemvirkninsgrader
(System efficiencies) which was initiated by Standard
Norge and paid by the Norwegian Water Resources
and Energy Directorate. It has also been supported by
the Primary Energy Efficiency project which is paid by
Nordic Energy Research and companies in the heating
field in Norway.
REFERENCES
[1] European Parliament and Council on energy
efficiency of buildings, Directive 2002/91/EC on
the energy performance of buildings (EPBD)
[2] Proposal on a recast of Directive 2002/91/EC on
the energy performance of buildings, 2009-11-25
Total system efficiency
Floor heating 0.87
Hot tap water 0.59
September 5
th
to September 7
th

243

[3] PEDERSEN, L. (2007) Load Modelling of
Buildings in Mixed Energy Distribution Systems,
Department of Energy and Process Engineering,
NTNU,(Norwegian University of Science and
Technology), Trondheim
[4] EN 15316 Heating systems in buildings Method
for calculation of system energy requirements and
system efficiencies Part 4-5: Space heating
generation systems, the performance and quality of
district heating and large volumes, 2007
[5] EN 15316 Heating systems in buildings Method
distribution systems., 2007
[6] EN 15316 Heating systems in buildings - Method
emission systems., 2007
[7] CEN: Efficiency of domestic electrical storage
water-heater German version pr EN 50440,2005
[8] VVS-tekniske klimadata for Norge, Norges
byggforskningsinstitutt, Hndbok 33

September 5
th
to September 7
th

244

HEAT LOAD REDUCTIONS AND THEIR EFFECT ON ENERGY CONSUMPTION
Christian Johansson
1
and Fredrik Wernstedt
2

1
Blekinge Institute of Technology, PO Box 520, SE-372 25, Ronneby, Sweden, chj@bth.se
2
NODA Intelligent Systems AB, Drottninggatan 5, SE-374 35, Karlshamn, Sweden, fw@noda.se

ABSTRACT
In this paper we investigate the consequences of using
temporary heat load reductions on consumer
substations, from the perspective of the individual
consumer as well as the district heating company. The
reason for using such reductions are normally to save
energy at the consumer side, but the ability to control
the heat load also lie at the core of more complex
control processes such as Demand Side Management
(DMS) and Load Control (LC) within district heating
systems. The purpose of this paper is to study the way
different types of heat load reductions impact on the
energy usage as well as on the indoor climate in the
individual buildings. We have performed a series of
experiments in which we have equipped multi-
apartment buildings with wireless indoor temperature
sensors and a novel type of load control equipment,
which gives us the ability to perform remotely
supervised and coordinated heat load reductions
among these buildings. The results show that a
substantial lowering of the heat load and energy usage
during periods of reductions is possible without
jeopardizing the indoor climate, although we show that
there are differences in the implications when
considering different types of heat load reductions.
INTRODUCTION
The main purpose of this paper is to investigate the
consequences of using temporary heat load reductions
on consumer substations within a district heating
network. The most common way to perform temporary
heat load reductions is to use night time set-back, i.e.
to lower the wanted indoor temperature during night
time while social activity is expected to be low.
Emerging technologies like Demand Side Management
(DMS) and Load Control (LC) also use temporary heat
load reductions in order to accomplish system wide
control strategies, although the characteristic of these
head load reductions differ significantly from night time
set-back.
In the context of this study we regard a heat load
reduction to be the whole process from the initial
change of heat load, through the return to normal heat
load, and until no evidence of the heat load reduction
can be noticed in the dynamics of the building energy
balance. This definition is based on the fact that the
heat load reduction will continue to exert an influence
on the buildings thermal buffer for some time even after
the heat load reduction in itself is ended. The length of
this interval is specific to each building and is related to
the thermal inertia of the building in question.
In this paper we study the consequences of using
different types of heat load reductions, and try to
analyse the way the thermal buffer of the building is
affected along with the actual heat load and energy
usage from both a local and a global perspective. We
study the performance of both long low-intensity heat
load reductions (e.g. night time set-back) as well as
short high-intensity reductions (e.g. those frequently
used in DMS schemes). The use of night time set-back
has received some attention in previous works, e.g [1],
and the possibilities to use the building as a heat buffer
has been evaluated [11], but heat load reductions such
as those used in DSM and LC have to the knowledge
of the authors not been thoroughly investigated.
Night Time Set-back
Night time set-back means to lower the wanted indoor
temperature during night time, with the purpose of
saving energy through reduced heat losses due to
decreased difference between indoor and outdoor
temperature. This is the most common way to perform
temporary heat load reductions, and many commercial
control systems support this feature. This is normally
done by a parallel displacement of the heat control
curve during night hours. During night time set-back the
wanted indoor temperature will be set to one, or a few,
degrees lower than during normal operations. There is,
however, an ongoing debate on whether night time set-
back actually gives an energy saving or not [4], and
most practical implementations of night time set-back
suffer from morning peak loads when the control
system returns to the original operational level. Still,
almost all control equipment companies sell equipment
that facilitates the use of night time set-back, and the
use of this technique is widespread.
Demand Side Management and Load Control
While night time set-backs are a solely local energy
saving technique, DMS and LC are usually performed
with a system wide perspective in mind. A building
September 5
th
to September 7
th

245

owner is normally only interested in lowering the
energy consumption, while the district heating company
is more interested in being able to optimize the whole
production and distribution process. Optimizing the
production normally translates to avoiding expensive
and, more often than not, environmentally unsound
peak load boilers or trying to move heat load demand
in time in order to maximize utility during combined
heat and power generation. Basically, from the
perspective of the district heating company it is a
question of finding a balance between lowering
expensive heat load demand while still selling as much
energy as possible. Implementing this on a system
wide scale requires complex coordination control
strategies that dynamically adapt to the state of the
district heating system [2]. On the local building level
this is implemented by performing temporary heat load
reductions. On a local level these reductions are
normally very short, i.e. one or a few hours, but they
can be of high intensity, even sometimes completely
shutting of the heat load during shorter periods of time.
This behaviour requires the control system to be highly
adaptive in relation to the dynamics of the buildings
thermal inertia in order to avoid jeopardizing the indoor
climate. By coordinating such local heat load
reductions among a large group of buildings it is
possible to achieve system wide DMS and LC.
Previous work
Most previous work regarding temporary heat load
reductions deals with night time set-back. This is a
technique that has been around for a long time, and is
based on the general idea that if you decrease the
difference between the outdoor and indoor temperature
in a building you will save energy. One of the first
large-scale evaluations of night time set-back was
performed in 1983 when buildings in Sweden, USA,
Belgium and Denmark were evaluated. This
experiment concluded that night time set-back did not
save as much energy as was expected, at most a few
percent for multi-apartment buildings [3]. In hindsight it
is possible to see that these meagre results were a
consequence of several interacting factors. First of all
the control systems of the time were not capable of
properly handling the transition from night time set-
back to the original operation mode, which causes a
considerable over-compensation of heat load when the
systems tries to find the new control level. This extra
boost in heat load during the mornings counteracts
large portions of the energy saving done during the
night. The theoretical part of the experiment also had a
few draw-backs, e.g. assuming optimally adjusted
radiator systems and linear relations between indoor
temperature and energy savings. Other articles show
that there is indeed a substantial level of energy saving
to be found by controlling the local heat load [5].
Most of the previous work done on the subject is based
on simulated results. This is expected since the
dynamic thermal processes within a building are
extremely complex and it is not surprising that
comparisons between measurements and calculations
sometimes show large discrepancies. It is noted that
most calculations are dependent on variables that
cannot be measured and verified, and that the building
time constant is really not a constant [6].
EXPERIMENTAL METHOD
In order to study the effects of temporary heat load
reductions we equipped a building with several wireless
temperature sensors in order to measure the
fluctuations in indoor temperature. The building in
questions is an office building with semi-light thermal
characteristics (light construct with concrete slab) and a
time constant of about 150 hours [7]. The indoor
temperature sensors were placed on different locations
within the building in order to get a good overview of
the thermal behaviour of the indoor climate. In addition
to the existing outdoor temperature sensor an extra
wireless sensor was also placed on the outside of the
building. Unlike the existing outdoor temperature
sensor the wireless one was placed in a position were
it was fully exposed to any possible sunshine. This
gave us an extra indication of the impact of free heating
through window areas, even though we did not have
any ability to measure the actual solar irradiance.
In order to control the district heating consumer station
we connected a load control platform for system wide
LC and DSM [8]. This platform is based on a novel
form of hardware and software which enables us to
manage the heat load of the substation without any
major alterations or any damage on the existing
hardware. The software system is based on the open
source Linux operating system and is equipped with an
application programming interface (API) for I/O. This
makes it easy to apply additional sensors, e.g. for
measuring the forward and return temperatures of the
radiator system. The platform also features
connections to a database system which enables real-
time logging and analyse of sensor data. The actual
heat load reductions are implemented by supplying the
existing control system with adjusted outdoor
temperatures, which gives us the ability to manage the
behaviour of the heat load without exchanging any
existing hardware. This adjusted outdoor temperature
can be managed with a resolution of at most 60
seconds. The computer platform uses either Ethernet
or GPRS modems to communicate with the database.
In our case we used the existing Internet access in the
building. In addition to this primary experimental
building we also collected and analysed data from
previously installed buildings using the same basic
computer platform.
September 5
th
to September 7
th

246

Energy and heat load usage was primarily evaluated by
studying the dynamic differences between the forward
and return temperature of the radiator system in
relation to the flow. These readings were then verified
by specifications from the district heating provider
regarding energy consumption and momentary heat
load usage.
Using this set-up we scheduled different types of
temporary heat load reductions and studied their
effects on the measured data. During this study we
studied three primary types of temporary heat load
reductions:
- Long Four to eight hours of continuous heat
load reduction with different intensity
- Short Up to one hour long heat load
reductions with different intensity
- Recurring Several short subsequent heat
load reductions with short pauses in between
When we studied the different types of heat load
reductions we took care in allowing the buildings
thermal process to return to its original state between
each reduction so that the reductions would not
influence each other. This was done in between each
reduction except in those cases when then purpose
was to explicitly study the interaction between
subsequent heat load reductions.
EXPERIMENTAL METHOD
Figure 1 shows the temperature difference between the
forward and return temperature in the radiator circuit
during a short heat load reduction.

Figure 1: dT in radiator circuit with short heat load
reduction
The heat load reduction starts at about 60 minutes and
continues until the 120 minute mark. Between the 120
minute mark and about the 160 mark the control
system performs a controlled heat load recovery in
order to avoid unwanted heat load peaks after the
reduction.
The same values are shown for a long heat load
reduction in Figure 2. The heat load reduction starts
slightly before the 600 minute mark and continues for
several hours until about the 900 minute mark. After
that the control system performs a controlled recovery
in order to return to the original operational state.
Figure 2: dT in radiator circuit with long heat load
reduction

Figure 3 shows the same values for a series of
recurring heat loads.

Figure 3: dT in radiator circuit with recurring heat load
reduction

Each of the heat load reductions in Figure 3 is one hour
long intersected by one hour long recovery periods.
The first reduction starts at the 60 minute mark and
continues until the 120 minute mark.
September 5
th
to September 7
th

247

Figure 4 shows the energy consumption in relation to
the outdoor temperature during week long periods with
and without heat load reductions implemented as LC.
The squares are from periods without LC and the
triangles are from periods with LC. LC in this regard
means that temporary heat load reductions are being
performed in recurring sets throughout the week as
long as the thermal inertia of the building allows it, i.e
without jeopardizing the indoor climate. In this example
the energy usage is about 8.2% lower during periods of
heat load reductions.

Figure 4: Energy usage in relation to outdoor temperature.
The squares are values during periods without LC, and
triangles show periods with LC

Figure 5 shows the heat load (kW) during 24 hours
when using reductions compared to not using
reductions. The control scheme is also added to the
figure in order to show when the reduction was
performed.

Figure 5: Heat load showing 24 hours without reductions
(black), 24 hours with reductions (dark grey) and control
scheme for reductions (light grey)
Figure 5 clearly shows that the reduction in heat load
closely follows the control scheme. The largest heat
load reduction is about 30% in this example.
Figure 6 shows recurring heat load reductions instead
of single long ones. It is clear that the building is able to
respond to the control scheme in this example also.
The largest heat load reduction during the recurring
scheme is about 25%.

Figure 6: Heat load reductions shown 24 hours without
reductions (black), 24 hours with reductions (dark grey)
and control scheme for reductions (light grey)

Figure 7 shows a range of indoor temperature readings
during periods with heat load reduction (triangles) and
during periods without (squares). The average
deviation during heat load reduction is about 0.29 while
the average deviation during periods without reductions
is about 0.19.

Figure 7: Indoor temperature during periods with heat
load reductions (squares) and during periods without
heat load reductions (hourglass)
Figure 8 shows readings from two different outdoor
temperature sensors during a time period of two days.
The graph shows the outdoor temperature sensor
which is connected to the actual consumer sub-station
in the building (black line). Normally these sensors are
placed somewhat in the shadow to avoid large
fluctuations due to solar radiation. We added another
temperature sensor (grey line) in order to estimate the
September 5
th
to September 7
th

248

impact of this solar radiation. Hence this sensor was
placed in full view of the sun. The first day was sunny
during most of the morning until midday, while the
second day was cloudier.

Figure 8: Outdoor temperature sensors placed in the
shade (black line) and in full view of the sun (grey line)
DISCUSSION
When dealing with temporary heat load reductions it is
important to include the whole process of the reduction.
This also includes what happens after the actual heat
load reduction has been performed. For example, when
just restoring the wanted control level after a long
reduction, e.g night time set-back, the forward flow
temperature in the radiator system will rise much faster
than the return flow temperature. This causes a
substantial, although temporary, heat load increase in
the radiator system which negates large portions of the
energy saving done during the actual reduction. Apart
from decreasing the local net energy saving this
behaviour is also less than desired from a system wide
perspective, since it causes massive heat load peaks if
done in many buildings simultaneously, e.g.
contributing to morning peak loads. In order to avoid
this it is important to factor in the whole process of the
reduction, and make sure that the control system
properly handles the transition from the reduction level
to the original level. The inability among most
commercially available control systems to properly
handle this over-compensation is most likely
contributing a great deal to the lingering controversy
whether night time set-back actually gives an energy
saving or not.
It is important to realize that the definition of an
acceptable indoor temperature is not about having the
indoor temperature at a certain precise level at all time,
but rather to have it within a certain, socially
acceptable, temperature interval at all time. This has
been discussed at great length in previous work [6].
The general idea is that a greater temperature interval
will lower the need of additional heating from the
radiator system, by coordinating the thermal inertia of
the building with freely available heat, e.g. heat from
sunlight or electrical appliances, to balance the heating
need. This notion is supported by our results as we
have shown that the thermal inertia of even a small or
medium sized multi-apartment building is considerable.
How people perceive the indoor climate is dependant
not only on the actual indoor temperature itself but also
on other factors like air quality, individual metabolism
and behaviour, radiation temperature and air
movement. In relation to this it can be noted that
previous work have shown that about five percent of
any group of people will always be unsatisfied by the
indoor climate [9], and that it is not possible to create a
perfect climate that will make everyone happy.
CONCLUSIONS
There is an ongoing debate whether night time set-
backs lead to an energy reduction or not. Results from
this study clearly show an energy saving in relation to
heat load reductions, although this assumes that the
control system is able to smoothly handle the transition
from reduction to normal operation. The results
showing energy saving is evaluated in relation to the
total energy usage which also includes tap-water
usage. Normally this is estimated to about 30% of the
total energy use in a multi-apartment building.
In prior studies of temporary heat load reductions the
focus has been on the fluctuations in the indoor
temperature as a way of evaluating the energy saving
[3]. This idea is based on the widespread notion that
any energy saving is linearly proportional to the
temperature difference between the indoor and outdoor
temperature. This model might be true in a steady state
simulation where the temperature difference is
assumed to have had time to permeate the air mass as
well as the entire building structure, but it is obviously
inadequate in a dynamic situation. We have instead
focused on the heat load and energy usage directly, i.e.
the difference between forward and return temperature
in relation to the flow within the radiator circuit. In most
of the buildings evaluated there has been a
considerable reduction of energy consumption without
any noticeable change in indoor temperature. The
reason that there does not need to be a measurable
change of the indoor temperature is due to the
dynamics of the thermal inertia of the building, e.g. the
time constant of a building is not a constant [6]. This
aspect comes into play when using very short heat load
reductions, at most one or a few hours long. During this
first part of the reduction it is mainly the actual air mass
that is influencing the indoor temperature drop since
this body has a low resistance to change, i.e. the short
time constant [10]. If the heat load reduction is
prolonged, like during a night time set-back, the
September 5
th
to September 7
th

249

building mass will start to interact with the air mass and
thus stabilizing the continuing temperature drop, i.e. the
long time constant [10].
The influence of external and internal free heat is large
enough that when these heat sources interact with
other parts of the thermal process it hides shorter heat
load reductions in the ambient temperature. This can
be seen in Figure 7 where it is shown that although the
average indoor temperature is not noticeably affected
there is still a somewhat larger deviation in the indoor
temperature which implies that there is indeed a higher
level of temperature flux within the air mass and that
this is triggered by the heat load reductions. The
control policies used during this work obviously set a
high bar for the control system to handle, but as the
average hardware develops it should be possible to
implement such techniques on a larger scale.
Figure 8 gives another clear indication of just how
substantial such sources of free energy can be. This
extra heating due to solar radiation through the
windows directly interacts with the mass of air inside
the building, thus raising the temperature.
In addition to being able to help save energy usage in a
building temporary heat load reductions also form the
backbone of DSM and LC, in which the goal is to
manage the heat load (kW) rather than the energy
usage (kWh).
FUTURE WORK
In the future we plan to further develop models in order
to dynamically estimate the temperature flux within
buildings and develop theoretical and practical
interfaces for incorporating this data dynamically into
the control systems.
ACKNOWLEDGEMENT
This work has been financed by Blekinge Institute of
Technology and NODA Intelligent Systems AB.
REFERENCES
[1] N. Bjrsell, Control strategies for heating systems,
University-College of Gvle-Sandviken.
[2] F. Wernstedt, P. Davdisson and C. Johansson,
Demand Side Management in District Heating
Systems, in Proc. Of Sixth International
Conference on Autonomous Agents and Multiagent
Systems, Honolulu, Hawaii, USA, 2007.
[3] L. Jensen. Nattsnkning av temperatur I flerbo-
stadshus, R64:1983, Byggforskningsrdet, 1983
(In Swedish).
[4] H. Lindkvist and H. Wallentun. Utvrdering av nio
fjrrvrmecentraler i Slagsta Report ZW 04/05,
ZW Energiteknik, 2004 (In Swedish)
[5] F.B. Morris, J.E. Braun and S.J. Treado Experi-
mental and simulated performance of optimal
control of building thermal storage, ASHRAE
Transactions, Vol. 100, No. 1, 1994
[6] E. Isflt and G. Brms. Effekt- och energi-
besparing genom frenklad styrning och drift av
installationssystem I byggnader, ISRN KTH/IT/M--
22--E. Institutionen fr Installationsteknik. Kungliga
Tekniska Hgskolan, 1992. (In Swedish)
[7] S. Ruud. Energimyndighetens program fr
passivhus och lgenergihus Remissversion 2009-
03-10. Forum fr Energieffektiva byggnader, 2009.
(In Swedish)
[8] F. Wernstedt and C. Johansson. Demonstrations-
projekt inom effekt och laststyrning. ISBN 978-91-
7381-041-8, The Swedish District Heating
Association, 2009. (In Swedish)
[9] J. Skoog, PM avseende komfort, F-Infrastruktur
AB, 2005. (In Swedish)
[10] C. Norberg. Direktverkande elradiatorers reglering
och konstruktion Vattenfall Utveckling AB, Rapport
nr F-90:5, lvkarleby, 1990. (In Swedish)
[11] L. Olsson Ingvarsson, S. Werner. Building mass
used as short term heat storage in Proceedings of
The 11th International Symposium on District
Heating and Cooling. Reykjavik, Iceland, 2008.

September 5
th
to September 7
th

250

VERIFICATION OF HEAT LOSS MEASUREMENTS
J.T. van Wijnkoop
1
, E. van der Ven
2

1
Liandon B.V,
2

ABSTRACT
Heat loss tests are performed on different samples of
the Thermaflex Flexalen 600 series and one
ST-PUR-PE sample at the Thermaflex heatloss
equipment and two German test facilities. At these
facilities two different testing methods are used. These
methods are both described in the European standard
[1] but show significant differences in the results. In this
paper the different methods of testing are described.
Furthermore the Thermaflex heat loss equipment is
verified with the test institute that uses the same testing
method.
INTRODUCTION
Last year Liandon developed a test-rig for Thermaflex
to measure heat loss of insulated plastic piping
systems. With this test-rig it is possible for Thermaflex
to test the in house produced pre-insulated, semi
flexible pipes in various diameters.
To verify the test results, the results of the Thermaflex
heat loss equipment are compared with the test results
of two acknowledged institutions. For this paper two
German institutions are chosen, since they both
measure in compliance with the European standard EN
15632 [1], however with different methods described in
this paper. In order to give an appropriate comparison,
knowledge of the testing methods of both systems is
required. In this paper the testing methods of all three
systems is covered, together with the comparison of
the test-results. Since the testing facilities use two
different methods described in the standard, the
comparison refers to the test methods and the test
results.
The objective of this paper is to compare the test
methods and test results of the two different test
institutes with the Thermaflex heat loss equipment and
verify the outcome. As in Heat loss of flexible plastic
pipe systems analysis and optimization
(E. Van der Ven et Al.) [4] and Performance of pre
insulated pipes (I. Smits et Al.) [6] these results are
used to compare different sizes of the Flexalen 600
series and competitive products.
The Thermaflex test-rig is newly developed for the
research of heat loss of pre-insulated pipes. The
novelty of this system is its ability to measure the
overall heat loss of different samples under similar
conditions, as defined by the European standard [1].
The ability to conduct equally based heat loss
measurement result in an objective comparison of
different types of (semi) flexible piping systems,
providing the opportunity to highlight strengths and
weaknesses of (competitive) piping systems.
Furthermore, in contradiction to most heat loss tests,
the test time in the Thermaflex test-rig is only a few
hours so the test can be performed during production.
This provides the opportunity to optimize the production
process real-time and measure the heat loss of the
product several times during a production run. This
guarantees the quality of the produced batch.
In addition the handling of the equipment is made easy,
so no specially trained staff is needed for testing,
making it possibly for operators to carry out the tests.
EUROPEAN STANDARD METHOD DESCRIPTION
The European standard EN 15632 [1] allows two
different methods of heat loss or thermal conduction
testing. These methods both state the same on internal
heating of the service pipe but vary on the method of
compensation for heat loss in axial direction.
The first method, the guarded end method, states no
axial heat transfer is permitted. This should be
accomplished by the use of end guards, an extra pair
of heating elements at both ends of the service pipe as
shown in Fig 1. By heating the ends separately to the
same temperature as the middle test section no heat
transfer will take place to the ends of the service pipe.
In this case a theoretical compensation is not required
since the test section only has losses in radial direction.
This method is used in the Thermaflex heat loss
equipment and at one of the institutes.

Fig 1, Longitudinal section guarded end heating probe

The second method described for compensating for
axial heat loss is the calibrated or calculated end
method. The calibrated end method will not be covered
in this paper since it is not used in our comparisons.
September 5
th
to September 7
th

251

The calculated end method states the ends of the
service pipe shall be insulated with a known thermal
conductivity as shown in Fig 2.

Fig 2, Configuration calculated end cap.

The service pipe is heated, using a heating element
with only one section. During the tests a thermal profile
is made of the outer casing of the sample, showing
lower values at the ends. After testing the heat loss is
compensated for the end loss with the van Rinsum or
Nukiyama theory. For this investigation only the van
Rinsum theory is used and therefore described.
According to the van Rinsum theory, the axial heat loss
causes a decrease in temperature not only towards the
ends of the service pipe, but in the test section as well.
With the use of the equations (1), (2), (3) this
temperature decrease in the test section can be
calculated and added to the measured value,
compensating the end loss. This corrected temperature
is used in equation (4) to calculate the overall thermal
conductivity. This method is used by one of the
German institutes.
calc
( )
o ln
D
2
D
0
|
\
|
|
.
2 t L T
0m
T
2
( )
:=
(1)

c
2 t
calc
A
1

1
A
2

2
+
( )
ln
D
2
D
0
|
\
|
|
.
:=
(2)

AT
0m
T
0m
T
0X
cosh X c ( )
:=
(3)

o ln
D
2
D
0
|
\
|
|
.
2 t L T
0m
AT
0m
+ T
2
( )
:=
(4)

calc
: approximate value of thermal conductivity
D
2
/D
0
: outer/inner diameters of casing and service pipe
A
1
, A
2
: areas of the heating probe, inner service pipe
1
,
2
, : thermal conductivity of heating probe, thermal
conductivity of medium in the service pipe, thermal
conductivity total sample.
X, L: distance to next measuring point from the middle,
sample length
T
0m
,T
0X
,T
0m
,T
2
: pipe temperature at the middle of the
test section, temperature at distance X of the middle,
temperature correction, temperature at insulation
surface.
VERIFICATION OF SAMPLES
To verify the outcome of the Thermaflex heat loss
equipment and the laboratory tests, three samples of
the Flexalen 600 piping system are tested on their
overall heat loss. These samples consist of 2 or 3 m of
the pre-insulated piping system. More information
about the Flexalen 600 system can be found in Heat
loss of flexible plastic pipe systems analysis and
optimization (E. van der Ven et Al.) [4]. Furthermore,
method comparison tests are performed on competitive
pre-insulated piping systems, a comparison of the
products themselves is given in Performance of pre
insulated pipes (I. Smits et Al.) [6].
The tests on the Flexalen 600 products are performed
by Thermaflex and by one of the acknowledged
institutes, using the different methods. To ensure the
effect of ageing in the Flexalen 600 system is the same
during all tests, the Flexalen 600 samples are tested
simultaneously. To exclude effects of the production
process both tested samples are half of a 6 meter stick.
An alternative method is used for Flexalen 50A25 and
competitive products. Here the same sample is tested
at the different test facilities.
The comparison of the results is based on the outcome
of heat loss per meter, calculated as described in the
European standard [1]. This loss per meter is only
conclusive on a very small part of the entire system.
Therefore the complete Flexalen 600 system will be
covered in paper Heat loss system optimisation (J.
Korsman et Al.) [3] and New economical connection
solutions for flexible piping systems (C. Engel et Al.)
[5].
In this report the following diameters of the Flexalen
600 piping systems are used for comparison of the
measurements:
- Flexalen 600:
- 50A25, two guarded end tests* and calculated end
test.
- 160A90, one guarded end test* and calculated end
test.
- 200A110, one guarded end test* and calculated
end test.
Competitive products:
- Sample 1 two guarded end tests
- Sample 2 two guarded end tests*
*At the time of writing the second test results were not
yet available.
September 5
th
to September 7
th

252

THERMAFLEX HEAT LOSS EQUIPMENT 600
The Thermaflex heat loss equipment is specially
designed for the Thermaflex Flexalen 600 series. One
of the major design goals was to develop a fast and
easy to use test rig with the precision of a laboratorial
test. These goals have resulted in a test rig that is able
to measure heat loss in a few hours, allowing direct
optimization during the production process, and is
operable by the production staff without the loss of
accuracy
Physical test facility
The physical part of the Thermaflex heat loss
equipment consists of three segments.
The first is the water cooled compartment in which all
tests are performed. This compartment is kept at a
constant temperature, (23 C), during each
measurement.
The second is a heat source, for which heating probes
are used. These heating probes are custom made by
equipping a two meter Thermaflex piping segment, of
all available diameters, with three heating coils.
The third part of the heat loss equipment is the control
unit. Here the heating probe is powered and all thermal
readings are done. By applying custom made software
all desired readings can be done. The final output is the
actual heat loss in W/m through the entire pre-insulated
Flexalen pipe, consisting of the service pipe, insulation
and outer casing.

Fig 3, Thermaflex heat loss equipment 600.
Measurement principle Thermaflex
The Thermaflex test rig is designed in compliance with
the European standard [1] and also the tests are
carried out according to ISO 8497 and EN 15632. In
the design of the heating probes the most realistic
method, the guarded end method, is used. According
to this method the heating probes are equipped with
three heating coils with separate power supply. As
shown in Figure 2, two 400 mm heating coils located at
each end of the 1000 mm test section. These two
sections provide a thermal insulation at both ends of
the test section since all three are kept under uniform
temperature, eliminating axial heat loss of the test
section. With this method it is possible to measure the
heat loss by measuring the power needed to maintain a
constant temperature of the test sample. In
contradiction to the measurements at the test institutes,
the Thermaflex heating probes temperature is
regulated by PID controlled power supplies. In the test
results Graph 2 the power consumption versus test
time is shown. This variable power supply makes it
possible to pre-heat the probes in a short period of
time, shortening waiting times considerably.
Furthermore the use of the actual pipe material as a
heating probe increases the accuracy. Moreover it
eliminates all additional heat loss by convection that
will be present with the use of smaller, not inner service
pipe connecting heating probes.
For testing competitive products with different
diameters these advantages are lost. However by the
use of thermal compartments in the service pipe the
test results can be guaranteed.
Thermaflex method of testing
For testing, the heating probe with the appropriate
diameter is inserted in the insulation covered with outer
casing, and inserted in the cooled test section. After
connecting the probe to the control unit the
measurement can be started. Different testing
conditions can be entered at this point such as the
inner pipe temperature, representing the internal
medium. When the test is started the heating coils heat
the inner side of the probe until the desired
temperature is reached. When the inner temperature is
considered constant and uniform throughout the three
heating coils, the actual measurement is started. To
ensure a constant temperature in the probe, a waiting
time is built in the software that will reset the
measurement if temperature exceeds preset
temperature values.
The heat loss measurement is done by measuring the
energy required to keep the probe at a constant
temperature, by measuring the current at constant
voltage in the heating coils, and calculating the power
consumption. Since the middle/testing coil is exactly
one meter in length the required energy represents the
exact heat loss through one meter of piping and
insulation in W/m. Since the actual piping material is
used during the measurement, there are no other
losses, nor advantages, than there will be in practise,
ensuring an objective measurement. Furthermore a
realistic fit of the insulation material is guaranteed. As
stated in the foregoing paragraph these advantages
are lost for divergent diameters. However during this
investigation the probes have proven suitable for
testing, as both testing institutes also use smaller
heating probes.
September 5
th
to September 7
th

253

Test results Thermaflex heat loss equipment
In this paragraph the test results are presented for the
tests carried out with the Thermaflex heat loss
equipment. For this study four different types of the
Flexalen 600 series were tested. The tests for the
Flexalen 600 series took place at three different
temperatures, 60, 70 and 80 C. The values at lower
temperatures are calculated using the linearization
method described in the European standard EN 15632
[1]. In the following tables and graphs the test results of
the Thermaflex heat loss equipment are presented.
The power usage during the testing cycle is shown in
Graph 2. In this graph the first 40 minutes represent the
heating and stabilisation time for the heating probe and
insulation, whereas the last 30 minutes is the actual
test time. Since, as the figure shows, the temperature
is constant, the power usage equals the heat loss
through the piping system in radial direction during the
last 30 minutes. The results, as given in Table 1, are
calculated by using the mean of the power
consumption during the last 30 minutes of the heat loss
test. The results in Table 1, are also displayed in
Graph 1 for the three tested samples.
Table 1, Results heat loss equipment for the Flexalen 600
products
Heat loss of the Flexalen 600 series in W/m tested on
the Thermaflex heat loss equipment
Product 40 C 50 C 60 C 70 C 80 C
50A25 3.6 6.4 9.3 12.0 15.0
160A90 6.2 10.1 14.0 17.9 21.8
200A110 6.5 12.0 17.5 23.0 28.5

40 50 60 70 80
0
10
20
30
Flexalen 50A25
Flexalen 160A90
Flexalen 200A110
Heat loss results thermaflex heat loss equipment
Temperat ure inner service pipe [C]
H
e
a
t

l
o
s
s

[
W
/
m
]

Graph 1, Results heat loss equipment Flexalen 600
products.

Graph 2, Power and temperature of the Thermaflex
heating probe.
Outcome Competitive products for comparison of
testing method:
For the comparison with test institute two, two samples
of competitive products are tested. As these samples
are ST-PUR-PE system, a correction has been made
for using the PB heating probe using the Wallentn [2]
method. First the thermal conductivity of the insulation
is determent by the use of equation (5), hereafter the
heat loss is recalculated without the heating probe,
using the temperature of the inner service pipe in
equation (6). The results are presented in Table 2 and
Graph 3.
i
ln
d
3
d
2
|
\
|
|
|
.
2t
T
p
T
c
( )
o
probe
st
ln
d
2
d
1
|
\
|
|
|
.

1
p
ln
d
6
d
5
|
\
|
|
|
.

1
c
ln
d
4
d
3
|
\
|
|
|
.

:=

W
m K
(5)

o
corrected
2t T
st
T
c
( )
st
ln
d
2
d
1
|
\
|
|
|
.
i
ln
d
3
d
2
|
\
|
|
|
.
+
1
c
ln
d
4
d
3
|
\
|
|
|
.
+
:=

W
m
(6)

Where:
T
p
,T
c
, T
st
=Probe, Casing and Steel pipe temperature
d
1
to d
6
= inner/outer diameters of service pipe, casing
and heatingprobe
st
,
i
,
c
,
p
= heat coefficient of service pipe, insulation,
casing and probe
probe
,
corrected
=probe power and corrected heat loss.

September 5
th
to September 7
th

254

Table 2, Results heat loss equipment for the competitive
products
Heat loss of competitive products in W/m tested on
the Thermaflex heat loss equipment
Sample 40 C 50 C 60 C 70 C 80 C
Sample 1 2.5 4.5 6.4 8.4 10.3
Sample 2 6.5 8.8 11.2 13.5 15.9
Sample 3

40 50 60 70 80
0
5
10
15
20
Competitive sample 1
Heat loss results thermaflex heat loss equipment
H
e
a
t

l
o
s
s

[
W
/
m
]

Graph 3, Results heat loss equipment for the competitive
products
TEST INSTITUTE ONE
This institute is specialized in measuring heat loss in
different types of insulation. The test facility used for
the Flexalen 600 system is specially designed for
measuring the heat loss of (pre-) insulated piping
systems. This means the facility is designed to
measure all different types and diameters.
Measurement principle institute one
The measurements are all based on the calculated end
apparatus, using the van Rinsum theory as correction,
as described in the paragraph European standard [1]
method description of this paper.
The physical part of the test facility is similar to the
Thermaflex test rig and also consists out of the three
elements: A temperature controlled compartment
where the tests are carried out at a constant
temperature of 23 C. FIW also uses heating probes as
a heat source but, since it is not specially designed for
the Flexalen 600 system, they are made to fit all
systems. To ensure the fit of the probes in all different
systems the diameters are smaller, and for durability
made of metal. Furthermore no heat guards are used.
This means the outer ends of the piping system are
insulated and the heat loss is corrected with a
calculated value. In the paragraph European standard
method description a more detailed description is
given. As can be seen in Fig 4 the heat distribution in
this case is not uniform along the test specimen,
proving the need for the van Rinsum correction.

Fig 4, Thermal image of the sample at institute one

In contradiction to the Thermaflex test rig, no integrated
computer controlled power supply system is used. The
power for the heating probe is first theoretically
calculated and manually set to this value. For the
temperature measurement thermocouples and a data
logger with computer link are used.
Method of testing
The heating probe is positioned in the centre of the test
pipe with positioning foam in three sections of the pipe.
On these foam blocks four thermocouples are placed in
0, 90, 180 and 270 degrees on the inner surface of the
service pipe. For the outcome of the pipe inner
temperature the mean of the four values is used. To
measure the temperature on the outside casing of the
insulation, five groups of four thermocouples are used
in the same configuration as the inner pipe. The
difference being that the thermocouples are placed
both on and in between the corrugations of the casing.
The test sample, with the heating probe, is placed in
the conditioned container thereafter the test can be
started. The power supply of the heater is turned on by
setting the voltage and current of the power unit to a
fixed value so the electrical power equals the
calculated heat loss.
Depending on the diameter of the test sample and the
test temperature the waiting time for the heating of the
sample is five to eight hours due to the low, fixed power
input. After a constant temperature of the outer casing
is achieved the actual test cycles start. Each test cycle
consists of a measurement of 30 min in which the outer
casing temperature is to be constant. If not the cycle
has to be restarted. In total ten cycles will be performed
on each sample. After the test the values are corrected
September 5
th
to September 7
th

255

for axial heat loss and the thermal conductivity, thermal
resistance and overall heat loss are calculated.
Test results institute one
As the actual measurement data are not available due
to the correction factor, only the calculated values can
be discussed in this paragraph. As soon as the actual
measurements become available this section will be
updated. Furthermore the results are not given for
exactly 60, 70 and 80 C due to the fixed power supply
with no temperature set point, the displayed results are
calculated heat loss values at the set temperatures to
make the data more interpretive. For this calculation
the linearization method described in the European
standard [1] is used. In Graph 4 the data from Table 3
is presented as a graph.
Table 3, Results test institute one for the Flexalen 600
products
Heat loss of the Flexalen 600 series in W/m tested at
test institute one
Product 40 C 50 C 60 C 70 C 80 C
50A25 5.6 8,8 11.9 15.1 18.3
160A90 9.1 15.1 21.1 27.1 33.0
200A110 9.8 15.1 20.5 25.8 31.2

40 50 60 70 80
0
10
20
30
40
Flexalen 50A25
Flexalen 160A90
Flexalen 200A110
Heat loss results test institute one
H
e
a
t

l
o
s
s

[
W
/
m
]

Graph 4, Results test institute one for the Flexalen 600
products
TEST INSTITUTE TWO
For the second test institute in this research, an
institute using the same guarded end method is
chosen. This makes it possible to provide a correct
comparison between the test results and not only the
testing method. The tests carried out by test institute
two at the time of writing are of competitive products
only as the facility was already running on full capacity.
An update to this paper will be made as soon as the
Flexalen 600 results will become available.
Testing method institute two
The method used by this institute is generally the same
as the method used by Thermaflex; however the test
facility itself is different.
The testing facility at institute two consists of a
temperature controlled room, kept at the prescribed
23 C. As a heat source a heating probe, consisting of
a 2 m test section and two 50 cm end guards is used.
At the time of writing no further information on the test
facility was available. This paragraph will be updated
when this information becomes available.
Method of testing
Prior to testing, the sample is prepared by placing
thermocouples in various locations on the inner service
pipe and outer casing. Subsequently the sample is
placed in the temperature controlled room and the
heating probe is inserted. By setting the power supply
to a calculated value for all three heating coils the
heating process of the sample is started. Because of
the low fixed value of the power supply, this heating will
take approximately 5 to 8 hours. After the desired
temperature is reached at the test section as well as at
the guarded ends, the actual test is performed. The test
consists of a power reading during a 30 min cycle
where het temperature of the test section and guarded
ends may not exceed the limit of an yet unknown
bandwidth.
Test results test institute two
The test results of institute two are given in Table 4 and
Graph 5. As not all data was available during writing
there tables and graphs will be updated.

Table 4, Results test institute two for the competitive
products

Heat loss of competitive products in W/m
tested at test institute two
Sample 40 C 50 C 60 C 70 C 80 C
90DN25 1.95 4.00 6.02 8.06 10.09

September 5
th
to September 7
th

256

40 50 60 70 80
0
5
10
15
Heat loss results test institute two
H
e
a
t

l
o
s
s

[
W
/
m
]

Graph 5, Results test institute two for the competitive
products
COMPARISON OF THE TEST RESULTS
Comparison of the Thermaflex flexalen 600 series:
Although both methods, guarded end and calculated
end, are approved and described in the European
standard [1], the difference between the results is
substantial as displayed in Graph 6. Moreover all
results vary more as the temperature difference
increases. This can be explained by the use of the
calculated end caps that conduct more energy at
higher temperature differences. As these end cap
losses increase, the corrected thermal conduction for
the sample also increases, resulting in a higher
calculated heat-loss.
40 50 60 70 80
0
5
10
15
20
25
30
35
Flexalen 50A25 Thermaflexresult
Flexalen 50A25 Inst itute one result
Heat loss comparison Thermaflex and test institute one
H
e
a
t

l
o
s
s

[
W
/
m
]

Graph 6, Comparison results of the heat loss equipment
and test institute one
Comparison of competitive products:
Although the Thermaflex heat loss equipment was
designed for Flexalen series, test results on
competitive products show consistency with test
institute two as shown in Graph 7. The heat loss
equipment values are just a little higher, which can be
explained by the need to cut the sample in order to
place the heating probe with thermocouples in the right
position. The difference between the outcome of the
test on sample 1 are 0.39 and 0.22 W/m at an inner
service pipe temperature of 60 and 80 C respectively.
These values are within the combined accuracy range
of both facilities. This comparison, although only based
on one test, proves the worthiness of the Thermaflex
heat loss equipment and will be updated as more data
comes available.
40 50 60 70 80
0
5
10
15
Heat loss comparison Thermaflex and test institute two
H
e
a
t

l
o
s
s

[
W
/
m
]

Graph 7, Comparison results of the heat loss equipment
and test institute two
CONCLUSION
During this research it has become clear that the
European standard [1] tolerates differences in heat loss
values by allowing different testing methods. The
outcome of the tests indicate that the result of the
guarded end cap method varies from the result of the
calculated end cap method, however no assumptions
can be made based on only one comparing
measurement. Further study that is being conducted at
this moment will provide more comparison data. This
will be updated with this data as soon as becomes
available. This new data could point out that the van
Rinsum theory is not suitable for accurate heat loss
measurement of plastic piping systems.
The comparison of the guarded end method results
from test institute two and the Thermaflex heat loss
equipment conclude that the results of the heat loss
equipment are correct and comply with the European
standard [1]. This validation makes the results of the
Thermaflex heat loss equipment valid for not only in
house testing but also for publication as done in Heat
loss of flexible plastic pipe systems analysis and
optimization (E. van der Ven et Al.) [4] and Perfor-
mance of pre insulated pipes (I. Smits et Al.) [6].
September 5
th
to September 7
th

257

FURTHER INFORMATION
Veerweg 1
5145NS Waalwijk
The Netherlands
Liandon B.V.
Dijkgraaf 4
6920AB Duiven
The Netherlands
ACKNOWLEDGEMENT
Acknowledgments go to both the test institutes for their
open and honest explanation of their testing methods
and facilities and for even showing the entire facility
and methods.
Furthermore acknowledgements go to all involved
employees of Thermaflex Isolatie B.V. and Liandon
B.V. who made this research possible. Special
acknowledgements go to P. Blom and P. van Rijswijk
for their devotion on all the heat loss measurements
they performed during this research in a short amount
of time.
REFERENCES
[1] NEN-EN 15632 and NEN-EN-ISO 8497
[2] P. Wallentn, steady-state heat loss from
insulated pipes, Lund Institute of Technology,
Sweden, 1991
[3] J. Korsman and G. Baars, Heat loss system
optimization, 12th ISDHC 2010
[4] E. van der Ven and R. van Arendonk, Heat loss
analysis and optimization, 12th ISDHC 2010
connection solution for flexible piping systems,
12th ISDHC 2010.
[6] I. Smits and E van der Ven, Performance of pre
insulated pipes, 12th ISDHC 2010.
September 5
th
to September 7
th

258

DISTRICT HEATING AND COOLING WITH LARGE CENTRIFUGAL CHILLER-HEAT
PUMPS
Ulrich Pietrucha
Friotherm AG, Switzerland

ABSTRACT
With prices for primary energy resources soaring, the
recovery of "waste energy" was getting into the focus of
attention within the last years. Also the global climate
change reminded us to limit the use of primary energy
resources to a minimum, thus exploiting "waste energy"
potentials wherever feasible. The process of upgrading
low grade waste heat is especially interesting where
large amounts of such energy are available at one
point, e.g. next to sewage water treatment plants,
alongside main sewers, in power plants or close to
ground water sources.
Even if the "waste energy" potential is abundant and
easily exploitable, the aspect of overall thermal
efficiency is considered crucial for the final decision to
invest in large heat recovery installations.
INTRODUCTION
Described are five applications of large centrifugal heat
pumps-chillers for the use in large district
heating/cooling systems.
Application 1: Heat recovery from raw sewage water
and hot water production at 90 C.
Application 2: Combined heating and cooling with a
raw sewage water heat pump/chiller installation. This
plant is operated successful since 1989.
Application 3: Combined heating and cooling: a
combination of cooling with simultaneously heat
production in summer and heat recovery from cleaned
sewage water in winter.
Application 4: Heat recovery from wet flue gas
cleaning process
Application 5: Combined heating and cooling in
Stockholm
1. SKOYEN VEST PLANT IN OSLO: HOT WATER
PRODUCTION AT 90 C
This is the world's largest heat pump plant using raw
waste water as heat source. It is installed in a cavern
alongside one of the main waste water channels in
Oslo. With 2 heat pumps a heating capacity of 27'600
kW is generated by recovering heat from raw waste
water.
Number of units 2
Type UNITOP 50 FY/34FY
Refrigerant R134a
Heat source medium Raw waste water
Raw sewage water inlet 10.0 C ... ~ 15 C
Raw sewage flow water flow 3800 m3/h
Heating water temp. in/out 60 / 90 C
Heating water flow 824 m3/h
Power at terminal 9750 kW
Heat capacity 27600 kW
Coefficient of performance 2.83 up to >3.0

One of the Skoyen heat pumps

2. THE SANDVIKA PLANT IN OSLO: COMBINED
HEATING AND COOLING FROM A RAW SEWAGE
WATER HEAT PUMP
This is the oldest combined chiller/ heat pump
installation in the world, producing simultaneously
cooling, taking out heat from raw sewage water and
producing heating capacity for the district heating
system.
The heat pumps are in successful operation since 1989
and each one has an additional heat exchanger, which
is used either as raw sewage evaporator or as raw
sewage water condenser.
Each heat pump has an overall operating time of about
160'000 hours, means the heat pumps were operated
since 1989 each year for more than 8'400 hours.
A 3rd larger heat pump was taken into operation in
2008.
September 5
th
to September 7
th

259

Combined chiller / heat pump at Sandvika plant

3. THE KATRI VALA PLANT IN HELSINKI:
COMBINED HEATING AND COOLING
This is the largest combined chiller heat pump
installation in the world producing simultaneously
60MWth cooling and 90MWth heating, i.e. total
produced thermal energy is 150MW. The required
electrical input is 30MW i.e. a superb COP of 5 can be
achieved (150MW / 30MW).
During Winter season the required cooling is done by
sea water, while heat is produced by using cleaned
waste water as heat source.

Typical Unitop 50FY heat pump (Qheat 15 to 23MW)

4. SYSAV MALM: HEAT RECOVERY FROM WET
FLUE GAS CLEANING PROCESS
SYSAV Malm in Sweden has built a new waste-to-
energy plant. An important part in this plant was the
installation of a 19MW heat pump using the flue gas
condensation as heat source. The heat pump is
supplying hot water with a temperature of up to 70 C
to the district heating system of the community of
Malm.
The two heat pumps are connected in series on the
heat source side and on the heat sink side; this
improves considerably the COP. There are operating
points with lower district heating temperatures where a
COP of up to 6.5 can be achieved.

5. NIMROD STOCKHOLM: COMBINED HEATING
AND COOLING
Due to the fact that with every cooling process there is
also waste heat generated, Friotherm AG, which has
worked since many years on chillers with heat
recovery, has worked out a concept which allows
various operating modes in order to operate the chiller /
heat pumps more efficient over a longer period and,
making therefore the investment more attractive:
There are 4 chiller / heat pumps installed in the Nimrod
plant. The centrifugal compressors are switched in
parallel for Summer cooling production of 48MW.
However during this period heat recovery is not
required as there is sufficient capacity available from
the existing heat pumps.
The same units are producing during Spring, Autumn
and Winter a cooling capacity 24MW with a full heat
recovery of 35.6MW at a temperature level of 78 C.
For heat recovery operation mode the centrifugal
compressors are switched in series.
Each chiller / heat pump consists of two centrifugal
compressors Type Uniturbo 33CX and 28CX and is
able to operate at the following modes, described
below:
5.1 Cooling only:
During Summer with high cooling demand, the waste
heat from the condenser is removed with sea water of
Generator

2 Friotherm heat pumps
Type UNITOP 28CX-71210U

34.2C
24.3C
50C
59.2C
Flue-gas
Steam
- turbin
e
Boiler
From district-
heating-network
Heat source capacity
15.5MW
Power consumption
3.5MW
Heat sink capacity
19MW

Waste-to-Energy plant SYSAV Malm Sweden
Flue-gas-
cleaning

Flue- gas-
condensing
September 5
th
to September 7
th

260

max. 22 C, therefore the condenser and sub cooler
are equipped with Titanium tubes.
The two compressors Uniturbo 33CX and 28CX are
then working in parallel, in a single stage mode, with a
single stage expansion, producing a cooling capacity of
up to 7MW plus 5MW = 12MW i.e. with 4 units a total of
48MW.
Depending on the cooling demand, one or the other, or
both compressors can be put in operation.
If needed, the part load of each chiller / heat pump can
be controlled down to 10% of its nominal capacity, with
a reasonable high efficiency, with the use of inlet guide
vanes. The chilled water temperature outlet is kept
constant to 5 C
5.2 Combination of cooling and heating:
During Spring, Autumn and Winter, with moderate
cooling demand of up to 24MW, but simultaneous need
of heating, the waste heat from the condenser is
supplied to the district heating network at a
temperature outlet of 78 C and a maximum heat
capacity of 35.6MW.
The two compressors Uniturbo 33CX and 28CX are
then working in series in two stage compression mode,
with two stage expansion using an economiser after
the first stage expansion.
The compressor Type Uniturbo 33CX with the larger
volume flow is working as 1st stage and the Type
Uniturbo 28CX with the smaller volume flow as 2nd
stage compressor.
The control system is controlling the required cooling
capacity; the surplus heat is supplied fully to the district
heating network at a temperature level of up to 78 C.
I.e. this operation mode delivers heat which can be
sold in addition to the cooling, with a total COP of
above 5.
The cooling only mode and the combination of heating
and cooling mode are explained in the below P&Is:
CONCLUSION
Reliability of technology, future developments and
challenges
About more than 140 heat pumps, producing hot water
with temperatures above 70 C, are installed world
wide since 1980. The heat pump plants described in
this article are only showing a small part of the
nowadays available applications.
Almost all of the installed heat pumps plants starting
from the early 1980's are today still in operation, which
is showing the high reliability of this technology.

Today developments are the extension of the heat
pump operation range in temperature and capacity to
exploit new heat sources and to extent the field of
applications.
The adaptation of the centrifugal heat pumps to new
refrigerants with GWP close to zero is already on the
way.
REFERENCES
Text and pictures from Friotherm AG / Switzerland.
2
8
8

Cooling capacity 12 MW
Nimrod single stage operation
2 compressors in parallel
Sea water

Heating capacity 9 MW
Cooling capacity 6 MW
28
33
Nimrod two stage operation
2 compressors in series
Heating water
September 5
th
to September 7
th

261

NEW ECONOMICAL CONNECTION SOLUTION FOR FLEXIBLE PIPING SYSTEMS
Christian Engel, Gerrit-Jan Baars


ABSTRACT
Most Energy Companies are facing the same problem:
Connection costs per house shall be cheaper and
faster to install to reach more customers. At the same
time high level of durability and a system free of
maintenance must be guaranteed.
This paper shall give an insight of the practical
experience with new solutions showing the economic
and ecological advantages in projects with several
Energy companies.
INTRODUCTION
District Heating & Cooling networks are a major cost
factor for Energy Providers and subject to permanent
search for cost improvements.
Flexible plastic pipe systems have been a major step
for cost reduction in low temperature networks. With
the new EN 15632 [1] the necessary basis for
certification of these systems has been laid. This is a
milestone in terms of acknowledgement for flexible
systems as a proven part of future network
developments.
As flow temperatures and pressures are reduced, the
field of application for flexible plastic systems is
increasing. Until recently only a small percentage of
District Heating companies have started to use plastic
pipes in their networks. These were kind of pioneers
who co-created systems together with the industry.
The first co-development of such a system was started
already in 1980 by the Austrian Electricity company
STEWEAG. They were looking for a pre-insulated
piping system as easy to install as an electric cable.

Photo no.:1 first Flexalen installation 1981

Together with the University of Leoben, the long term
durability of two types of plastic medium pipes, made of
PB and PE-X were investigated. The research made by
Dipl.Ing. E.Kramer and Univ.Prof.Dr.J. Koppelmann [2]
was based on OIT (oxygen induction time), tear
strength, elongation at break and internal pressure
tests to determine the lifetime of plastic pipes at 80, 95
and 110 C. The final results were in favour for pipes
made of PB. The calculation of the lifetime for PB pipes
was based on a typical temperature profile used in
secondary district heating networks of STEWEAG. The
lifetime expectancy was stated with 36 years.
The decision of STEWEAG was made for PB pipes due
to their more homogenous structure, superior flexibility
and allowance for welded joints. More than 250 km of
this system have been installed since 1981 in
secondary networks operated by STEWEAG. See also
Univ. Prof. Dr. E. Hnninger, STEWEAG [6].
WHAT CAN BE SOLVED WITH FLEXIBLE PLASTIC
SYSTEMS
Apart from the high and long term investment costs, the
following main problems had to be solved as well:
- Corrosion problems in conventional Systems made
of Steel/PUR/PE or Cu/PUR/PE or Cu/Mineral
wool/PE
- Heat loss due to wet and aged insulation
- System shut downs for maintenance and repair

Photo no.: 2 corroded steel pipe connection
THE NEW FLEXIBLE PIPE GENERATION
In 2001 the Dutch Energy Provider NUON started a
development co-operation with Thermaflex to create
even more flexible and moisture resistance piping
systems. Target was again to reduce the connection
costs for new district heating projects. The new system
September 5
th
to September 7
th

262

developed is called FLEXALEN 600, the improved
version of the system used by STEWEAG. The system
consists of a PB (Polybutene) medium pipe and PO
(Polyolefin) insulation foam welded to a HDPE (high
density Polyethylene) outer casing.
With a new inline production process it was possible to
weld the moisture resistant insulation to the outer
casing. The targets of a corrosion proof and moisture
resistant insulation were met.

Photo no.:3 Flexalen 600 longitudinal cut

FLEXALEN has been the first system to pass a
certification and 3
rd
party control by KIWA, which is
similar to the new EN 15632.
DECREASING INSTALLATION COSTS
The most obvious advantage is the chance to reduce
the installation time with flexible systems supplied in
coil lengths of 100m and more. Compared to rigid
systems the following relative costs have been realized
in actual projects
Table 1

Pre-insulated
steel pipes
FLEXALEN
Material costs 100% 90150%
Installation time 100% 2025%
Trenching 100% 5070%
Total 100% 6085%

Material costs are depending on the dimensioning of
the system in the first place. In case of optimization of
pipe sizes and lengths according to the advantages of
PB pipes and connection systems, as described later
on, the material costs can be reduced for Flexalen.
Installation costs are proven in practical experiences
since almost 30 years. Flexalen systems can be
installed 5 times faster than rigid systems.
Lower costs in trenching is related to the fact that
Flexalen systems are supplied in double line systems
up to d63 there trenches can be compared to cable
ducts. Another fact is that connections are most of the
time only necessary at branches, for sticks trenches
need to be suitable for execution of the welding and the
insulation process.

Photo no.: 4 Steel compared to Flexalen

Although material cost for plastic pipes are higher
especially for larger dimensions, the total installed
system costs are lower, especially when using double
pipe systems wherever possible (see Photo nr. 5).
BRANCH SOLUTIONS
Until now 2 types of branch solutions have been used.
On site welded solution with Half-shells plus insulation
to cover (see Photo no.:4). This technology has been
used for smaller networks. Due to homogenous
welding techniques, either with polyfusion or with
electrofuction fittings, the branches are corrosion free
and offer the same inner diameter as the pipes.

Photo no.: 5 Polyfusion welded Flexalen branch

Pre-insulated Tees are another way to secure more
reliable network quality due the reduced number of
joints to be made on site (see Photo no.: 6). The
straight connections are insulated with special kits with
a robust slide over HDPE tube, which is sealed to the
outer casing with heat shrinks.
September 5
th
to September 7
th

263

Photo no.: 6 Pre-insulated branch

The pressure for even more economic solutions for
connections has led to further innovations in close co-
operation between the Dutch Energy Provider Eneco
and Thermaflex.
PRE-FABRICATED NETWORKS
Compared to the branch solutions described before,
these new solutions take full advantage of flexible
welded systems, in order to further reduce the number
of joints on sites.
A new type of pre-fabricated network has been
developed. High flexibility and a minimum of
connections was the goal.
The first application was the district heating network
Capelle a/d Ijssel in the surroundings of Rotterdam for
renovation in difficult circumstances under houses
(high ground water level) to replace corroded heating
and sanitary distribution systems.
The space under the houses is so small that neither
welding nor mechanical connections can be carried out
in a safe way. Steel welding is even forbidden under
these conditions.

Photo no.: 7 Narrow space for district heating lines under
buildings in NL

The result was the Flexalink (Flexalen T-Link) solution
a very small, flexible and 100% watertight system, pre-
fabricated and pressure tested by Thermaflex NL.

Photo no.: 8 Flexalen T-Link

This solution combines the following advantages:
- Factory made welding and branch insulation all
watertight and pressure tested.
- Connections are made under clean manufacturing
circumstances. No weather influences, no failure
costs.
- Customer made connections according to the real
situation.
- Light weight and flexible for easy sliding into the
trench or under the house
- Fast Installing time (first project experience 10
houses/day)
- Reducing connection costs in new building
projects.
- Less system parts on the building area.
This development covered all the wishes and
requirements of the Dutch Energy Provider ENECO.
They have already ordered this system for 800 house
connections in 2010.
September 5
th
to September 7
th

264

Photo no.: 9 Installation of Flexalen T-Link

Photo no.: 10 Installation of Flexalen T-Link

These 2 photos (no 8 and 9) demonstrate how easy it
was to slide the connection into the duct under the
house.
Also further connections between the pre-fabricated
sections were made before sliding the entire system
under the house. Only the last connection had to be
made in the duct. See also photos no.:11 & 12.

Photo no.: 11 pipes under the building

Photo no.:12 Welding of T-Link to the next section

All experiences so far have been very positive and
have led to further applications already.
RENOVATION IN PURMERENT
The situation of the current district heating network in
Purmerend (Energy supplier Stadsverwarming
Purmerend) is very critical. Due to higher ground water
levels than expected the current metal pipe systems
have corroded and need to be replaced.
As the network has been installed under the basement
of the attached housing schemes, the space for the
installer is very tight and it is not allowed to use any
steel welding process in these circumstances. Flexalen
T-Link has been identified to be the ideal solution.
Purmerend has ordered this system for 300 house
connections for 2010 already.
For the renovation market in the Netherlands, this
solution has shown big advantages. This solution is
now available for Energy Provider worldwide not just
for renovation, but also for new networks.
PRE-FABRICATED NETWORKS FOR NEW
PROJECTS
The conditions in new building situations are much
easier and this solution can help to reduce the
connection costs.
Both Dutch Energy Provider, Eneco and Nuon are
investigating this new solution for new building projects.
September 5
th
to September 7
th

265

Fig. 1 network scheme suitable for T-Link

Especially for networks with short distances between
the branches and the connection to the houses, a high
degree of pre-fabrication can be offered.
One solution is the T-Link as described before. Another
one can be a main line up to 100 m with factory welded
and insulated Tees. This reduces the work on site to
just 1 welding for the house connection line. This
solution is interesting for longer distance house
connections.
THE IMPORTANCE OF THE NETWORK DESIGN
Flexible PB piping systems offer important advantages
compared to other plastic and steel systems in terms of
layout and design.
Compared to steel pipes flexible PB systems can be
laid more direct as the system is flexible and fully self-
compensating. Expansion loops and elbows can be
saved. The saving in pipe length can be calculated with
710%.
PB systems offer low friction loss and show no
calcification or incrustation during the lifetime. The
polyfusion welded fittings have at least the same inner
diameter as the pipe and offer the same high abrasion
resistance. Taking this into account, some extra
security factors used in pipe dimensioning can be
eliminated.
PB systems can be operated with much higher flow
speed; hence smaller dimensions can be used for the
same load requirement. See also J. Korsman, I.M.
Smits, E. van der Ven [4].
With relation to the topic of this paper, the following
additional savings can be made during the network
design:
- Looking for a new building area, mostly streets with
block of houses there are two possibilities:
Installing under the floor or Installing in the streets.
- For every house connection under the floor only
two welds and two insulation sets are necessary.
Reduction of the installing time/costs by 5060%.
- For every house connection in the street, a
pre-fabrication e.g. for 810 house-connections
built in into one 100m coil can safe installing
time/costs totally including excavating the trenches
of 70%
Taking all these possibilities for savings and
optimization into account, the next most important topic
for Energy Provider, the efficiency of the network in
operation, can be tackled as well. Due to the possible
reduction in network length and pipe diameter, the
overall heat loss can be reduced as well. See also
results from the work of I.M. Smits, J. Korsman, J.T.
van Wijnkoop and E.J.H.M. van der Ven [5] and J.T.
van Wijnkoop, E.J.H.M. van der Ven [3].
September 5
th
to September 7
th

266

CONCLUSION
New connection solutions, which meet the
requirements of Energy Companies in terms of lower
investment costs, faster installation time and durability,
have been co-created with leading Energy Suppliers in
Austria and The Netherlands.
These solutions are based on flexible and weld-able
plastic systems and have been used successfully up to
29 years in secondary networks with maximum
operation temperatures of 95 C (peak temperature)
and maximum pressure of 8 bars.
The latest development is going into the direction of a
higher degree of pre-fabrication, by including the entire
connection line to the houses as well as parts of the
main line into one piece, made up and fully pressure
tested in the factory.
The experiences in recent projects are showing
installation times 510 times faster compared to
conventional pre-insulated steel. The number of
connections to be made on site is significantly reduced.
Successful projects with Energy Suppliers in The
Netherlands are confirming the advantages of this new
connection solution.
ACKNOWLEDGEMENT
Acknowledgement go to the innovative engineers in
Energy Providers like STEWEAG, NUON and ENECO,
who are drivers for co-creation of new solutions for the
benefit of the entire industry.
REFERENCES
[1] EN 15632 District heating pipes, Pre-insulated
flexible pipe systems, Requirements and test
methods
[2] Dipl. Ing. E. Kramer, Univ.Prof. Dr. J. Koppelmann,
Untersuchung zur Dimensionierung einer flexiblen
Fernwrmeleitung aus Kunststoff, University
Leoben, Austria; 1984.
[3] J.T. van Wijnkoop, E. van der Ven, Verification of
heat loss measurements, 12
th
ISDHC 2010.
[4] J. Korsman, I.M. Smits, E.J.H.M. van der Ven
Heat loss analysis and optimization of a flexible
piping system, 12
th
ISDHC 2010.
[5] I.M. Smits, J. Korsman, J.T. van Wijnkoop and
E.J.H.M. van der Ven, Comparison of competitive
(semi)flexible piping systems by means of heat
loss measurement, 12th ISDHC 2010.
[6] Univ. Prof. Dr. E. Hnninger, Sekundrnetze
frdern die Fernwrmeanwendung, STEWEAG,
Fernwrme International 14/85.
[7] C. Engel, Polybutene The alternative material for
heating and domestic hot & cold water systems,
PLASTIC PIPES IX, Edinburgh 1995.
September 5
th
to September 7
th

267

COMPETITIVENESS OF COMBINED HEAT AND POWER PLANT TECHNOLOGIES
IN ESTONIAN CONDITIONS
E. Latov
1
and A. Siirde
1

1
Department of Thermal Engineering
Tallinn University of Technology, Tallinn, ESTONIA

ABSTRACT
The goal of this paper is to evaluate competitiveness of
market ready combined heat and power (hereafter
called as CHP) technologies for CHP expansion
potential locations in Estonian. The main criteria to
indicate preference of CHP technology is a heat price
by which the internal rate of return is equal to investors
expectations.
Calculation results shows, that in spite of the
advantages of gas engines (relatively low investment
costs and high electrical efficiency) the calculated heat
prices are the highest. Heat price for expected 7% IRR
is 5361 EURO/MWhheat depending on heat demand.
It is mainly because of relatively high natural gas price.
Under 5 MWel ORC is competitive to steam
turbine/engine technology. Heat prices are lower for
14 EURO/MWhfuel, depending on heat demands.
Heat prices for places with annual heat demand under
20 000 MWh are mainly above 45 EURO/MWhfuel
(average heat prices for biomass boiler houses in
Estonia is between 4045 EURO/MWh). Developing of
CHP plants in such areas is feasible in the case of
grant payments for investments. CHP plant
development based on wood chips or peat could be
feasible without grant payments in the places where
heat demand exceed 30 00040 000 MWh annual.
Carefully selected CHP technology and capacity can
afford higher IRR when keeping competitive heat
prices.
The most feasible places for CHP expansion in Estonia
are Maardu, Viljandi, Rakvere, Valga, Haapsalu, Vru,
Paide and Plva.
INTRODUCTION
This paper draws on ongoing project Analysis on the
technical and economic consequences of renewable
energy based CHP systems in new areas with the
lowered useful heat demand or after implementation of
energy conservation measures in the areas with older
buildings within the project Primary Energy Efficiency
partly financed by NER, which contributes to the effort
of enhancing the primary energy efficiency (PEE) and
reducing CO2 emissions in the energy sector.
Present-day world energy policy is based on two main
directions: energy efficiency and environmental
protection. Efficient CHP production is one of the
energy consumption effective methods, where CHP
production from the renewable fuels is preferable. [1]
During the last 2 years a few CHP plants working on
woodchips and peat were build in Estonia. A few of
biomass CHP plants are under active development. All
of them are planed or constructed in major Estonian
cities and are based on backpressure steam turbine
technology. At the same time feed-in tariffs as well as
possibilities to get grants for expanding of CHP and
usage of renewable fuels makes CHP expansion more
attractive for locations with a lower heat demands.
Steam turbine technology is a classic for CHP plants.
But in relatively small-scale boilers and district heating
systems use of steam turbines is connected to
economically less efficient operation (commonly higher
specific investment costs, O&M costs and lower
electrical efficiency) where use of other alternative CHP
technologies could be preferable.
The goal of this paper is to evaluate competitiveness of
market ready CHP technologies for CHP expansion
potential locations in Estonian. The main criteria to
indicate preference of CHP technology is a heat price
by which the internal rate of return (hereafter called as
IRR) is equal to investors expectations.
The paper is structured as follows. After an overview of
places where construction of CHP plants can be
reasonable the paper provides principles for evaluation
of CHP technologies competitiveness. Next sections
provide an overview of the CHP technologies which
can be used in CHP plants and descriptions of main
fuel sources for energy production in Estonia. The last
section provides heat price calculation examples based
on proposed principles for evaluation of CHP
technologies competitiveness.
LOCATIONS OF POTENTIAL BIOFUELED CHP
PLANTS IN ESTONIA
Fig. 1 shows major Estonian cities and municipalities
which are distributed by the annual heat demands.
Places where CHP plants are already constructed or
under construction, as well as in a state of active
development are marked separately.
Fig. 1 reflects well known principles, where the
consumers with higher annual heat consumptions are
more preferable.
September 5
th
to September 7
th

268

Fig. 1 Distribution of Estonian cities and municipalities by the annual heat demands

PRINCIPLES FOR EVALUATION OF CHP
TECHNOLOGIES COMPETITIVENESS
The revenues of a CHP company are generated from
the heat and electricity sales. Theoretically they must
cover the operation and maintenance (hereafter called
as O&M) costs of the CHP plant completely and
provide an expected IRR. Main CHP plant related costs
and incomes are shown in Fig. 2.

Fig. 2 CHP plant incomes and costs

Fuel costs, pollution charges and ash handling costs are
mainly depend on used fuel properties and are
estimated in section Fuel sources for energy production.
CHP technology related fixed operation and
maintenance costs depend on selected CHP technology
and are defined in % from the investment costs annual.
They are estimated in section CHP technologies.
Electricity sells depends on amount of produced
electricity (computable value) and fuel prices. Fuel prices
are estimated by taken into account feed-in tariffs
described in Electricity Market Act [2].
Income from the heat sells depends on amount of sold
heat (computable value) and heat price.
Knowing investment costs (specified in section CHP
technologies), and other above mentioned costs and
incomes the power plant operation annual net cash flows
can be calculated and IRR defined.
The principle for evaluation of CHP technologies
competitiveness is based on finding such heat price
which will cause an expected (proposed) IRR, where
calculation/estimation rules for the other cash flows
components are clearly defined.
CHP TECHNOLOGIES
There are numerous CHP technologies that can be
theoretically used for small scale CHP systems, but not
all of them are economically and technically feasible.
The list of main CHP technologies ordered by market
readiness and common heat outputs are shown in Fig. 3.
It is important to consider the market ready solutions first
of all, such as a steam turbine (hereafter called as ST),
steam engine (hereafter called as SE), ORC
technology (hereafter called as ORC) and gas engine
(hereafter called as GE). Hereafter SE and ST are
considered jointly, where capacities less than 1 MWel
correspond to SE by default.
For CHP plant economical calculations it is important to
know such CHP plant parameters as efficiencies, price
and O&M costs.
Above mentioned parameters are obtained and
systemized on the basis of information regarding CHP
plants collected from different information sources such
as [3, 4, 5, 6, 7, 8].
September 5
th
to September 7
th

269

Fig. 3 Main prime mover CHP technologies

Values for investments as well as nominal electrical
capacities used to calculate fuel prices depend on CHP
nominal electrical capacities and CHP technology, as
shown in Table 1.
Table 1. Values for investments and nominal electrical
capacities for selected CHP technologies
Capacity
Specific
investment
costs
Electrical
nominal
efficiency
MW
el
MEURO/MW
el
%
0,1 ST/SE 10,3 10
0,1 GE 1,6 32
1 ST/SE 5,1 15
1 ORC 5,8 15
1 GE 1,0 40
5 ST/SE 3,2 22
5 ORC 4,5 16
5 GE 0,8 41
10 ST/SE 2,9 22
10 ORC 4,2 16
10 GE 0,8 42
Technology

In this paper investment means all costs before CHP
plant commissioning.
For the evaluation of CHP competitiveness the efficiency
drop working at partial load is taken into account. It is
assumed, that minimal CHP heat load for all
technologies is 25% from the nominal heat load. It is
assumed, that electrical efficiency working at minimal
heat load is 35% for steam engine/turbine, 80% for gas
engine and 85% for ORC from the nominal electrical
efficiency.
It is assumed, that CHP technology related fixed O&M
costs for SE/ST, ORC and GE are relatively 2.5%, 2%
and 3.5% from the investment costs annual.
FUEL SOURCES FOR ENERGY PRODUCTION
Main fuel sources for under 10 MWel CHP plants in
Estonia are natural gas, peat and wood chips.
Fuel prices
The fuel prices taken as basis for heat price calculations
are as follows:
- Peat price 11.7 EUR/MWhfuel. Proposed price is
based on average peat price levels obtained from
Tootsi Turvas AS, the biggest peat milling and
exporting enterprise in Estonia.
- Wood chips price 12.8 EUR/MWhfuel. Proposed
price is based on latest data, published by Estonian
Institute of Economic research in their web based
price information system [9].
- Natural gas 35 EUR/MWhfuel. Proposed price is
an average price for the latest data published by
Statistics Department of Estoni [10].
September 5
th
to September 7
th

270

Ash handling costs
Ash handling costs calculations are based on
assumptions, that:
- Peat ash content is 5%. Average calorific value is
3.3 MWh/t [11];
- Wood chips ash content is 1%, calorific value
2.4 MWh/t;
- Natural gas combustion does not emit any ash;
- Regarding to information obtained from different
landfill owners, an average for year 2012 expected
ash removal costs (ash transportation to landfill, and
storing) are 45 EUR/t.
- The combustion plant is equipped with dry ash
removing system.
Taking into account above mentioned information the
ash handling costs per MWh of fuel energy content for
the peat and wood chips are ~0.19 and 0.72 EURO
respectively.
Pollution charges
Pollution charges and levels are calculated base on the
Environmental Charges Act [12], Regulation No 99/2004
[13] and No 94/2004 of Estonian Minister of Environment
[14].
The method described in [13] takes into account different
combustion technologies, flue gas cleaning
technologies, control devices as well as capacities to
define emission factors of pollutants.
Table 2. Summarised results of the heat price calculations for different CHP expansion scenarios
Heat
demand
Maximum
heat capacity
Heat Electrical
MWh MW MW
h
MW
e
MEURO Without
grant
With
grant
Without
grant
With
grant
5000 1,5 ST/SE Peat 0,83 0,13 1,31 55 --- 67 ---
5000 1,5 ST/SE Woodchips 0,83 0,13 1,31 53 39 65 44
5000 1,5 Gas engine Natural gas 0,83 0,63 0,77 61 --- 66 ---
10000 3 ST/SE Peat 1,65 0,27 2,54 54 --- 65 ---
10000 3 ST/SE Woodchips 1,65 0,27 2,54 51 37 62 43
10000 3 ORC Peat 1,65 0,34 2,93 53 --- 66 ---
10000 3 ORC Woodchips 1,65 0,34 2,93 49 32 61 38
10000 3 Gas engine Natural gas 1,65 1,36 1,29 56 --- 60 ---
20000 6 ST/SE Peat 3,30 0,60 4,46 49 --- 59 ---
20000 6 ST/SE Woodchips 3,30 0,60 4,46 46 34 56 39
20000 6 ORC Peat 3,30 0,70 4,94 46 --- 57 ---
20000 6 ORC Woodchips 3,30 0,70 4,94 42 29 53 34
40000 12 ST/SE Peat 6,60 1,44 7,08 41 --- 49 ---
40000 12 ST/SE Woodchips 6,60 1,44 7,08 38 29 45 32
40000 12 ORC Peat 6,60 1,44 8,10 40 --- 49 ---
40000 12 ORC Woodchips 6,60 1,44 8,10 36 25 44 29
80000 24 ST/SE Peat 13,20 3,90 14,57 39 --- 47 ---
80000 24 ST/SE Woodchips 13,20 3,90 14,57 35 --- 42 ---
80000 24 ORC Peat 13,20 2,96 15,20 38 --- 46 ---
80000 24 ORC Woodchips 13,20 2,96 15,20 34 --- 42 ---
Heat price, EURO/MWh
IRR 7% IRR 12%
District heating area
Technology Fuel
CHP plant capacity
Investment
To avoid complexity of the analysis to be issued from
different combinations of capacities, combustion
technologies, fuel gas cleaning and control equipment
it is assumed that:
- Thermal capacity of combustion plants is below
50MW;
- Selected combustion technology provides lowest
emission level than the others in [13] mentioned
combustion technologies;
- Combustion plant is equipped with the most
effective control systems mentioned in [13];
- Combustion plant is equipped with the most
effective flue gas treatment technology mentioned
in [13].
Calculated levels for pollution charges for year 2013
are:
- ~0.07 EUR/MWhfuel for wood chips;
- ~0.95 EUR/MWhfuel for peat;
- ~0.43 EUR/MWhfuel for natural gas.
September 5
th
to September 7
th

271

HEAT PRICE CALCULATIONS
Calculations of heat prices are provided in
correspondence with principles described in section
Principles for evaluation of CHP technologies
competitiveness.
Heat prices are evaluated for different scenarios.
Scenarios include described heat demands, considered
fuels and technologies.
Heat prices are calculated for 7% and 12% IRR.
Heat price for CHP plant developing scenarios which
satisfy the requirements described in regulation [X],
which define conditions for grant payments to expand
renewable energy production and construction of CHP
plants in Estonia, are calculated separately.
For calculating heat prices in addition to information
from previous paper sections, some other figures have
to be specified:
- Cash flows are calculated for 20 years;
- CHP starts energy production in the beginning of
2013;
- Expected rate of inflation is 1.5%;
- Heat loses in district heating network are 15%;
- Heat load profile is estimated based on heat load
model described in [15] taking as a basis the heat
load duration curve shape of Tallinn.
The results matrix of heat price calculations is shown in
Table 2.
CONCLUSION
The technologies for smaller CHP applications are
more expensive (specific price) and less efficient than
those for larger CHP plants.
At present peat is considered as a good alternative for
wood chips. Lower fuel price (11.7 EUR/MWh) smooth
over higher than for wood chips ash handling costs and
pollution charges. At the same time wood chips are
more preferable because of higher feed-in tariffs for
produced electricity.
The advantages of gas engine CHP plants are
relatively low investment costs and high electrical
efficiency. But because of high natural gas price
(MWhfuel price is 2.53 times higher than for wood
chips and peat) and relatively high fixed O&M costs the
calculated heat prices are the highest. Heat price for
expected 7% IRR is between 53 and
61 EURO/MWhheat depending on heat demand.
Under 5 MWel ORC is competitive to SE/ST
technology. Calculated heat prices are lower for
14 EURO/MWhfuel, where higher fuel price difference
corresponds to places with lower heat demands.
Heat prices for places with annual heat demand under
20 000 MWh are mainly above 45 EURO/MWhfuel
where an average heat prices for biomass boiler
houses are between 4045 EURO/MWh [16].
Developing of CHP plants in such heat demand areas
is feasible in the case of receiving of grant payments
for investments.
CHP plant development based on wood chips or peat
could be feasible without grant payments in the places
where heat demand exceed 3000-40000 MWh annual.
Carefully selected CHP technology and capacity can
afford higher IRR when keeping competitive heat
prices.
The most feasible places for CHP expansion in Estonia
are Maardu, Viljandi, Rakvere, Valga, Haapsalu, Vru,
Paide and Plva.
Calculation results are valid for assumed cases only.
Other particular cases should be calculated
individually.
REFERENCES
[1] C. Dtsch and A. Jentsch, District heating (DH) in
areas with low heat demand density (HDD):
A chance for the integration of renewable energy
sources (RES), 10th International Symposium on
District Heating and Cooling, 35, September
2006, p. 2
www: http://www.lsta.lt/files/events/20_doetsch.pdf
[20.01.2010]
[2] Electricity Market Act www:
https://www.riigiteataja.ee/ert/act.jsp?id=13279771
[14.05.2010]
[3] Schwaiger, H., Jungmeier, G, (2007) Overview of
CHP plants in Europe and Life Cycle Assessment
(LCA) of GHG emissions for Biomass and Fossil
Fuel CHP Systems CIBE Conference
Cognration biomasse dans l'industrie et sur les
rseaux de chaleur opportunits retours
d'exprience-perspectives
[4] Obernberger, I., Thek, G, Techno-economic
evoluation of selected decentralised CHP
appications based on biomass combustion in IEA
partner countries Graz (2010)
[5] Bryson, T., Major, W., Darrow, Ken. Assessment of
On-Site Power. Opportunities in the Industrial
Sector, Carlsbad (2001) www:
http://www.uschpa.org/files/public/Assessment%20
of%20Onsite%20Power%2001.pdf
[14.05.2009]
[6] Kirjavainen, M., Sipil, K., Savola, T. Small-scale
biomass CHP technologies. Situation in Finland,
Denmark and Sweden, VTT Processes (2004)
September 5
th
to September 7
th

272

www:
http://www.opet-
chp.net/download/wp2/small_scale_biomass_chp_t
echnologies.pdf [14.05.2010]
[7] Institute for Thermal Turbomachinery and Machine
Dynamics, Cogeneration (CHP) Technology
Portrait, Vienna (2002) www:
http://www.energytech.at/pdf/techportrait_kwk_en.pdf
[14.05.2010]
[8] U. S. Environmental Protection Agency Combined
Heat and Power Partnership, Biomass Combined
Heat and Power Catalog of Technologies, (2007)
www:
http://www.epa.gov/chp/documents/biomass_chp_
catalog.pdf [14.05.2010]
[9] Estonian Institute of Economic research www:
http://www.ki.ee [14.05.2010]
[10] Statistics Estonia www: www.stat.ee [14.05.2010]
[11] Paappanen, T., Leinonen,A. Fuel peat industry in
EU, 2005, p. 134 www:
http://turbaliit.ee/index.php?picfile=21 [14.05.2010]
[12] Environmental Charges Act, [14.05.2010] www:
http://www.riigiteataja.ee/ert/act.jsp?id=13316043
[14.05.2010]
[13] Procedure and Methods for Determining Emissions
of Pollutants from Combustion Plants into Ambient
Air www:
http://www.riigiteataja.ee/ert/act.jsp?id=789462
[14.05.2010]
[14] Vlishu eralduva ssinikdioksiidi heitkoguse
mramismeetod www:
http://www.riigiteataja.ee/ert/act.jsp?id=127572 15
[14.05.2010]
[15] Latov, E., Siirde, A. (2010). Heat load model for
small-scale CHP planning. In: Proceedings of
International Conference on Renewable Energies
and Power Quality: International Conference on
Renewable Energies and Power Quality
(ICREPQ10), Granada (Spain), 23-25th March,
2010., 2010.
[16] Estonian Competition Authorities approved district
heat maximum prices (without VAT) to end-users
www:
http://www.konkurentsiamet.ee/file.php?15416
[14.05.2010]

September 5
th
to September 7
th

273

DISTRIBUTION OF HEAT USE IN SWEDEN
Margaretha Borgstrm, Sven Werner
1
School of Business and Engineering
Halmstad University, PO box 823, S-301 18 Halmstad Sweden

ABSTRACT
The current heat use refers normally to the average
heat use in a country or a sector during the course of a
year. But it is also important to be aware of the
distribution of high to low use when estimating the
potential for reducing total heat use.
Energy statistical data published in the annual report
from Statistics Sweden have been supplemented by a
deeper analysis of distribution of heat use and
systematic causes regarding high heat use.
The aim of this paper is to explain the variation in heat
use with respect to construction year, degree days and
energy efficiency measures.
In the Swedish energy efficiency debate, many voices
refer to systematic causes for high heat use. However,
the results from this study do not support this opinion,
since the use distribution mostly comes from individual
causes. The most important implication of the study
results is that systematic policy measures will have a
low impact on the total national energy efficiency.
INTRODUCTION
Multi-family residential buildings and service sector
premises constitute 80% of the customer stock in the
Swedish district heating systems. The level of future
heat use in these buildings will thus have a strong
influence on the future district heating economy and the
corresponding investment demand. It is therefore of
interest to collect information and make analyses of the
costumer heat use and how the heat use will develop in
the future.
Specific heat use in multi-family buildings and service
sector premises has decreased considerably since the
1970s. In 2006, the specific heat use in multi-family
buildings has decreased by 38% compared to the heat
use in 1972. The lower heat use is due to increasing
energy prices and more energy efficient buildings.
An extensive study of the current heat use for buildings
in Sweden has been performed. The input information
for this study was constituted by the anonymous
responses to the annual survey of energy use in multi-
family buildings and service sector premises performed
for 2006 by Statistics Sweden, [1] & [2]. The responses
provided input data from 11253 buildings having a total
area of 77.6 million square metres. By using scaling
factors, estimates could be made for the entire country,
corresponding to 310 million square metres in multi-
family buildings and service sector premises. The
survey sample thus constituted a sizable portion of the
entire building stock.
This energy statistical data, published in the annual
reports from Statistics Sweden, have been
supplemented with a deeper analysis of the distribution
of the heat use and the systematic causes regarding
high heat use. Independent variables for explanation of
variations were number of degree-days, construction
year, ventilation system, energy efficiency measure,
and co-use of heat supply. High and low users were
also analysed by location, construction year, heat
supply method, ownership, and building size [3]. In this
short paper, the specific heat use will be presented by
its distribution, construction year, degree days and
energy efficiency measures.
1. Distribution of heat use
The total distribution of specific heat use as a function
of the percentage of the building area of all multi-family
buildings and service sector premises in Sweden is
shown in Fig. 1.
0
50
100
150
200
250
300
350
400
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Share of all national building space
Heat use
kWh/m
2
Multi-family buildings
Premises

Fig. 1 Heat use distribution during 2006 as a function of
the share of all national building space. The diagram is an
estimation for all multi-family and service buildings in
Sweden.

The area under each curve is the total heat used in
multi-family buildings and service sector premises
during 2006. The figure shows that 13% of the area in
multi-family buildings had a specific heat use of more
than 200 kWh/m
2
, and 12% of the area in service
sector premises had a specific heat use of more than
200 kWh/m
2
. This result shows that there are no major
differences between the percentages of the building
September 5
th
to September 7
th

274

area with high heat use in multi-family buildings and
service sector premises.
The results in Fig. 1 also show that 11% of the building
areas in multi-family buildings, and 31% of the building
area in service sector premises have a specific heat
use lower than 100 kWh/m2.
Buildings with heat use 200 kWh/m2 or more have
been further analysed and the results are presented in
the following section considering construction year.
2. Construction year
Fig. 2 shows specific heat use in multi-family buildings
as a function of construction year. The figure also
includes the average value each year, together giving
the total average specific heat use of 152 kWh/m2.
There are no major differences in heat use in buildings
constructed before 1980. After 1980, the heat use was
approximated 15% lower than the average heat use for
all buildings in Sweden.
0
50
100
150
200
250
300
350
400
1930 1940 1950 1960 1970 1980 1990 2000 2010
Construction year
Heat use
kWh/m
2

Fig. 2 Specific heat use as a function of construction year
for 4285 multi-family buildings

0
50
100
150
200
250
300
350
400
1930 1940 1950 1960 1970 1980 1990 2000 2010
Construction year
Heat use
kWh/m
2

Fig. 3 Specific heat use as a function of construction year
for 4061 service sector buildings.
The heat use in service sector premises is shown in
Fig. 3. Also in these buildings, the average heat use
after 1980 is lower (about 10%) than the average heat
use in all service sector premises in Sweden.
The relationship between construction year and high
heat use in buildings has been analysed. The definition
of high heat use is 200 kWh/m2 or more. Fig. 4 shows
the results for multi-family buildings. There were a total
of 179.3 millions square metres in multi-family buildings
in 2006 and 13% of the heated area had heat use of at
least 200 kWh/m2.
Of special interest are buildings built during the period
196574, when a large part of the existing buildings in
Sweden were built. During this period there were no
requirements for low energy use in buildings.
In multi-family buildings built during the period
196574, 30% of the total area had heat use of at least
200 kWh/m2 and for buildings built in the period
194160. 42% of the total building area had heat use
of 200 kWh/m2 or more.

Fig. 4 Total square metres where heat use is higher or
equal to 200 kWh/m2 in multi-family buildings categorised
by construction year.

Fig. 5 Total square metres where heat use is higher or
equal to 200 kWh/m2 in service sector premises
categorised by construction year.

The relationship between high heat use in service
sector premises and construction year is shown in
Fig. 5. During 2006, 15.6 million square metres had
heat use of at least 200 kWh/m
2
. Service sector
premises built between 1965 and 1974 had high heat
use in 3.5 million square metres.
September 5
th
to September 7
th

275

The results show that the period 19651974 did not
have a dramatically higher heat use in the construction
year analysis.
3. Degree days
The climate in Sweden varies with a much colder
climate in the northern part compared to the southern
part. Since the statistical data consist of buildings from
different parts of Sweden, the influence of the local
climate on the heat use in buildings can be analysed.
This has been done by analysing the correlation
between the number of degree days for the location of
a building and the corresponding specific heat use.
The number of degree days, according to the Swedish
definition, varies from approximately 3000 in the south
up to 7000 in the north of Sweden. Each building in the
analysis was connected to one of 14 climate areas.
y = 15,63x
0,28
0
50
100
150
200
250
300
350
400
2000 2500 3000 3500 4000 4500 5000
Degree-days
Heat use
kWh/m
2

Fig. 6 Specific heat use for 5111 multi-family buildings as
a function of the number degree days in each climate
area.

y = 10,37x
0,30
0
50
100
150
200
250
300
350
400
2000 2500 3000 3500 4000 4500 5000
Number of degree-days
Heat use
kWh/m
2

Fig. 7 Specific heat use for 6041 service buildings as a
function of degree days in each climate area.

Fig. 6 and Fig. 7 show the specific heat use as a
function of degree days for multi-family buildings and
service sector premises. The figures also show the
average curve and its equation for specific heat use as
a function of degree-days. You should also note that
the exponent in the fitted equations has only the
magnitude of 0.3 instead of the 0.5 exponent obtained
in the theoretical analysis of the optimal wall insulation
as a function of degree-days.
The results show that the average difference between
Northern and Southern Sweden was small, implying a
small climatic impact on heat use. The main conclusion
from this analysis is that the individual variation in each
climate area is much higher than the local impact of
climate. This astonishing conclusion can have several
different explanations:
- Higher awareness and consequences of low
building heat resistances in Northern Sweden
- Lower regional GDP in Northern Sweden giving
higher incentive to reduce heat costs
- More frequent snow cover in Northern Sweden
giving extra heat resistance during the winter.
4. Energy efficiency measures.
The statistical data shows the energy efficiency
measures during the period 19952005. The energy-
efficiency measures were:
a. Supplementary insulation
b. More energy efficient windows
c. Balancing heating- and ventilation systems
d. Electrical efficiency measures
e. Heat recovery in ventilation systems
In multi-family buildings, one or several energy
efficiency measures were implemented for an
estimated floor area of 57.6 million square metres
during the period 19952005. No energy efficiency
measures had been performed for an estimated floor
area of 92.2 million square metres during the same
period.
In service sector premises, with an estimated floor area
of 37.2 million square metres, one or several measures
had been taken during the period 1995-2005. During
the same period, no measures had been taken for an
estimated floor area of 70.3 million square metres.
The most common measures in multi-family buildings
and service sector premises were balancing of heating-
and ventilation systems.
In many buildings, a combination of two or several
energy efficiency measures had been taken in the
same building. In some buildings, up to five measures
have been taken in the same building.
The average heat use in multi-family buildings and
service sector premises in relation to measures taken
is shown by bars in Fig. 8 and Fig. 9. The horizontal
lines show the average heat use in buildings, in which
no energy efficiency measure was performed.
September 5
th
to September 7
th

276

Fig. 8 Average heat use in multi-family buildings in
relation to the measures performed. The measure figures
correspond to the measures defined in the text.

Fig. 9 Average heat use in service sector premises in
relation to the measures performed. The measure figures
correspond to the measures defined in the text.

As shown in the figures 8 and 9, there were no
substantial differences in heat use between buildings
where energy-saving measures had been taken and
those where they had not. The conclusion from this
analysis is that the measures taken during these 10
years were taken by late-comers rather than by early
adopters, since heat use after measures were taken
generally corresponds to the average level for all
buildings.
CONCLUSION
The main conclusions from the analysis were:
- Individual variations dominate compared to
systematic causes regarding the specific heat
use in multi-family and service sector buildings.
- The district heating companies can help their
customers by identifying them as high, medium
or low users of heat.
- On the short term, a significant potential exists
for lower heat use in the Swedish multi-family
and service sector buildings.
- More efficient heat use in buildings will probably
represent the most important competitor to
district heating supply in the future.
- In the Swedish energy efficiency debate, many
voices refer to systematic causes for high heat
use. However, the results from this study do not
support this opinion, since the distribution of
heat use mostly comes from individual causes.
The most important implication of the study
results is then that systematic policy measures
will have a low impact on total national energy
efficiency.
REFERENCES
[1] Statistics Sweden, Energistatistik fr
flerbostadshus 2006 (Energy statistics for multi-
family houses during 2006). Statistiska
Meddelanden EN16SM0702.
[2] Statistics Sweden, Energistatistik fr lokaler 2006
(Energy statistics for premises during 2006).
Statistiska Meddelanden EN16SM0703.
[3] Andreasson M, Borgstrm M, Werner
S, Vrmeanvndning i flerbostadshus och lokaler
(Heat use in multi-family buildings and premises
2006) Fjrrsyn report 2009:4, Stockholm 2009.
Available at www.svenskfjarrvarme.se
September 5
th
to September 7
th

277

DAMAGES OF THE TALLINN DISTRICT HEATING NETWORKS AND INDICATIVE
PARAMETERS FOR AN ESTIMATION OF THE NETWORKS GENERAL CONDITION
Aleksandr Hlebnikov
1
, Anna Volkova
1
, Olga Duba
2
, Arvi Poobus
1
, lo Kask
1
1
Department of Thermal Engineering, Faculty of Mechanical Engineering,
Tallinn University of Technology, Kopli 116, 11712 Tallinn, Estonia
2
Tallinna Kte, Punane 36, 13619 Tallinn, Estonia
ahleb@staff.ttu.ee, anna.volkova@ttu.ee

ABSTRACT
District heating networks in Estonia are mostly old and
in bad condition. The state of the district heating
networks of Tallinn is typical for the rest of Estonian DH
networks. The paper includes analysis of the Tallinn
district heating networks. Valid data about damages in
district heating systems received for the last 12 years
were used for an analysis of the networks damages.
Different types of network damages are analysed:
external corrosion, internal corrosion, defect of
installation, factory defects, defect of construction and
other reasons. The number of damages for the different
elements of networks is compared in the paper:
armature, compensator, construction and pipes. Main
factors, which influence damages in district heating
networks, are the age of networks, the quality of
construction works and the network operation
conditions.
The damage quantity dependence on the age of
networks is also defined and analysed in the paper.
The number of damages can be diminished by
reducing the average age of networks. This is possible
by replacing old pipelines and other network system
elements. Pipes average age changes for 20 years
period are simulated according different intensities of
renovation works.
INTRODUCTION
District heating (DH) allows centralized heat production
for an area and hot water transportation to the buildings
through a network of pipes. District heating systems
offer the potential to use energy-efficient and
renewable heat generation technologies, such as
cogeneration technologies which implement both fossil
fuels, as long as biomass and waste [1]. District
heating system is traditional in Estonia. It has formed
approximately 70 per cent of all heating in the country.
The share of heat produced by combined heat and
power production stations is approximately one third. At
the same time, the technical situation of the district
heating networks (and production equipment) is poor.
[2] Unsatisfactory condition of DH networks and
unreliable heat supply can doubt on future of district
heating and the consumers can make a choice towards
a different heat supply alternative. Often the
decentralized heating is not an effective solution for
regional heat supply strategy and it decreases potential
of combined heat and power production.[3].
Nowadays DH systems operate both in big cities and in
small towns, which means, that there is enough heat
load for the installation of new cogeneration equipment.
But before new energy sources installation it is
important to define and analyse the situation with DH
networks.
The purpose of this paper is to define the valid
condition of typical old networks in Estonia, to define
the reasons of damage occurrence on the basis of
operational data and to make forecasts for operation of
a DH network for the next 20 years. The paper includes
analysis of Tallinn district heating networks. The valid
data about damages in district heating systems
collected during past 12 years was used for analysis of
networks damages.
THE PRESENT CONDITION OF TALLINN DISTRICT
HEATING SYSTEM
networks. In Tallinn the heat is transmitted to the
consumers through a 406-kilometres long heating
network including the 93 km of pre-insulated pipes
(23%). District heating systems of Tallinn were
constructed mostly during the 1960-1980 period and
their average age is 22 years.
The AS Tallinna Kte enterprise makes operation of
the bigger part of district heating networks and boiler-
houses of Tallinn.
District heating systems of Tallinn consist of five
districts of the central heat supply: Kesklinna district
(total length ~92 km, length on the balance of
AS Tallinna Kte ~76 km), Lne district (total length
~162 km, length on the balance of AS Tallinna Kte
~141 km), Lne district local networks (total length
~12 km, length on the balance of AS Tallinna Kte
~11 km), Lasname district (total length ~114 km,
length on the balance of AS Tallinna Kte ~106 km),
September 5
th
to September 7
th

278

Maardu district (total length ~25 km, length on the
balance of AS Tallinna Kte ~14 km). [4]
District heating systems of the areas Kesklinna and
Lasname are connected through the pump station
Laagna. The total length of heating networks is 406 km
from which on the balance of AS Tallinna Kte there
are 348 km, or 85,7%.
The following CHP stations and boiler-houses supply
heat to the districts of Tallinn: the CHP Iru (natural gas,
190 MWel, 748 MWth), the boiler house lemiste
(natural gas, 232 MWth); the CHP Vo (wood chips, 25
MWel, 65 MWth); the boiler house Mustame (natural
gas, 390 Wth); the boiler house Kadaka (natural gas,
290 MWth).
Besides the abovementioned there are some small-
scale boiler houses. In Fig. 1 is displayed the basic
scheme of Tallinn heat supply.
District heating systems of Tallinn were constructed
mostly during the 19601980 period and their average
age is 22 years.
Fig. 1 The basic scheme of Tallinn district heating system
The state of DH networks varies for the different
districts of Tallinn.
In Lasname the construction of district heating
systems began in 1970, and the network length is
~106 km at present time. Assuming the actual load the
heating systems of Lasname district are the most
overloaded in town.
The length of main pipelines DN10001200 is ~19 km,
the length of pipes DN400-800 is ~4,4 km. The share of
the main networks is quite big and it is ~22% of total
network length in Tallinn. Thermal isolation is made of
glass wool according to old soviet building norms and it
is the reason of big heat losses in the network. The
heat losses in Lasname network in 2008 were 21%
from the total produced heat.
The interconnected district heating systems of boiler-
houses Mustame, Kadaka and Karjamaa (not in
operation at present time) are related to the Lne area
(districts Mustame and isme). Initially there had
been two separate networks which were merged later
on as a result of growth. In the area Lne the
construction of district heating systems began in 1960.
The length of the Lne area network is ~141 km. The
diameters of the main pipelines are less than those in
the Lasname area.
The length of the main pipelines with diameter
DN400900 is ~27,8 km. The heat losses of the
network in 2008 were 16% from the total produced
heat.
The speciality about the heating system of the area
Lne is that in past there was an open system of hot
water supply. The water added to the system had no
time to purify sufficiently and oxygen and water
hardness led to an intensive internal corrosion of pipes.
In Kesklinn area the network construction began in
1959. Initially the heat supply was carried out by the

Mustame
boiler-house
390 MW
390 MW

Kadaka
boiler-house
290 MW
lemiste
boiler-house
232 MW
(in reserve)
Iru CHP
748 MW (190 MW)
Mustame network
325 MW
Kesklinna
network
180 MW

Maardu
network
Laagna
pump station
200 MW

Vo CHP
65 MW

Lasname
network
268 MW
September 5
th
to September 7
th

279

Tallinna Soojuselektrijaam heat and power station and
later on by the boiler-house lemiste. The district
heating system of Kesklinn area is the oldest in Tallinn.
The average age of the Kesklinn area network is
25 years, the total length is ~76 km.
The length of the main pipelines with diameter DN400-
900 is ~13,8 km. The share of main pipelines in
Kesklinn area network is ~18,1%. Relative heat losses
of Kesklinn network are within the limits of 15...18%. In
comparison with other areas the relative heat losses
are less. The reasons for this are: the bigger network
loading, the not oversized pipes and the significant
share of preinsulated pipes.[5]
THE ASSESSMENT OF DAMAGES
The analysis of networks damage statistics for Tallinn
is made on the basis of valid data collected during the
past 20 years.
The distribution of damages of Tallinn district heating
network is shown in Fig. 2 according the periods of
construction. It is obvious that the most critical situation
is with the sites constructed during the 19801985
period. It can be explained by the poor quality of both
construction works and materials used in construction.
During that period the networks were being constructed
in a hurry and with lack of proper supervision.
0
50
100
150
200
250
300
350
400
-
1965
1965 -
1970
1970 -
1975
1975 -
1980
1980 -
1985
1985 -
1990
1990 -
1995
1995 -
2000
2000 -
2005
2005 -
2008
years of construction
damages during
1998-2009

Fig. 2 Damages of Tallinn district heating networks
according the periods of construction

In Fig. 3 the places of damage in the network elements
are shown: armature, compensators of thermal
lengthening, construction and pipes. The major part of
all damages was the pipes.
During the 19972003 period there were many
problems with armature and compensators; after 2003
the quantity of damages to these elements had
considerably decreased. The oldest thermal
lengthening compensators work since 1959. The
resource of axial compensators is no more than
30 years and by today they are already worn out. The
probability of failures sharply increases. By today the
84% of all compensators should be replaced. Some
parts of the old locking armature also have to be
replaced. The service life of armature has exceeded
25 years. Armature and compensators are partly
renovated; however some pieces of it are old and also
require replacement. [5]
0
50
100
150
200
250
300
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
armature
compensator
construction
pipes

Fig. 3 Places of damage in Tallinn district heating
networks elements
In Fig. 4 the nature of damages is summarized. There
are no data about the character of damage for all the
areas of Tallinn within past 10 years, thats why the
damage allocation by character of damage is shown for
a five year period.
In Tallinn network the significant part of damages is
caused by external corrosion of pipes. Main reasons of
external corrosion are the bad waterproofing of
underground channels and chambers and the
collapsed drainage. Amongst other reasons are the
defects of pipe supports and the destruction of
concrete channels.
0
10
20
30
40
50
60
70
80
90
2004 2005 2006 2007 2008
external corrosion
internal corrosion
deffect of construction
deffect of installation
wrong service
other reasons

Fig. 4 Nature of damages in Tallinn district heating
networks

The second main cause of damages is the internal
corrosion. In 2004 many pipes damaged by internal
September 5
th
to September 7
th

280

corrosion were revealed. Internal corrosion is the most
serious problem in Lne network where an open
system of hot water supply earlier has been used.
Besides the damages caused by defects of installation,
defects of construction, factory defects and improper
maintenance, the other reasons have also been
registered.
The main factors, which have an affect on the damages
in district heating networks, are the age of networks,
the quality of construction works and the network
operation conditions. The two latter can be regulated
by control authorities and proper legislation, however,
the influence of these factors has been reduced in
comparison with the 19701990 period. Then quality of
construction works was very low, drainage systems
were installed incorrectly or were not installed at all and
isolation materials were not qualitative. As regards
district heating operation conditions, the
aforementioned open vented hot water supply system
used in some networks has led to intensive internal
corrosion of pipes.
One important reason for damages reduction is that in
recent years the networks have significantly reduced
pressure. The network works in a stable temperature
mode, the reliability of heat sources is improved and
the quantity of equipment emergency stops forced by
sharp fluctuations of the heat-carrier temperature has
decreased.
Other operation condition factor which influenced
number of district heating system damages was higher
water temperatures in networks (up to 130 t C) than
nowadays (up to 110 t C). Finally we can conclude
that such factors as quality of construction works and
quality of network operation are close to their optimum
at present time in comparison with previous years.
Damage quantity also depends on the age of networks.
The number of damages can be reduced by reducing
the average age of the networks. This is possible by
replacing the old pipelines and other networks systems
elements.
Reconstruction and replacement works are made in
Tallinn, but the intensity of replacement is rather low
and not enough for a stable system operation. It is
important to define, how intensive the network
reconstruction should be.
Data for the three past years were used for defining the
damage dependence (number of damages/km/year) on
the age of networks. Data about damages were
collected for 7 age groups (05 years, 510 years,
1015 years, 1520 years, 2025 years, 2530 years,
3035 years).
Using least squares analysis, a regression equation for
this dependence was defined.
0496 . 1 8985 . 1 0096 , 0
2
+ + = A A D (1),
where
D Number of damages/100 km per year
A Age of networks
Before using this regression for further calculations, we
should check if this equation is appropriate. One of the
main parameters for estimation of regression equation
is the correlation coefficient. It is considered, that the
correlation is good in case when R>0.8. In the case of
damage dependence on pipes age, R is 0.802.
R2=0.643, which means that the equation
characterizes the 64,3% of damage number changes,
but the 35,7% of changes are characterized by another
factors. There is still an influence of other factors,
which can not be changed, such as construction and
installation problems in the past.
Data about damages allocation by the group and
approximation of these data is shown in Fig. 5.
The regression equation can be used for the damage
forecasts in future.
0
10
20
30
40
50
60
70
0 5 10 15 20 25 30 35
Age of networks
damages/100 km
per year

Fig. 5 Damage number dependence on the age of pipes in
district heating systems for the 20052007 period

As it has been mentioned before, the age of networks
depends on the intensity of renovation works.
In Fig. 6 the length of all repaired sites is shown split by
years.
Since 1980 the serial repair of Tallinn district heating
system is being carried out.
Basically the investments have been directed towards
the increase of reliability and the reduction of quantity
and duration of faults in heat supply. It has been
invested a lot in the locking armature.
For the past 10 years ~35 km of district heating
pipelines have been replaced, which is 10% of total
length of the district heating systems in Tallinn area.
September 5
th
to September 7
th

281

The annual replacement of pipes is in average about
3,06 km per year, which is less than 1 percent from the
length of Tallinn DH system pipelines.
0
1
2
3
4
5
6
7
8
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
Length, km

Fig. 6 Length of replaced pipelines by years in Tallinn
district heating network
THE FORECASTS FOR DISTRICT HEATING
SYSTEM AGE
One of the tasks was to assess, how big the renovation
works should be in order to stop increasing the average
age of pipes. A simulation model, which uses both real
data and also some assumptions, was created for such
estimation.
Assuming that the length of pipes (360,67 km) will not
change during the forecast period and that the annual
scope of renovation works will remain the same during
whole of the period means that the length of renovated
pipes also will not change. Besides its was assumed
that every year just the oldest pipes would be
renovated; however in reality the renovation works are
based on the pipes actual state estimation.
Allocation of pipes ages for starting point (2008) is
shown on Fig. 7 [5].

0
5
10
15
20
25
4
9
4
6
4
3
4
0
3
7
3
4
3
1
2
8
2
5
2
2
1
9
1
6
1
3
1
0 7 4 1
age, years
length, km

Fig. 7 Length of DH networks by pipes age (in 2008)

The average age of pipes for each year was calculated
according equation (2)
j
b
a i
i j
b
a i
i
j
a i
i
j
l
l l c j i j l i j l
Aav

= = =

=
) ( ) ( ) ( ) (
(2)
i=b, if
j
b
a i
i
l l <
=
; i=c, if
j
c
a i
i
l l >
=

where
Aav is average age of pipes in j year
li is length of pipes, constructed in i year
I year of construction;
J current year;
a year of construction of the oldest pipes, operating in
the current year.
As a result of simulations, seven forecasts for pipes
average age were calculated according different
intensity of renovation works: for current intensity of
renovation (3,06 km/year) and for intensities when 1%,
1,5%, 2%, 2,5%, 3% and 4% of total DH system length
would be annually renovated. The forecasts were
simulated for the 20 year long period.
The results of simulation are shown in Fig. 8.
5
10
15
20
25
30
35
40
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
2
0
1
7
2
0
1
8
2
0
1
9
2
0
2
0
2
0
2
1
2
0
2
2
2
0
2
3
2
0
2
4
2
0
2
5
2
0
2
6
2
0
2
7
2
0
2
8
2
0
2
9
2
0
3
0
2
0
3
1
2
0
3
2
2
0
3
3
2
0
3
4
2
0
3
5
2
0
3
6
2
0
3
7
2
0
3
8
2
0
3
9
2
0
4
0
age, years
1%
2%
3%
4%
1,50%
2,50%
current

Fig. 8 Pipe age forecasts for different intensity of network
renovation works

As it can be seen from Fig. 8 in case the renovation
stays on the same level, the average age of pipes will
grow till reaching 39 years in 2040. In case the length
of annually changed pipes is 1% or 1,5% higher, the
average age will still rise, but in a less steep way.
When the 2% of DH system length is annually
renovated there will be the minimal changes in age
during first 5 years, after that the age will start rising
and only after 15 years it will begin to decrease.
If renovation intensity is 2,5% of the length or higher,
the average age will not rise at all or will decrease. For
reducing the damages occurrence probability
influenced by the networks age, the amount of repaired
September 5
th
to September 7
th

282

sites should be at least 9 km/year. This way the
process of ageing will slow down and also the average
age will stabilize on a certain mark. One of the possible
solutions is to replace the pipes with higher intensity of
34% until reaching the 1720 years average age and
then reduce the length of renovated pipes per year to
the 22,5% of the whole length of DH network.
CONCLUSIONS
networks. Thats why the result of damage analysis
made for the DH network of Tallinn can be used for the
other networks in Estonia.
The AS Tallinna Kte enterprise makes operation of
85% from the length of district heating networks in
Tallinn. Tallinna Kte data about the damages were
used for assessment.
Places of damages in the DH system are following:
armature, compensator, pipes and construction. Most
of the damages happened in the pipes.
As regards the character of damages, the typical
damages are caused by external corrosion, internal
corrosion, defect of construction, defect of installation
and wrong service. The major part of damages is
caused by external corrosion of pipes.
The age of networks, the quality of construction works
and the network operation conditions are the most
important factors, which influence the damages in
district heating networks. The number of damages can
be reduced by reducing the average age of the
networks. This is possible by replacing the old pipelines
and other networks systems elements. The intensity of
replacement works during last 25 years was less than
one percent from the whole length of pipes.
Seven forecasts for pipes average age according
different intensity of renovation works were simulated:
for current intensity of renovation (3,06 km/year) and
for intensities when 1%, 1,5%, 2%, 2,5%, 3% and 4%
of total DH system length would be annually renovated.
It was concluded, that for maintaining the networks
average age at least at former level, the rate of old
pipelines replacement should exceed the 2,5% of the
whole length of DH system.
AKNOWLEDGMENT
This work has been partly supported by the European
Social Fund within the researcher mobility programme
MOBILITAS (20082015), 01140B/2009
REFERENCES
[1] Cogeneration and district energy sustainable
energy technologies for todayand tomorrow,
International Energy Agency, 2009.
[2] Long-term Public Fuel and Energy Sector
Development Plan until 2015, Riigi. Teataja, RT I,
23.12.2004, 88, 601
[3] Hlebnikov, A.; Siirde, A. The major characteristic
parameters of the estonian district heating
networks, their problems and development. // The
11th International Symposium on District Heating
and Cooling: University of Iceland, 2008, 141148.
[4] Tallinna kte webpage, www.soojus.ee
[5] A. Hlebnikov "The analysis of efficiency and
optimization of district heating networks in
Estonia", Doctoral Thesis, Tallinn University of
Technologies, 2010.
September 5
th
to September 7
th

283

EFFICIENCY OF DISTRICT HEATING WATER PUMPING IN FINLAND
Antti Hakulinen
1
, Jarkko Lampinen
1
and Janne Lavanti
1
1
Pyry Finland Oy

ABSTRACT
The objective of this study was to determine the
savings potential in district heating pumping in Finland.
A measurement method was also developed to quickly
estimate the efficiency of district heating pumping.
The work was based on the data gathered from district
heating statistics. The work is divided into two parts.
The district heating statistics reveal a number of district
heating networks whose consumption of energy
needed for pumping is exceptionally high. These
companies should clarify the reasons for that.
In addition, companies with an exceptionally low
consumption of pumping energy should check their
measurements and data gathering routines.
On average the electricity needed for district heating
pumping should not be over 0.5 per cent of the total
energy supply (=sold+losses). If the density
(supply/length of the network) of the district heating
network is less than 3 GWh/km, the energy needed for
pumping may rise. In any case the proportional
pumping energy should be lower than 1 per cent of
total energy supply.
The Finnish potential for saving in district heating
pumping is estimated to be 20 per cent of the current
pumping energy i.e. 30 GWh/a. This is equivalent to a
yearly saving of approximately 2 million.
PART 1.
INTRODUCTION
In the Finnish district heating systems no typical
pumping arrangements have been used at heat
production plants or at booster pump stations. The
ways of dimensioning and connecting pumps have
varied a lot. This has led to incorrect dimensioning and
connections of pumps, which in turn has caused higher
investment costs and greater pumping energy usage
than expected, operational problems and in the worst
case many interruptions in the use of the network. The
Finnish district heating system is based on the variable
flow operation (consumer driven scheme)
1. COSTS
1.1 The investment costs of a pump (including
motor and controls)
The calculated investment costs of a pump including
motor, control, and pump are shown Fig. 1.
The investment costs of a pump
0
10 000
20 000
30 000
40 000
50 000
60 000
70 000
80 000
90 000
0 100 200 300 400 500
power kW
Total
Pump
Frequency converter
Motor

Fig. 1. The investment costs of a pump.
1.2 Total costs of pumping
Total costs of pumping include capital, maintenance
and energy used in pumping. The pump lifetime costs
are mainly energy costs as we can see from figure 2 on
the next page. The lifetime costs are calculated with
the following assumptions: energy price 60/MW/h,
operating lifetime 15 years, utilization period of
maximum load 5000 h/a, interest rate 5 per cent and
the O&M 1.2 per cent of the investment.

Fig. 2. An example of lifetime costs of three different sized
pumps.
Pump I, power: 16 kW
14 %
2 %
84 %
Capital
O&M
Energy
Pump II, power: 131 kW
7 %
1 %
92 %
Capital
O&M
Energy
Pump III, power: 283 kW
5 %
94 %
1 %
Capital
O&M
Energy
September 5
th
to September 7
th

284

As we can see the pumps efficiency plays a huge role
because the lifetime costs mainly consist of the
operational energy (8494 per cent). For that reason a
lot of attention should be paid to the efficiency when
making the investment. Pumps with a low efficiency
may eat into the savings of the investment many fold.
1.3 Booster pump station and costs
A booster pump station should be considered when the
primary pumps of an energy station do not have
enough capacity to ensure the pressure difference at
the last customer. Typical reasons for the building of a
booster pump station can be: long transmission lines,
expansion of network, optimization of pumping energy
and controlling of pressure level.
The investments of a booster pump station including
pump, motor, frequency converter, building, automation
systems, etc. are shown in the following figures 3
and 4.
The investment costs of a booster pump station
300 000
320 000
340 000
360 000
380 000
400 000
420 000
440 000
460 000
480 000
500 000
0 50 100 150 200 250 300 350 400 450 500
power [kW]

Fig. 3. The investment costs of a booster pump station
(only 1 pump).

The investment cost of a booster pump station (flow + return)
360 000
410 000
460 000
510 000
560 000
610 000
660 000
710 000
760 000
0 50 100 150 200 250 300 350 400 450 500
power [kW]

Fig. 4. The investment costs of a booster pump station
(with 2 pumps).
2. PUMPING ARRANGEMENTS
At the primary station it is usually sensible to divide the
pumping between a few pumps, for example 2*60 per
cent or 2*70 per cent in parallel connection with
individual rotating speed controls. In that way the
pumping of maximum heat load can be managed and
there is a room for possible expansion of the district
heating network. The other pump will act as a summer
pump so that the efficiency of pumping remains high
also when the heat load is low.
By dividing the pumping capacity between many
pumps it is possible to save pumping energy even if
pumping is handled from one point or from the heat
production plant and the booster pump station. The
possibilities to divide the pumping must be examined
case by case by taking into account every single thing
that might have an effect on the costs.
3. MAXIMUM WATER FLOW
The actual cooling of the district heating system in
operational conditions of pumping should be taken into
account when determining the calculated maximum
water flow. It is worthwhile to specify the water flow
according to slightly worse cooling than the actual
conditions require so that there is some design margin
for unusual conditions.
4. OPERATION POINT
To change the rotating speed of a pump with a
frequency converter is a good way regarding energy
efficiency because the pumps efficiency often remains
on high level within the whole adjusting area but the
need for power reduces strongly when the rotating
speed goes down.

Fig. 5. An example functional diagram of a pump.

An example functional diagram is shown in Fig. 5.
When the rotating speed changes, the efficiency
remains good regardless of the changing rotating
speed. The pumping of a district heating network
follows this theoretical situation very well. However
when choosing a district heating pump it is important to
September 5
th
to September 7
th

285

pay attention to its rotating speed which should be at
the minimum from 50 to 60 per cent of the nominal
rotating speed.
PART 2.
INTRODUCTION
The goal of the second part was to motivate the district
heating companies to analyse their pumping methods
and, hopefully, to lower their pumping costs.
Total savings potential in district heating pumping in
Finland was also estimated.
This part is based on the Finnish district heating
statistics of the year 2007 [1]. The statistics cover the
data of nearly 200 district heating companies but only
about 60 of them have reported the electric power used
in district heating pumping.
The biggest companies have reported the pumping
energy, thus, the pumping figure is available to
companies which supply almost 70 per cent of all
district heat in Finland.
STATISTICAL FINDINGS
The used pumping energy in different companies was
analyzed by comparing the pumping energy to the
following parameters:
- Heat supply (sold heat + losses)
- Length of the district heating network
- Heat density (supplied heat energy divided by the
length of the DH network)
The following parameters were also examined but no
clear correlation was to be seen, and the results are
therefore not reported in this paper:
- Heating output density (daily maximum heating
output divided by the length of the DH network)
- CHP production
- Share of small (< 30 kW) consumers
- Peak load utilization time
- Losses of DH network
1. Heat supply
Heat supply is the same as sold heat + losses.
The assumption used was that the bigger the company
the smaller the proportional pumping energy.
The situation is presented in figures 6a and 6b, where
only the companies which supply less than
2500 GWh/a (Helsinki not included) are shown.
Fig. 6a. Example Electricity used for pumping in relation to
the size of a district heating company, heat supply
02 500 GWh/a.
Fig. 6b. Electricity used for pumping in relation to the size
of a district heating company, heat supply 0300 GWh/a.

The average pumping energy is 0.6 per cent of the
heat supply. The bigger the company, the smaller the
proportional pumping energy.
Companies with exceptionally high pumping energy are
marked with a circle.
2. Length of the district heating network

Fig. 7a. Electricity used for pumping in relation to the
length of the district heating network, 01 300 km
Specific pumping energy vs. length of the DH net
Length 0 - 500 km
0.1 %
0.3 %
0.5 %
0.7 %
0.9 %
1.1 %
1.3 %
1.5 %
1.7 %
0 100 200 300 400 500
Length of the DH net, km
S
p
e
c
i
f
i
c

p
u
m
p
i
n
g

e
n
e
r
g
y

(
e
l
e
c
t
r
i
c
a
l

p
o
w
e
r

/

h
e
a
t

s
u
p
p
l
y
)
Specific pumping energy vs. heat supply
Supply 0 - 300 GWh/a
0.0 %
0.2 %
0.4 %
0.6 %
0.8 %
1.0 %
1.2 %
1.4 %
1.6 %
0 50 100 150 200 250 300
Heat supply, GWh/a
S
p
e
c
i
f
i
c

p
u
m
p
i
n
g

e
n
e
r
g
y

(
e
l
e
c
t
r
i
c
a
l

p
o
w
e
r

/

h
e
a
t

s
u
p
p
l
y
)
Specific pumping energy vs. heat supply
Supply 0 - 2 500 GWh/a
0.0 %
0.2 %
0.4 %
0.6 %
0.8 %
1.0 %
1.2 %
1.4 %
1.6 %
0 500 1000 1500 2000 2500
Heat supply, GWh/a
S
p
e
c
i
f
i
c

p
u
m
p
i
n
g

e
n
e
r
g
y

(
e
l
e
c
t
r
i
c
a
l

p
o
w
e
r

/

h
e
a
t

s
u
p
p
l
y
)
September 5
th
to September 7
th

286

Fig. 7b. Electricity used for pumping in relation to the
length of the district heating network, 0500 km.

Fig. 7c. Electricity used for pumping in relation to the
length of the district heating network, 070 km.

For big companies the proportional pumping energy is
almost constant 0.5 per cent of heat supply.
The longer the DH network, the smaller the proportional
pumping energy. The result is partly the same as in the
previous chapter: the bigger companies have smaller
proportional pumping energies.
If a company seems to have a high proportional
pumping energy in figures 7a7c it may be due to poor
heating density (lots of pipes in areas with not so much
consumers).
3. Heat density
Heat density is the heat supply divided by the length of
the district heating net.
Fig. 8. Electricity used for pumping in relation to the heat
density.

It is natural that in a DH network with not too many
pipes in proportion to sold heat the need for pumping of
DH water is lower.
FURTHER INFORMATION:
Pyry Finland Oy
PL 93 (Tekniikantie 4 A)
FI-02151 Espoo
Finland
antti.hakulinen@poyry.com
CONCLUSION
District heating networks enlarge and change
continuously and therefore the conditions of pumping
will also change. For that reason, it is important to
check every now and then if the actual operating point
of the pump is as designed and what the efficiency of
the present operating point is. The pumping could still
work technically well but the pumps could be operating
with low efficiency.
The most important issues in designing and operating
of district heating pumping are:
- A sufficient but not too big pressure difference
must be guaranteed for customers.
- There must be enough pressure in all parts of the
network at all circumstances and at the same time
the maximum pressure level must not be
exceeded.
- When designing pumping it is important to study
all possible pumping cases.
- Good operating point should be verified when
designing and operating pumps.
Pumping energy is dependent on certain parameters.
The best parameter is considered to be heat density on
which pumping energy is clearly dependent. And this is
a quantity every district heating company measures.
0.0 %
0.2 %
0.4 %
0.6 %
0.8 %
1.0 %
1.2 %
1.4 %
1.6 %
0 200 400 600 800 1000 1200 1400
S
p
e
c
i
f
i
c

p
u
m
p
i
n
g

e
n
e
r
g
y

(
e
l
e
c
t
r
i
c
a
l

p
o
w
e
r

/

h
e
a
t

s
u
p
p
l
y
)
Length 0 - 70 km
0.1 %
0.3 %
0.5 %
0.7 %
0.9 %
1.1 %
1.3 %
1.5 %
1.7 %
0 10 20 30 40 50 60 70
S
p
e
c
i
f
i
c

p
u
m
p
i
n
g

e
n
e
r
g
y

(
e
l
e
c
t
r
i
c
a
l

p
o
w
e
r

/

h
e
a
t

s
u
p
p
l
y
)
Specific pumping energy vs. heat density
0.1 %
0.3 %
0.5 %
0.7 %
0.9 %
1.1 %
1.3 %
1.5 %
1.7 %
0 1 2 3 4 5 6 7
Heat density, GWh/km
S
p
e
c
i
f
i
c

p
u
m
p
i
n
g

e
n
e
r
g
y

(
e
l
e
c
t
r
i
c
a
l

p
o
w
e
r

/

h
e
a
t

s
u
p
p
l
y
)
September 5
th
to September 7
th

287

The figure below is the same as the Fig. 8 added with a
red line to help the reader estimate the pumping energy
of his own plant. If the pumping energy is above the red
line some measures ought to be taken.

Fig. 9. Electricity used for pumping in relation to the heat
density of the district heating network + trend line

Figure 9 shows that on average the electricity needed
for district heating pumping should not be over 0.5 per
cent of the total energy supply (=sold+losses). If the
density (supply/length of the network) of the district
heating network is less than 3 GWh/km, the energy
needed for pumping may rise. In any case the
proportional pumping energy should be lower than
1 percent.
The following figure illustrates an example case in
which the heat density is over 2.5 GWh/km. The figure
can be utilized when estimating the losses in real
money if the proportional pumping energy is over the
average of 0.5 percent.
Value of "excess" pumping energy
Heat density > 2.5 GWh/km, Value of power 60 EUR/MWh
0
50
100
150
200
250
300
350
400
100 600 1100 1600 2100
Heat supply, GWh/a
V
a
l u
e
o
f
e
x
c
e
s
s
p
u
m
p
i n
g
e
n
e
r
g
y
, 1
0
0
0
E
U
R
/ a
Proportional share of pumping energy 0.8 %

For example, if the heat supply of the company is 1.1
TWh/a and the proportional pumping energy is
0.7 percent, the losses of unnecessarily high pumping
energy is 130 000 per year.
Some of the pumping energy is converted to heat. This
decreases the value of the losses.
In total, the potential savings in all Finnish district
heating companies are approximately 20 percent of the
current pumping energy, i.e. 30 GWh/a. This is
equivalent to a yearly saving of approximately
2 million.
REFERENCES
[1] DH statistics 2007, Energiateollisuus ry, 2008
Specific pumping energy vs. heat density
0.1 %
0.3 %
0.5 %
0.7 %
0.9 %
1.1 %
1.3 %
1.5 %
1.7 %
0 1 2 3 4 5 6 7 8
Heat density, GWh/km
S
p
e
c
i
f
i
c

p
u
m
p
i
n
g

e
n
e
r
g
y

(
e
l
e
c
t
r
i
c
a
l

p
o
w
e
r

/

h
e
a
t

s
u
p
p
l
y
)
0.1 %
0.3 %
0.5 %
0.7 %
0.9 %
1.1 %
1.3 %
1.5 %
1.7 %
September 5
th
to September 7
th

288

MODELLING DISTRICT HEATING COOPERATIONS IN STOCKHOLM AN
INTERDISCIPLINARY STUDY OF A REGIONAL ENERGY SYSTEM
D. Magnusson
1
, D. Djuric Ilic
2

1
Department of Thematic Studies Technology and Social Change, Linkping University,
SE-581 83 Linkping, Sweden
2
Department of Mechanical Engineering, Division of Energy Systems, Linkping University,
SE-581 83 Linkping, Sweden

ABSTRACT
In this paper, a combination of methods from social
science (interviews) and technical science (modelling)
have been used to analyse the potential for
cooperation in the present and future district heating
system in Stockholm. The aim of the paper is to explore
barriers and driving forces for energy cooperation in the
Stockholm district heating system and to analyse the
potential for combined heat and power generation in
the system. In the study it was found that with better
connectivity in existing systems, the annual system
cost would decrease by approximately 10 million , and
with new CHP plants a similar potential exists. There is
also a large potential for decreasing the local and
global emissions of CO2 with CHP plants. The results
from the interviews showed that the existing
cooperation has a long history and is working well
today. The advantages are higher supply security and
economic benefits, while disadvantages are a need for
more administration and control because of a more
complex system. That the barriers to cooperation are
seldom technical is another conclusion. With the
combination of methods, we have gained a better
understanding of the actual potential for the
development of the system.
NOMENCLATURE
CO
2
carbon dioxide;
LECO
2
local emissions of CO
2
;
GECO
2
global emissions of CO
2
;
CHP combined heat and power;
BCHP CHP plants fuelled by solid biomass;
NGCHP CHP plants fuelled by natural gas;
TGC tradable green certificates;
GHG greenhousegas.
1. INTRODUCTION
Swedish district heating has a long history and is today
one of the dominant heating forms with approximately
55% of market share, and an annual energy production
of approximately 55 TWh[1]. The first system was built
in Karlstad in 1948 and during the following decades
the largest cities built their own systems, as was the
case in Stockholm [2]. Because of the large amount of
energy in the systems, the fuel used in the plants has a
major impact on greenhouse gas (GHG) emissions,
and there is also a large potential for using combined
heat and power (CHP) technology in the systems. CHP
technology is becoming more important as a part of
creating sustainable energy systems, which for
example can be seen in the EU directive for promotion
of cogeneration [3]. In Sweden, as well as in
Stockholm, large investments are made in building new
CHP plants, in large part thanks to the electricity
certificate system [1]. Another important potential with
CHP generation is through the Electricity Directive of
1996, in which the EU prescribed common rules for
creation of an open and competitive electricity market
[4]. With a fully integrated electricity market, the
Swedish prices of electricity can be expected to
increase. However, as long as they are lower than
Europes there is a large potential for exporting
electricity. From a marginal power production
perspective, which will be discussed further in the
paper, there is a potential for decreasing global
emissions of CO
2
, if the exported electricity comes from
non-fossil fuels.
A large enough system is an important prerequisite for
investment in CHP plants, in order to take advantage of
the economy of scale of district heating and CHP
generation. In Stockholm, the largest urban region in
Sweden, there are already well-developed district
heating systems. The systems started as smaller units
that gradually have been interconnected and today
consist of three large networks. However, since there
are eight different energy companies in the city region,
a working cooperation between the energy companies
is important. With this in mind we will analyze how the
actors perceive existing and future cooperation. The
study is conducted with an interdisciplinary approach
where interviews have been combined with modelling
the systems' performance with present and possible
future interconnections, present plants and future CHP
plants, and finally with a hypothetical introduction of
natural gas. The aim of the paper is to explore barriers
and driving forces for energy cooperation in the
September 5
th
to September 7
th

289

Stockholm district heating system and to analyse the
potential for CHP generation in the system.
2. CASE STUDY
There are three large district heating networks in
Stockholm that deliver more than 12 TWh of heat
annually, produced in some 70 heating plants [5].
Table I shows the heat production, types of base
production and installed heat and electricity capacity in
those networks. Six of the plants in the system are
CHP plants with total installed electricity capacity of
about 600 MW, which gives a possibility for production
of over 2 TWh of electricity annually [5].
Table I. Major district heating networks in Stockholm. [5]
South-
central
North-
west
South-
east
Heat production in
the year 2005 [TWh]

9.4

2.2

0.53
Installed heat
capacity [MW]

4000

700

300
Installed electricity
capacity [MW]

493

105

20
Base production CHP
waste,
CHP
coal
BCHP NGCHP

3. METHODS
A combination of methods from social science and
technical science has been used; modelling with
MODEST and semi-structured interviews with
representatives from the largest energy companies.
MODEST is an energy-system optimisation model with
time-dependent components that was developed at
Linkping University in Sweden. MODEST uses linear
programming to calculate the most profitable
combination of existing and potential new facilities and
shows which investment options are financially
viable [5].
3.1 The structure of the interviews
The interviews were conducted during the spring of
2009, with representatives from the five largest energy
companies producing and/or distributing district heating
in the Stockholm region. These are Fortum, Norrenergi,
Sderenergi, E.ON and Vattenfall. The representatives
were chosen by the companies themselves, since they
could better decide who would be most appropriate to
answer questions regarding interconnections,
cooperation and future strategies. We decided to let the
respondents remain anonymous, since one of the
interviewees wanted this. The interviews were semi-
structured, as we had similar questions for most of
them, although with some differences depending on the
company. Because a semi-structured interview is a
qualitative method, the possibility of using open
questions is an advantage, and since we are interested
in a specific situation, the interviewees have the
chance to give their opinion. It also gives the
opportunity to analyze the answers in different ways, to
understand the opinions expressed [7].
3.2 Modelling Stockholms district heating system
Based on the data from Open district heating network
in greater Stockholm [5] a model of Stockholms
district heating system has been constructed.
Purchases and sales prices of electricity, taxes and
tradable green certificates (TGC) are included in the
model (Table II) as well as the operating and
maintenance costs for all plants and fuel prices.
However, due to agreements with the contact persons
from the district heating companies, the prices for fuel
are not presented in the paper.
Table II The average annual purchases and sales prices
of electricity, including all taxes and TGC. [8], [9], [1]
Current price of electricity [/MWh]
Purchase Sale Sale with TGC
included
70.10 35.46 67.56
European price of electricity [/MWh]
Purchase Sale Sale with TGC
included
83.30 48.65 80.77

Carbon dioxide emissions used in this paper are shown
in Table III [10]. However, since the greenhouse effect
is a global problem, carbon dioxide (CO
2
) emissions
are not simply analysed from a local perspective but
also in regard to a global perspective. The global
emissions of CO
2
(GECO
2
) of the system are calculated
with the assumption that electricity produced in the
plants is going to replace marginal power production in
the integrated European electricity market. Since coal-
fired condensing power plants have the highest
variable cost compared with other sources of electricity
in the EU, they work as the marginal power
production [11]. When assuming that the coal-fired
condensing power plants have an electricity efficiency
of 33%, each megawatt-hour of electricity generated in
such a plant releases approximately one tonne of CO
2
.
According to that, any increase in electricity production
in Stockholms district heating system can lead to
reduced production in the marginal coal condensing
power plants, and consequently to a reduction of
one tonne of CO
2
emissions. However, it is necessary
to mention that considering the EU Emissions Trading
Scheme (EU ETS), the decrease of CO
2
emissions in
September 5
th
to September 7
th

290

electricity production sector does not necessarily to
lead to reduction of GECO2 [12]. But the marginal
electricity concept still has significance for future
measurement of and planning for future limitations of
CO
2
emissions and the future trading system.
Table III. Net emissions of CO2 [10].
Fuel Emissions |kg/MWhfuel|
Oil 280
Coal 330
Waste 100
Biomass 0
Electricity 950
Natural gas 230

3.3 Description of chosen scenarios
Nine different scenarios have been analysed
considering the possible future cases (Table IV), with
special attention to economic and environmental
aspects.
The existing district heating system (scenario 1) and
the system with three new CHP plants that are planned
to bee built according to the interviews and documents
(scenario 4) have been analysed. Since the base
productions in the networks differ, the differences
between the productions costs in different parts of the
system are notable. Because of that, in both cases
(scenarios 1 and 4) the influences of a better
connectivity between networks have been studied
(scenarios 2 and 5).
Table IV. List of the chosen scenarios.

Sc.
Plants
in the
district
heating
system

Connectivity

Electricity
price

TGC
1 existing existing Nordpool exist
2 existing one
network
1

Nordpool exist
3 existing existing EU exist
4 + new
CHPP
existing Nordpool exist
5 + new
CHPP
one network Nordpool exist
6 BCHP one network Nordpool exist
7 BCHP one network EU exist
8 BCHP one network EU do not
exist
9 NGCHP one network EU -
1) Interconnections between the south-central and the
north-west networks have been introduced as well as
interconnections between south-central and south-east
networks. Capacities for existing pipes have been
increased.
Since electricity generation will probably be the primary
production in all district heating companies in the
future, when the Swedish electricity price becomes as
high as the typical European price, in scenarios 6-9 our
research focuses on the cogeneration potential in
Stockholm's district heating system. Scenario 1 has
been used as a reference scenario for scenario 6.
Scenario 3, where the influences of a higher electricity
price on the system with the existing plants have been
analysed, has been used as a reference scenario for
scenarios 79. Scenarios 6-9 are analysed as possible
future cases that may exist more than 10 years from
today. Because of that, all plants in the scenarios are
new so the investment costs for all plants are
considered. While in the scenarios 6, 7 and 8 the
system consists of 31 CHP plants fuelled by solid
biomass (BCHP), there are a total of 46 CHP plants
fuelled by natural gas (NGCHP) in scenario 9. In
scenario 9 it is assumed that the natural gas network
exists along the Swedish east cost.
The characteristics of the CHP plants that have been
integrated in the model of the district heating system in
scenarios 49 are presented in Table V [13].
Table V. The characteristics of the new integrated CHP
plants in scenarios 49 [13].

Sc.
Technical characteristics
Fuel Electrical
output
Fuel
efficiency
*
MWe %

45
biomass 30 110 0.45
waste 20 91 0.32
biomass 80 110 0.46
68 biomass 80 113 0.51
9 natural gas 150 89 1.41
Economic characteristics
Process
plant cost
Operating and maintenance
/KWe % of PPC /MWh fuel

45
2 745 1.5 2.45
5 440 3 9.31
2 110 1.5 2.45
68 2 110 1.5 2.45
9 715 2.5 0.9
* electrical/thermal output
3.4 Previous studies
Two studies regarding Stockholms district heating
system were done in the years 2005 [10] and 2006
[14], and the results showed that benefits for better
connectivity between some parts of the system existed.
September 5
th
to September 7
th

291

It was also shown that if all plants in the system are
replaced with BCHP plants, with an electricity-to-heat
output ratio 0.46, up to 10TWh electricity can be
produced and the potential for decrease of GECO
2
of
the system would be 3 tons CO
2
annually. If all plants
in the system are replaced with NGCHP plants, with an
electricity-to-heat output ratio 1.2, the electricity
generation in the system can increase to 11TWh and
the potential for decrease of the GECO
2
of the system
would be about 5 tons CO
2
annually. However, since
these two studies were done, a new connectivity
between networks has been built and the total installed
electricity capacity in the system has increased by 20%
[5]. Furthermore, new CHP technologies are constantly
being developed, which enable greater electricity
efficiency and consequently greater benefits from
economic, energy and environmental viewpoints.
Regarding interconnection and cooperation of DH
systems, some studies have been conducted in a
Swedish context. However, none of them have focused
on cooperation between energy companies. They have
instead focused on cooperation between energy
companies and industry.
Thollander et al. [15] found that technical aspects are
seldom barriers to cooperation. The barriers are rather
risk, different aspects of information during
negotiations, and other social factors such as inertia
among personnel. Driving forces have been economic
factors such as an aim for lower costs and means of
control, as well as environmental values. In a study with
a similar aim, Fors [16] found the same results, that
technical aspects are seldom barriers. Information
during negotiations, stable contracts and the
importance of involving the personnel at the plants in
the process are important factors. It is also important
that the cooperation benefits both parties. Grnkvist
et al.[17] reached a similar conclusion in a study that
emphasises the importance of the willingness of people
on both sides to cooperate. The main advantages of
the cooperation are lower costs and benefits for the
environment, while the main disadvantages are less
flexibility as both parties work under contracts.
Historically, interconnection of technical systems has
been seen in the theory of Large Technical Systems as
one way for systems to grow. Systems start in a local
context, but when the technology is transferred to other
geographic areas, the systems grow and can then be
interconnected as they often have grown into each
other. Interconnection of systems can also be
explained through the fact that larger systems have a
higher load factor and better economic mix [18], [19],
[20].
4. RESULTS OF THE SCENARIOS
The results from the scenarios are presented in
Table VI and Table VII.
According to the optimisation results, if better
connectivity is introduced, some economic benefits
exist. In both cases the case with only existing plants in
the system and the case where the new plants are
introduced in the model (scenarios 2 and 5) the
decrease in system costs would be about 10 million
annually. The potential for decrease of the
environmental impact of the system is more notable. If
better connectivity were introduced in the system
today, the biomass share in total fuel use would be 8%
higher and consequently both the local emissions of
CO2 (LECO2) and GECO2 of the system would be
about 0.25 million tons lower annually. The potential for
decrease of GECO2 of the system if better connectivity
is introduced after the building of new CHP plants
(scenarios 4 and 5) is 0.4 million tons annually.
Table VI. Results for the scenarios economic aspects.

Sc.

Annual
system
costs

CHP heat
production
share

Electricity

Annual
production
The
income
from
electricity
million % TWh Million
1 258 47 2.30 122
2 245 47 2.31 125
3 243 48 2.35 150
4 204 58 2.96 164
5 192 62 3.15 176
6 403 424
7 344 100 6.39 482
8 546 281
9 504 100 17.66 777

As the electricity price increases, the system would
earn extra income from the electricity sold, and thus the
heat production cost would decrease (scenarios 1, 3).
This gives an even bigger advantage to CHP
generation compared with pure heat production.

September 5
th
to September 7
th

292

Table VII. Results for the scenarios environmental
aspects.

Sc.

Biomass
share in
the
system

LECO
2

GECO
2
of
the system

% [million
tons/year]
[million
tons/year]
1 48 2.50 0.32
2 52 2.25 0.06
3 49 2.46 0.23
4 52 2.12 0.69
5 55 1.91 1.08
6
7 100 0 6.07
8
9 0 7.80 8.98

The income from the electricity sold in scenario 3 is
about 30 million higher then the income in scenario 1,
and because of that the system cost is 6% lower. The
difference between the electricity production in
scenarios 1 and 3 is not significant, but in spite of that,
the decrease of GECO2 of the system in scenario 3 is
almost 100%. The reason is higher biomass share in
the total fuel used in the system in scenario 3, and
consequently lower LECO2 in the system.
The introduction of three new plants in the system
(scenario 4) would lead to a significant reduction of the
heat production cost compared with the system today.
The income from the electricity sold would be 35%
higher and, as a result, the annual system costs would
be 20% lower. This confirms that heat production in
CHP plants has a major influence on the economic
efficiency of the district heating system. With the
assumption that the electricity produced would replace
the marginal electricity in the European electricity
market, reduction of GECO2 of the system would be
almost 1 million tons annually.
If all plants in the system are BCHP (scenarios 68) or
NGCHP (scenario 9) plants, the annual electricity
production would be as high as 4.5% and 12% of the
total electricity production in Sweden, which was about
145 TWh in the year 2008 [1]. The annual income from
the electricity sold in those scenarios is much higher
then the income from the electricity sold in the other
scenarios. In the scenarios with typical European
electricity price, (scenarios 79), the income from the
electricity sold is 220%, 90% and even 420% higher
then in scenario 3, where the system with the existing
plants is analysed with the higher electricity price. It is
also notable that in scenarios 6, 7 and 9 the annual
income from electricity is higher than the annual
system costs. However, since all plants in those
scenarios are new, the total investments are high.
Because of that, if the analysed time period is just
10 years, the annual system costs are much higher
then today.
The lowest GECO2 of the system are in the scenarios
where all plants in the system are BCHP (scenarios 6-
8) and NGCHP (scenario 9) plants. In those two cases
GECO2 in Sweden, which is about 60 million tons
annually [21] would be reduced by approximately 9%
and 15% respectively, with the assumption that the
electricity produced would replace the marginal
electricity. LECO2 in the system is highest in the
scenario where all plants are NGCHP but at the same
time GECO2 of the system is lower because of the high
electricity production.
5. RESULTS FROM THE INTERVIEWS
In the following section the results from the interviews
will be presented. The interconnections between the
systems make it possible to cooperate regarding heat
production and distribution.
5.1 The system today
The interviews show that the interconnections have a
historical background. Most of them were made during
a period when a regional energy company called
STOSEB (Greater Stockholm Energy Company)
existed, where the municipalities, which to a large
extent owned the systems then, were represented. The
main reason for the interconnections then was supply
security. When the systems were interconnected, the
companies could help each other during stops, and this
is still the case. All representatives say this, and the
representative from Sderenergi expresses it this way:
At the same time it is a common good. It is good that
the systems are interconnected. It is an extra security if
one plant should stop for some reason [22].
The advantages historically and foremost today are
also economic. The emissions trading makes it
advantageous, since the companies can use the
production better by making capacity trades and even
out the production cost between the companies:
We see that we can use existing production more
effectively. Most of the trades are a trade to mid-price
so to speak. You can say that we split the profit.
Capacity trading (effektkp) is also common. Like we
have here with Sderenergi, we have partly
a production cooperation and partly we buy capacity.
They have more capacity than they need today [23].
September 5
th
to September 7
th

293

The main advantage with capacity trading is to avoid
peak load, which often is oil-based, which is costly both
for the fuel price but also because of the emissions.
Another point is that, as the Fortum representative
said, it is possible to even out effect between systems.
One system may have cheaper base load than the
other, and for tax reasons it may be cheaper to buy
from the other than to use peak load.
One factor that is pointed out for a successful
cooperation is that both parties can benefit from it. As
in all business, it is important that the cooperation be
correct from a business standpoint and that both parts
are satisfied [24].
The extent of cooperation varies between the
companies. Some have more extensive cooperation
with daily trades, like Fortum and Sderenergi, while
others, for example Fortum and E.ON, do not trade
every day. In the latter case, they normally do not trade
as much during winter, although sometimes when peak
load is needed it is decided quickly [25]. Another
advantage with the interconnections is that the
companies can cooperate regarding revisions of the
plants. While one company has revision during
summer, the other can produce for the other company.
The factors that are seen as barriers are seldom
technical. The companies think that the technical
problems often can be solved while making the
interconnection and at that point there is a need to
negotiate certain aspects. For example, who provides
the electrical energy for the pumps and takes
responsibility for the regulation of the water pressure in
the culverts and repairing the system in a joint part of
the system? However, this is often solved:
Yes, the other things we can handle while building the
technical parts. At that point we hopefully have
identified all technical barriers so that they can be
taken into account. They should not appear during
production. Settlement of account and such things,
they are not a big problem although complicated.
However, it is nothing that makes you pass on a
profitable cooperation [23].
In the above quote, we see one of the disadvantages
with today's cooperation, on which all the companies
agree, and that is the settlement of accounts. It is
complicated to control the systems and the trades, and
it requires staff to do so.
5.2 Barriers towards more co-operations
In the interviews the companies expressed satisfaction
with the present cooperation. Few actual barriers as
such were expressed, except the ones that todays
situation creates. For example, it is almost
geographically impossible to expand the systems to
smaller systems nearby. As could be seen in the
modelling, the systems are also already well
interconnected:
Yes, the principal structure is already established. ()
It is this connection, between the central and the
northwest system, it is the only one. That is not solved
yet [23].
This particular connection would interconnect the two
main systems, and has been discussed in some
investigations [26], [27]. However, it is yet to be done.
This connection is most important for Fortum, as for
example E.ON thought that it made little difference to
them.
The other main connection still missing is connection
between the south system and Vattenfall's system in
the southeast. Vattenfall thinks that the question has
been raised on occasion, although never realized. They
give no specific reason for this; they state that all
cooperation is important and that different
investigations have shown the advantages, although it
is difficult to quantify what it means practically [28].
Stockholms Energi (now Fortum) previously owned one
of the plants, and there were plans to interconnect the
systems then. Fortum gives no explanation for why the
interconnection has not been done earlier or now.
Although no direct comments regarding the lack of
interconnection were made, one of the interviewees
who previously worked at Vattenfall said that there was
an opinion at Vattenfall that they prefer to keep to
themselves, without interconnections, and should not
work towards cooperation. Comments without a
specific direction were also expressed in interviews that
there was a lack of will to cooperate from some
companies. There is also a history of rivalry between
Vattenfall and the former Stockholms Energi [29]. It is
possible that this rivalry stills exists. Fortum also
expressed opinions about the fact that other companies
are building their own CHP plants instead of trying to
find regional solutions.
5.3 Building CHP in the system
As seen in the scenarios, in the near future in Sweden
many CHP plants are planned and will start to be built.
In Stockholm most of the companies have plans for
CHP, and two of them have already built in the last
years, for example Igelsta (Sderenergi) and Jordbro
(Vattenfall). Other companies are making plans, such
as Norrenergi, EON and Fortum. The reasons for
building CHP are varied, but the most clear is that they
see economic advantages in selling electricity, and our
stagnating heat load ahead. By selling electricity there
is a possibility to keep profits high, even with a
stagnating heat load. The system is also relatively old
and well established; the potential for further
connections are getting smaller as saturation in the
September 5
th
to September 7
th

294

heating market for district heating makes it more
difficult to expand:
A rough rule of thumb has been that the expansion with
new customers that have been, (...), has been eaten up
by the efficiency we could achieve together with the
customers in their buildings. So basically, the heat load
has been static in our area for quite some time. ()
[The reason for building CHP] is the electricity. We,
as the producing company, have the problem that we
can not expand. We have our two customers and
district heating is not a new thing in the municipalities
so the chance of getting new customers is limited [22].
The other representatives are of a similar opinion, that
a stagnating load can be expected, and CHP is a way
to keep profits high. The Swedish certificate system
also makes it advantageous to build new bio-fuelled
CHP-plants. Another reason, arguably of a more
rhetorical character, is that building CHP is more
economically and environmentally correct since the fuel
efficiency is higher with CHPs. As scenarios 48 show,
there is major potential for reducing local and global
CO
2
emissions.
In the interviews we also asked questions about the
possibility of an introduction of natural gas in the
region. Investigations have been made earlier by the
above mentioned STOSEB; however, the plans never
came to reality. Generally the representatives did not
think that an introduction would happen. Since most of
them also have strategies to be climate neutral, natural
gas probably is not an option. The large investments in
infrastructure are another barrier:
These are such large infrastructure investments and
natural gas is not especially cheap either. It is difficult
to come in with natural gas in this energy system. It is
rather stable [23].
What the representative here points at is also the
inertia in the system. In LTS terms it is called
momentum: as the system is stable, it is difficult to
change the structure [18], [19].
6. CONCLUDING DISCUSSION
The study has shown the advantages of an
interdisciplinary approach. Advantages with
interconnections and CHP have been shown in the
modelling; however, as there are many different actors
involved, there is a need for a will to cooperate. The
interconnections have a historical background, with an
aim for higher supply security, and today most of them
continue to cooperate, despite the fact that the
structure and ownership of the companies in some
cases have changed since the deregulation of the
electricity market in 1996. As previous studies have
shown, the main advantages with cooperation have
been economic, as is also the case in this system. The
actors say that they can optimise the system's
performance, and our scenarios have shown that more
cooperation could benefit them even more
economically. Even though the gain is not extremely
high, since the lower system cost would be
approximately 5%, there is potential. However, since
there seems to be reluctance to cooperate between
some actors, it is difficult to fulfil the potential.
Advantages with the cooperation are said to be a
possibility to even out the production in the system and
thus avoid peak load. The disadvantages with the
cooperation are the need for more administrative work
to control the system and the trades; the control of the
system becomes more complex. This study also
confirms previous studies that have pointed out that
technical aspects are seldom barriers to cooperation.
Most things are solved while the systems are being
interconnected, and the will of the persons involved to
cooperate is important.
There is a large potential in building new CHP plants,
both from an economic and an environmental
perspective. If all the plants in the system were
replaced by BCHP or NGCHP, the electricity produced
could make up to 4.5 or 12% of total Swedish electricity
production, based on the fact that total production in
Sweden in 2008 was 146 TWh [1]. The reason that the
difference between the electricity productions in those
two cases is so large is a big difference between the
electrical/thermal outputs (see Table V). The intro-
duction of NGCHP is a less likely future since it can be
considered only with the assumption that the natural
gas network already exists along the Swedish east
cost. On the other hand, introducing more BCHP in the
district heating system would increase the systems
dependence on biomass availability and the heat
production cost would become highly sensitive to the
solid biomass cost. The actors are highly aware of the
potential for CHPs. Since they are expecting a
stagnating heat load, the sale of electricity is a way to
keep profits high. However, none of them think that
natural gas will become a reality in the near future, and
even if it did, the introduction is expected to be
somewhat problematic, since the fuel can be
considered fossil fuel and substantial infrastructure is
needed.
The study has shown a potential for decreased LECO
2

and GECO
2
. The largest potential from a local
perspective is from BCHP; so, since the LECO
2
would
be low and with high electricity production, the potential
for lower GECO
2
would exist. The high electricity-to-
heat output ratio in NGCHP has a high potential for
decreasing GECO
2
of the system. If all plants in the
system would be replaced with NGCHP the GECO2 of
the system would be -9 million tons annually. However,
in that case LECO
2
would be much higher than today.
September 5
th
to September 7
th

295

Modelling of a system gives one side of the truth, as
does interviewing the actors involved. When combining
the methods there is a possibility of getting a better and
deeper understanding of the actual potential for
cooperation. The historical and social aspects cannot
be neglected; they can in many cases explain why
potentially beneficial cooperation is or is not done,
while modelling can show the actual potential.
7. ACKNOWLEDGEMENTS
This article has been carried out in two PhD projects in
the Energy System Program, financed by the Swedish
Energy Agency. The authors would also like to thank
Jenny Palm (Linkping University) and Louise Trygg
(Linkping University) for valuable comments on the
paper.
8. REFERENCES
[1] Statens energimyndighet (Swedish Energy
Agency), Energy in Sweden 2009. Swedish
National Energy Administration, ET 2009:30,
Eskilstuna, Sweden (2009)
[2] S. Werner, Fjrrvrmens utbredning och utveckling
(The development and expansion of district
heating), Vrmefreningen, Stockholm (1989)
[3] European Union, Directive 2004/8/EC, Directive on
the promotion of cogeneration based on a useful
heat demand in the internal energy market,
(2010), homepage:
http://www.managenergy.net/products/R81.htm,
2010-04-28
[4] European Union, , Directive 96/92/EC Second
report to the Council and the European Parliament
on harmonisation requirements concerning
common rules for the internal market in electricity,
(2010), homepage:
http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?
uri=CELEX:32003L0054:EN:NOT , 2010-04-28
[5] B. Dahlroth, ppnade fjrrvrment i
Storstockholm (Open district heating network in
greater Stockholm), Stockholm, Sweden:
Fastighetsgarna Stockholm, (2009)
[6] D. Henning, Optimisation of Local and National
Energy Systems. Development and use of the
MODEST model. Dissertation No. 559. Linkpings
universitet, Linkping (1999)
[7] S. Kvale and S. Brinkmann, Den kvalitativa
forskningsintervjun (The qualitative research
interview), Studentlitteratur, Lund (2009)
[8] Nordpool. Nordpool electricity spot market, (2009),
homepage: http://www. nordpool.com. 2009-12-18.
[9] M. Melkerson and S-O. Sderberg, Dynamiska
elpriser elprissttning p en integrerad europeisk
elmarknad (Dynamic electricity prices pricing in
an integrated European electricity market in
Swedish), Sweden: Institute of Technology, Dept of
Mech Eng, Linkping University, Linkping (2004)
[10] N. Levinson and R. Freiman, Optimal kraftvrme
och ntinvestering I Stockholms fjrrvrmesystem
(Optimal CHP and district heating network
investments in Stockholm), Linkping University,
Linkping (2005)
[11] Statens energimyndighet (Swedish Energy
Agency). Marginal elproduktion och CO
2
-utslpp i
Sverige (Marginal electricity production and CO
2
-
emissions in Sweden, in Swedish), Swedish
National Energy Administration, ER 14:2002,
Eskilstuna, Sweden. (2002)
[12] E. Dotzauer, Greenhouse gas emissions from
power generation and consumption in a nordic
perspective, Energy Policy 2010, Vol. 38(2), pp.
701-704.
[13] H. Hansson, S-E. Larsson, O. Nystrm, F. Olsson
and B. Ridell, El frn nya anlggningar (Electric
Power from New Plants), Elforsk, Stockholm
(2007)
[14] M. Danestig, A. Gebremehdin and B. Karlsson,
Stockholms CHP potential An opportunity for
CO
2
reductions? Energy Policy 2007, Vol. 35(9),
pp. 4650-4660.
[15] P. Thollander and I-L. Svensson, Vgen till
framgngsrika vrmesamarbeten en fallstudie
(Road to succesful heating co-operations a case
study), In: L. Trygg, L. et al (2009) optimala
fjrrvrmesystemi symbios med industri och
samhlle, Rapport 2009:13, Svensk fjrrvrme
(Swedish district heating association) (2009)
[16] J. Fors, Spillvrme frn industri till fjrrvrment
sammanfattning av intervjuer p 5 orter, (Excess
heat from industry to district heating systems)
Rapport 2004:5, Svensk fjrrvrme (Swedish
district heating association) (2004)
[17] S. Grnkvist and S. Sandberg, Driving forces and
obstacles with regard to co-operation
betweenmunicipal energy companies and process
industries in Sweden, Energy Policy 2006, Vol. 34,
pp. 15081519.
[18] T.P. Hughes, Networks of Power: Electrification in
Western Society 1880 1930, John Hopkins
University Press, Baltimore (1983)
[19] B. Joerges, Large Technical Systems: Concepts
and issues, In: R. Mayntz and T.P. Hughes, The
September 5
th
to September 7
th

296

Development of Large Technical Systems,
Campus Verlag, Frankfurt (1988)
[20] J. Summerton, Changing Large Technical
Systems, Westview Press, Boulder, CO (1994)
[21] SCB (Statistics Sweden). Utslpp av vxhusgaser,
(2010), homepage:
http://www. scb.se. 2010-02-17.
[22] Sderenergi, Production manager, 090318
[23] Fortum, Site manager and Senior advisor, 090325
[24] Norrenergi, Production manager, 030304
[25] E.ON., Group manager production, 090323
[26] STOSEB, STOSEB 92 Energiframtider fr
Stockholms ln (STOSEB 92 Energy futures for
Stockholms County), STOSEB, Stockholm (1992)
[27] Fortum & Stadsbyggnadskontoret, Mjlighets-
studie: Ntintegration Storstockholm, (Possibility
study: Netintegration in greater Stockholm) Fortum
& Stadsbyggnadskontoret, Stockholm (2005)
[28] Vattenfall, Head of business development and
Senior advisor, 090320
[29] STOSEB, 25 Energiska r Om Stor-Stockholms
Energi AB (25 Energic years about greater
Stockholms Energy AB), STOSEB: Stockholm,
(2003)

September 5
th
to September 7
th

297

CUTTING COSTS OF DISTRICT HEATING SYSTEMS
BY USING OPTIMIZED LAYING TECHNIQUES
Alexander Goebel
1
, Dr. Stefan Holler
1

1
MVV Energie AG, Mannheim, Germany

ABSTRACT
The soil covered plastic jacket pipe is the common
state of the art laying technique in the district heating
sector: A preferable shallow trench is dug out and
backfilled with cable sand after the installation of the
two pipes. Alternative possibilities concerning the
digging of the trench, the backfill and the piping itself
are evaluated in this paper. Results show, that an
optimized laying technique can save construction or
running costs under the right boundary conditions:
Backfill materials with insulation properties can reduce
the heat losses by about 25 %. Using glass-reinforced
plastic pipes (GRP) instead of steel pipes leads to
pump energy savings of about 40 %.
INTRODUCTION
In the first place, excavation costs could be cut by
digging smaller and shallower trenches. However, this
is only possible if the location of the construction site is
appropriate. In an urban area the situation is com-
pletely different from a rural area concerning space and
regulations. The paper describes the boundary con-
ditions and compares different methods from the tech-
nical as well as the economical perspective using the
example of the district heating system in Mannheim,
Germany.
The second approach which will be presented in the
paper is the potential to reuse the excavated material
and to use self-compacting material when refilling the
trench. Furthermore, it is also possible to use new
materials with better insulation properties in order to cut
down heat losses. In the paper the different properties
of the new materials will be compared and evaluated.
A third possibility to reduce costs is the use of
specialized piping systems wherever possible.
Nowadays a wide range of products is available on the
market. In many cases a specialized system fits some
applications better than a standard system does. Not
only insulation properties but also compensation,
ductility and friction losses are important characteristics
of modern piping systems. In the paper it will be shown,
how costs could be reduced by using less or no
compensation measures (cold laying, flexible pipes,
fibre pipes), by avoiding welding measures (flexible
pipes for house connections, fibre pipes) or by reducing
friction losses (fibre pipes).
MATERIALS AND METHODS
The cost saving potentials of the alternatives, con-
cerning the digging of the trench and the backfill, are
mainly evaluated by outlining the results of research
reports. Calculations are used in order to estimate the
insulation properties of special backfill material. Also
the cost saving potentials of pipes with low friction
losses are evaluated with simple equations.
RESULTS
Reuse of the excavated material [1], [4]
Earlier research activities have proven, that plastic
jacket pipes could be used with backfill material
showing a greater grit size than cable sand. Special
protection material is available not only for the muffles
but also for the pipes. Field tests have shown, that
there are promising money saving potentials because
of the significant reduction of transport and disposal
costs. A consideration of reusing the excavated soil is
also reasonable from an environmental point of view.
Following points are important, when it comes to an
evaluation of this possibility at an individual
construction site:
- the grading of the excavated material
- sandy or cohesive ground
- compacting properties
- the friction between the ground and the jacket pipe
- protection measures for muffles and the pipe
- underground construction regulations
- a place for the storage of the excavated material
(beside the trench, container or any place near the
construction site)
- formation of dust (especially in the summer)
- contamination
- pH-value
- an improvement of the excavated material with
lime (especially with cohesive ground)
- a removal of the coarse material
- a separation of the material, if the ground is split up
in different layers
The use of self-compacting material [1], [4], [5]
It is important to distinguish between the following two
types of self-compacting material:
- stabilised sand mix
September 5
th
to September 7
th

298

- excavated material mixed with water and special
additives in order to get a self-compacting
behaviour
The use of self-compacting materials offers a wide
range of advantages and applications:
- it is possible to dig out a narrower trench, because
no machines are needed for the critical compaction
around the pipes
- the backfill process is significant faster after the
trench is filled up, it takes normally only one day
until the material is hard enough to walk on
- self-compaction is more reliable within difficult
conditions (many crossing pipes etc)
- without the use of compaction machines, buildings
nearby the construction site are stressed less (no
vibrations)
- there is less inconvenience for residents living
nearby the construction site, because of the noise
reduction
- in combination with the pipeline laying technique,
the sheeting can be omitted, because nobody
needs to work in the trench
A common problem is the local availability of the
technology. The price is also an issue, if the reason of
the application is the approach to save money.
Another problem concerning the dimensioning of the
compensation measures is the bad predictability of the
friction between the jacket pipe and the self-
compacting material. Depending on whether the pipes
are taken into service during or after the hardening
time, which is about a month long, a more or less
crucial tunnel effect is observed [4].
The reuse of the excavated soil as base material is
more elegant, than the stabilised sand mix, because of
the recycling aspect. Research projects have even
shown that sharp particles are less problematic,
because they are enclosed in the self-compacting
mass. An advantage of the stabilised sand mix is the
easier application.
If the district heating line does not run under a street,
compaction measures around the pipes can be avoided
simply by watering the cable sand, which is filled in
layers into the trench.
Cost saving potentials of backfill material with
insulation properties
If a reduction of the heat losses comes into consi-
deration, the change to a higher insulation series is
evaluated. Calculations show that in most cases an
economical justification is not given for this measure.
The idea of filling the trench around the pipes with
material that provides an additional insulation seems to
be promising. Like in the case of the self-compacting
material, the local availability is the greatest problem.
An economical justification is only achievable, if the
transport costs are low and the heat price is high. From
a technical point of view, the compaction behaviour has
to meet the requirements and regulations. The jacket
temperature must not exceed the maximum of 50 C
and the friction between pipe and the material should
be in the common range.
A calculation method for heat losses of plastic jacket
pipes is described in EN 13941 ANNEX D [2]. Figure 1
shows the influence of the thermal conductivity of soil
s
on the heat losses. Normally the value of
s
lies in
between 1,0 and 2,0 W/m*K [2]. The curve becomes
very non-linear below a value of 1,0 W/m*K. This
indicates that it is necessary to customise the
calculation method in order to get realistic results. The
heat losses are cut down by 30%, if the
s
is reduced
from 1,5 to 0,35 W/m*K.
40
45
50
55
60
65
70
75
0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Thermal conductivity of the soil
S
[W/m*K]
h
e
a
t

l
o
s
s
e
s

o
f

t
h
e

f
l
o
w

a
n
d

r
e
t
u
r
n

p
i
p
e

f

+

r

[
W
/
m
]

Fig. 1 Heat losses of a district heating line as a function of
s (DN 250, 120/50 C, Z = 0,6 m, C = 0,55 m,
i = 0,03 W/m*K)
The insulation material should solely be integrated in
the calculation as an additional thermal resistivity
(R
,embedment
), since the soil around the pipes is not
made completely out of it. The heat dependency of the
insulation foams thermal resistivity should also be
taken into account. Figure 2 illustrates, what is meant
with additional insulation layer.

Fig. 2 The different layers of the heat conductivity problem
September 5
th
to September 7
th

299

0
20
40
60
80
100
120
140
160
180
0 200 400 600 800 1000
nominal diameter [DN]
h
e
a
t

l
o
s
s
e
s

o
f

t
h
e

d
i
s
t
r
i
c
t

h
e
a
t
i
n
g

l
i
n
e

[
W
/
m
]
10%
15%
20%
25%
30%
h
e
a
t

l
o
s
s

r
e
d
u
c
t
i
o
n
without insulation material
with insulation material
reduction in %
embedment jacket steel h return i s
return
R R R R R R R
U
, , , ,
1
o
+ + + + + +
=
The modified calculation procedure is made up of the
following equations:
o t
o

=
i
D
R
1
(1)
|
|
.
|
\
|

=
i
o
steel
steel
D
D
R ln
2
1
,
t
(2)
|
|
.
|
\
|

=
o
insulation
i
i
D
D
R ln
2
1
t
(3)
|
|
.
|
\
|

=
insulation
C
C
jacket
D
D
R ln
2
1
,
t
(4)
|
|
.
|
\
| +

=
C
embedment C
embedment
embedment
D
s D
R
2
ln
2
1
,
t
(5)
|
|
.
|
\
|
+
+

=
embedment C
s
s
s
s D
R Z
R
2
) ( 4
ln
2
1
0

t
(6)

\
|
|
|
.
|
\
|
|
|
.
|
+
+
+

=
2
0
) ( 2
1 ln
4
1
embedment C
s
embedment
h
s D
R Z
R

t
(7)

(8)
(9)

( ) K T
average i C i i
50 0001 , 0
, 50 ,
=

(10)

( ) |
.
|
\
|

+
+ = +
soil
return flow
return flow r f
T
T T
U U
2
o o
(11)

( )
jacket steel i fluid average i
R R R R Q T T
, , , o
+ + + =

(12)

Fig. 3 Reduction of the heat losses for DN 15 to DN 1000
The average temperature of the insulation was
calculated with following equation and put back into
(10). A VBA script was used to iterate five times.
Fig. 3 shows the results of the calculation. The
insulation material was taken into account with a value
of 0,33 W/m*K (
embedment
). Around and in between the
flow and the return pipe a space of 0,2 m for each pipe
size was chosen (s
embedment
). The depth of cover had a
value of 1 m (Z). Like in the previous example, the flow
temperature was at 120 C and the return temperature
was at 50 C.
Fig. 3 shows, that savings are significant lower with
small diameters. Also the specific thickness of the PUR
insulation, which differs because of standardised jacket
pipe diameters, has an impact.
The heat losses of a DN 250 pipe are reduced from 67
to 51 W/m (24%). This means, that the heat loss
reduction is 6% less compared to Fig.1.
embedment jacket steel h flow i s
flow
R R R R R R R
U
, , , ,
1
o
+ + + + + +
=
September 5
th
to September 7
th

300

Since the use of an insulating backfill is more efficient
with huge diameters, a DN 700 pipe was chosen for an
example scenario. An annual average for the flow and
return temperature was taken into account. For the
calculation of the required backfill volume in the
embedment, 0.2 m space in every direction of the pipes
was estimated. It is important for the calculation to take
only the additional costs into account. That means the
price difference between cable sand and the insulation
material including the transportation costs.
Table 1 Scenario for insulation material

A common value of 6% was chosen for the required
rate of return (i).
The net present value C
0
was calculated with the
following equation:
+ + =
T
t
t
t
i R I C
1
0
) 1 ( ) (
(13)
The internal rate of return shown in Fig. 4 was
calculated with the IRR- function in Excel.
0
10,000
20,000
30,000
40,000
50,000
60,000
0.0% 1.0% 2.0% 3.0%
growth rate of the heat price
n
e
t

p
r
e
s
e
n
t

v
a
l
u
e

(
2
0

y
e
a
r
s
)

[
]
0.0%
2.0%
4.0%
6.0%
8.0%
10.0%
i
n
t
e
r
n
a
l

r
a
t
e

o
f

r
e
t
u
r
n
NPV
internal
rate of
return

Fig. 4 NPV and IRR of the scenario defined in Table 1
depending on the additional specific costs of the insulation
material.

The results in Fig. 4 show, that the additional specific
costs should be below 17 /m
3
in order to get a positive
value spread, assumed that the required rate of return
is 6 %. A reduction of specific costs of 5% (16 /m
3
)
results in an increase of the value spread by 1%. The
net present value after 20 years rises about 10,000 .
The growth rate of the heat price is difficult to predict,
but has an important influence within the given period
of 20 years. Presumably the heat price is mainly
influenced by emission trading, governmental subsidies
and the development of the fossil fuel price.
Other scenarios may estimate higher growth rates, but
in order to get realistic results, the rate was varied from
0% to 3,5%.
Table 2 gives an example of materials with low heat
conductivity that could be interesting to use as backfill.
It is obvious to look for natural products, because of the
price and environmental regulations.
Parameter Value Unit
Length of the district
heating line:
5000 m
Nominal diameter: DN 700
Average flow
temperature:
95 C
Average return
temperature:
50 C
Annual hours of
operation:
8760 h
Heat price (at the time of
the invest):
15 /MWh
th

Required volume of
insulation material:
1.85 m
3

Additional specific costs: 16 - 17 /m
3

Required volume for the
whole line:
9250 m
3

Heat loss with use of the
material:
69.6 W/m
Heat loss without use of
the material:
90.7 W/m
Energy savings: 23.3 %
Annual savings of the
whole line:
924 MWh
th

Additional investment (I): 148,000 157,000
Required rate of return: 6.0 %
Time of cash flow: 20 a
Annual growth rate of
the heat price:
0 3.5 %
17 /m
3

16 /m
3

16 /m
3

17 /m
3

September 5
th
to September 7
th

301

Table 2 Heat conductivity of different materials

Flexible pipes
Flexible pipe systems, which are defined in EN 15632
[3], are mainly distinguished by the material of the
service pipe:
- plastic (e.g. PE-Xa, Polybuten)
- copper
- mild steel
- corrugated stainless steel
Flexible pipes have a significant higher operating
pressure (16 or 25 bar) [3], if the service pipe is made
of metal. Also the maximum and continuous operating
temperatures differ much. Because of this fact,
systems with plastic service pipes could normally not
be used within huge district heating networks. In a
smaller network with lower flow temperatures, which
e.g. was built to distribute the heat of a small block
heating station, a system with a plastic service pipe
might have an application.
Flexible systems with a corrugated service pipe have
significant higher friction losses, which has to be taken
into consideration (dimensioning).
When it comes to money saving potentials, the most
important properties of flexible pipes are the following:
- less welding measures
- self-compensating
- less insulation work
- less work concerning the monitoring system
- less head access holes, because of the reduced
welding measures
- less risk of leaks, because of less weld joins
- faster laying of the pipes
Pipes with low friction losses
Service pipes made of glass-reinforced plastic (GRP)
have significant less friction losses than steel service
pipes. Because of their chemical resistance, GRP
pipes are used mainly in the chemical industry. It is
important do distinguish between filament-wound pipes
and centrifugally cast pipes. Because of the Poissons
effect almost no compensations measures are needed,
if filament-wound pipes are used. Centrifugally cast
pipes need to be compensated, but have an even
smoother inner surface, which means the lowest
possible friction losses. Also the temperature
resistance is a little bit higher. The greatest problem of
GRP pipes is the fact that the service life is cut down
by high temperatures in combination with high
pressures (derating factor).
Fig. 5 shows the possible savings, if a GRP pipe with a
surface roughness of k = 0,01 mm is compared to a
steel pipe with a roughness of k = 0,2 mm.
The following equations were used:
The calculations of the Reynolds number:

v
d w
Re

=
(14)
The value of the kinematic viscosity (v) was taken with
2.941*10-7 m2/s, the density () with 958.77 kg/m3
(water with 100 C and a pressure of 10 bar) [7]
The pipe friction factor was calculated with the
following equation [6]:
25 , 0
Re
68
d
k
0,11 |
.
|
\
|
+ =
(15)
The pressure loss was calculated with the following
equation:
2 d
L
p
2
w
= A
p
(16)

W/m*K
hard plaster [9]
600 kg/m
3
0.18
900 kg/m
3
0.30
1200 kg/m
3
0.43
1500 kg/m
3
0.56
light sediment natural stone [9] 0.85
porous rock, e.g. lava [9] 0.55
natural pumice [9] 0.12
bitumen [10]
2100 kg/m
3
0.70
as matter, 1050 kg/m
3
0.17
membrane, 1100 kg/m
3
0.23
expanded volcanic rock (perlite) [11]
loose perlite, 50 - 130 kg/m
3
0.07
perlite compressed with filaments,
170 - 200 kg/m
3

0.06
Thermosand

[8] 0.33
September 5
th
to September 7
th

302

The equation for the pump power:
n
p V A
=

pump
P
(17)

Equation (17) shows, that the correlation between
pump power and pressure loss is linear. The savings
are expressed as a percentage. They do not depend
on the diameter or length of the pipe, because only the
friction factor differs.
32.5%
35.0%
37.5%
40.0%
42.5%
45.0%
47.5%
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
flow speed [m/s]
r
e
d
u
c
t
i
o
n

o
f

p
r
e
s
s
u
r
e

l
o
s
s
Fig. 5 Pump energy savings of a GRP pipe

Another important aspect of GRP pipes are the joints:
the pipe ends are glued together with a two component
adhesive, which is heated up for the curing process.
This can be an advantage, because welding measures
on a construction site are often problematic (lack of
space, wind). Statistics show that in most cases leaks
are caused by bad weld seams [12].
CONCLUSION
A consideration of an alternative laying technique is
usually worth the work, because the money saving
potential is often higher than expected. It depends
strongly on the single project and the local boundary
conditions (heat prices, rural or urban area, availability
of technologies/materials etc), whether a different
technique makes sense from an economical point of
view.
Containing a bunch of alternatives in district laying
techniques, Table 3 gives a rough overview with a
simple rating. Techniques appearing in the table, which
are not discussed is this paper, are listed there,
because they are also belonging to the alternatives
and will be evaluated in future studies. When the word
alternative is used, it means every technical aspect,
which differs from the standard laying technique
defined in the abstract.

Table 3 Overview of alternative laying techniques
++ highly recommended to take into
consideration from an economical point of
view
+ a closer look seems promising
0 an economical benefit can be achieved, if
special boundary conditions are given
- because of technological or economical
reasons, the use of the technology is not
recommended
-- the technology is not available or can not
be applied reasonable under the given
boundary conditions

September 5
th
to September 7
th

303

G
R
P

p
i
p
e
s

f
l
e
x
i
b
l
e

p
i
p
e

p
i
p
e
l
i
n
e

l
a
y
i
n
g

t
e
c
h
n
i
q
u
e

s
e
l
f
-
c
o
m
p
a
c
t
i
n
g

m
a
t
e
r
i
a
l

r
e
u
s
e

o
f

e
x
c
a
v
a
t
e
d

s
o
i
l

c
o
m
b
i
n
e
d

l
a
y
i
n
g

w
i
t
h

o
t
h
e
r

s
u
p
p
l
y

p
i
p
e
s

s
t
a
c
k
e
d

l
a
y
i
n
g

i
n
s
u
l
a
t
i
o
n

a
s

a

c
a
s
t
i
n
g

c
o
m
p
o
u
n
d

i
n
s
u
l
a
t
i
n
g

b
a
c
k
f
i
l
l

t
r
e
n
c
h
-
l
e
s
s

l
a
y
i
n
g

t
e
c
h
n
i
q
u
e
s

nominal diameter
< DN 150
+ ++ 0 ++ + 0 + ++ - +
nominal diameter
> DN 150
++ -- ++ + ++ 0 ++ 0 ++ ++
new district heating line ++ + + ++ ++ ++ ++ - + ++
renovation measure 0 - 0 ++ ++ 0 -- ++ + --
construction site in an
urban area
+ ++ -- ++ 0 + + + 0 ++
construction site in a
rural area
+ + ++ 0 ++ ++ ++ + + 0
yet to be built housing
estate
+ + ++ + + ++ ++ - + 0
existing housing estate + ++ -- ++ + + + - 0 ++

NOMENCLATURE
s
the coefficient of thermal conductivity for
the soil
i
the PUR insulation
i,50 C
the PUR insulation at 50 C
C
the jacket pipe
steel
the service pipe
embedmen t
the insulating material used in the
embedment
heat transfer coefficient
R
h
insulance of the heat exchange between
flow and return pipe
R
s
insulance of the soil
R
i
insulance of the insulation material
R
,steel
insulance of the service pipe
R
,jacket
insulance of the jacket pipe
R
insulance of the convective heat transfer

inside the pipe
R
0
surface transition insulance
R
,embedment
insulance of the insulating material used
in the embedment
D
i
inner diameter of the service pipe
D
0
outer diameter of the service pipe
D
insulation
outer diameter of the PUR insulation
D
C
outer diameter of the jacket pipe
T
flow
flow temperature
T
return
return temperature
T
soil
temperature of the soil
T
i,average
average temperature of the PUR
insulation
T
fluid
flow or return temperature
U
flow
heat loss coefficient for the flow pipe
U
return
heat loss coefficient for the return pipe
September 5
th
to September 7
th

304

Z distance from surface to the middle of the
pipe
C distance between the centre lines of the
two pipes
s
embedment
space between the pipe and the trench
wall
f
+
r
heat loss per pipe pair

C
0
net present value
I investment
R
t
net cash flow (annual savings)
i required rate of return
T given period
t the time of the cash flow

k surface roughness of the service pipe
Re Reynolds number
w flow speed
d inner diameter of the service pipe
v the kinematic viscosity
the pipe friction factor
p pressure loss
L length of the pipe
the density of the heating water
P
pump
pump power
V
flow rate
pump efficiency
REFERENCES
[1] Alexander Goebel, Alternative Fernwrme-
Verlegesysteme, Mannheim (2010)
[2] European Committee for Standardization,
EN 13941 ANNEX D, Brussels (2009)
[3] European Committee for Standardization,
EN 15632, Brussels (2009)
[4] Dipl.-Ing. Heinz-Werner Hoffmann, Dipl.-Ing.
Torsten Ghler and Dr.-Ing. Manfred Klpsch,
Fernwrmeleitungsbau mit Recyclingmaterial, MVV
Forschungsbericht, Mannheim (2006)
[5] Dipl.-Ing. Heinz-Werner Hoffmann and Zoltan
Dioszeghy-Gnter, Neuartige Verlegetechniken fr
das Kunststoff-Verbundmantelrohr-System Band 1,
MVV Forschungsbericht, Mannheim (1995)
[6] Fratzscher et. al., Energiewirtschaft fr
Verfahrenstechniker, VEB Deutscher Verlag fr
Grundstoffindustrie, Leipzig (1982)
[7] VDI-Wrmeatlas, Verein Deutscher Ingenieure,
Heidelberg (2006), Dba 5, Dba 13
[8] KE KELIT Kunststoffwerk Gesellschaft m.b.H.,
Thermosand (Broschre), Linz (2006), p. 6
Heidelberg (2006), Ded 12
Heidelberg (2006), Ded 10
[11] Heinz Schmid, Excel mit VBA in der
Wrmetechnik, C. F. Mller Verlag, Heidelberg
(2008), p. 26
[12] Dipl.-Ing. (FH) Frank Espig, Schadensstatistik KMR
2007 des AGFW, article published in the
EuroHeat&Power (2008), issue 10

September 5
th
to September 7
th

305

ANALYSIS OF HEAT TRANSFER IN HEAT EXCHANGERS BY USING
THE NTU METHOD AND EMPIRICAL RELATIONS
O. Gudmundsson, O. P. Palsson and H. Palsson
1

Faculty of Industrial Engineering, Mechanical Engineering and Computer
Science Hjardarhagi 2-6, IS-107 Reykjavik, Iceland

ABSTRACT
Heat exchangers are widely used in domestic and
industrial applications involving transfer of energy from
one fluid to another, for example in district heating
systems. The wide usage underlines the importance to
have a good technique to detect if the effectiveness of
an heat exchanger is diminishing. There are number of
things that can cause diminishing effectiveness of an
heat exchanger, for example fouling, changes in fluid
properties as well as corrosion. In many cases the
fouling is a particular problem, for example when
geothermal water is used. Geothermal water is very
mineral rich which can cause serious fouling problems.
The method presented in this paper is simple and easy
to use and can be used to detect a diminishing heat
transfer coefficient in many types of heat exchangers,
in this paper the method is used on cross flow heat
exchanger. The method uses measurements of the
inlet and outlet temperatures as well as the mass flows,
these measurements are usually easy to gather under
normal operation. The method uses the well known
Number of Transfer Units (NTU) method as well as
empirical relations to estimate the overall heat transfer
coefficient, which is then statistically analyzed. The
data used in this study was gathered from a simulated
cross-flow heat exchanger where the overall heat
transfer coefficient was gradually decreased to
simulate diminishing effectiveness of the heat
exchanger. The conclusion of this study shows that the
derived detection method can detect fouling based on
the data from a simulated cross-flow heat exchanger,
with a good accuracy and consistency. Further analysis
on real data is scheduled.
INTRODUCTION
Heat exchangers are widely used in domestic and
industrial applications involving transfer of energy from
one fluid to another. General classification of heat
exchangers are parallel flow, counter flow and cross
flow. Their size and complexity can also vary greatly.
Their operating conditions can be classified into two
main classes, steady operation where mass flow and
temperatures are relatively constant and dynamic
operation where mass flow and temperatures can vary
greatly with time.
During operation it is important to have some
knowledge of the condition of the heat exchanger. For
a steady state condition it has proven to be relatively
simple, since analytical and empirical relations can be
derived for different heat exchanger types and used for
all necessary calculation regarding time invariant
conditions, see e.g. [1]. If a dynamic operation exists it
becomes more complex to monitor the condition of the
heat exchanger and more complex models are used,
see e.g. [2] and [3].
In this study, a mathematical model is used that has
been developed to simulate accurately the temperature
and flow transients in a cross flow heat exchanger. The
model is based on the finite volume method (FVM)
where a mathematical representation of a general
cross flow heat exchanger is solved numerically. One
possible application of such a model is to generate
data that can be used to compare and tune more
simple dynamic models based on either black box
methods or state space modelling. An important
application in this context involves methods to detect
fouling in heat exchangers under dynamic operation.
Description of the model can be seen in [4].
Fouling in heat exchanger can be categorized in the
following categories, precipitation fouling, chemical
reaction fouling, corrosion fouling, particulate fouling,
biological fouling and freezing fouling. Usually fouling is
a combination of the categories. The fouling process in
heat exchanger can be described as a process where
the separating metal inside the heat exchanger
accumulates deposits from the fluids. This is very
common and poses problems and results in reduced
efficiency of the heat exchangers. There are numerous
methods available to address the effect of fouling, see
[58]. Finally, decrease in the thermal efficiency of a
heat exchanger due to property changes in a working
fluid will have similar effect on the heat exchanger as
fouling.
There are number of ways to detect fouling but
according to [9], classical methods involve
a) examination of the heat transfer coefficient,
b) simultaneous observations of pressure drops and
mass flow rates, c) temperature measurements,
d) ultrasonic or electrical measurements and
e) weighing of the heat exchanger plates. Methods
ac) require the heat exchanger to be operating in
steady state condition, d) can only monitor local fouling
and e) requires the process to be stopped. These
restrictions can be too strict or costly. Another
September 5
th
to September 7
th

306

approach is to model the heat exchanger and look for
discrepancy between model predictions and what is
actually measured, see [10] and [4]. The method used
in this study falls into category a). To make the method
valid for dynamic operating conditions, empirical
relations for the mass flow rates are furthermore used.
Although district heating systems usually operate in
relatively steady state it can be argued that methods
that work well to detect diminishing efficiency under
dynamic operation should work very well under steady
state condition.
DATA USED
The data used in this study was the same data as was
used in [4]. The data was generated by a simulator
representing an unmixed cross flow heat exchanger.
The advantage of using simulated data is that it is
possible to control when and how much fouling will
occur in addition to controlling the inlet temperatures
and the mass flows. The data used had temperatures
for the hot side in the interval [53, 67] C and the cold
side [12, 27] C, the mass flow rates for the hot and
cold side were in the interval [0.30, 1.45] kg/s.
Description of the simulator can be found in [4].
Fouling
During design a heat exchanger is commonly designed
to operate under mild fouling by assuming a fouling
factor in the interval 0.0001 to 0.0007. According to [11]
and [12] there is usually an induction time before a
noticeable amount of fouling has accumulated. In [13] it
is shown that the fouling will grow with increased rate
during the fouling period. Figure 1 shows the evolution
of the fouling factor from the time the heat exchanger
starts to accumulate fouling until the simulation is
stopped. A dimensionless time is used to make easy
comparison between different lengths of data series.

Figure 1. Evolution of the fouling factor from the time

The simulated data sets used in this study include 200
sets without fouling and 200 sets with fouling, the data
sets are further divided equally between slow and fast
fouling. In the fouled cases the data set was without
fouling for the first 25% and then the fouling factor was
allowed to progress to a maximum of R
f
=0.00033,
which corresponds to 25% decrease in the overall heat
transfer coefficient.
THE DETECTION METHOD
The fouling detection is done by estimating the overall
heat transfer coefficient, U, by using NTU relations and
monitor the means of U for shift that can be related to
diminishing efficiency either because of accumulation
of fouling or property changes of the working fluid.
NTU method is commonly known and a description of it
can be seen in [1].
It is known that effectiveness of a heat exchanger can
be calculated by
(1)
The minimum fluid is the fluid that has the minimum
value of the production of mass flow and specific heat,
. Effectiveness for a unmixed cross flow heat
exchanger can also be calculated by the following
relations of the effectiveness to NTU.
(2)
In normal use, the overall heat transfer is usually
unknown and it is therefore not possible to calculate
NTU directly. It is therefore necessary to estimate NTU
from the relation between NTU and the effectiveness.
The estimation is done by minimizing a score function
with respect to NTU. The minimization
was done by using the minimization routine fmincon in
Matlab, see [14].
The parameter NTU is defined by
(3)
From Eq. (3) it is easy to derive the formula for U
(4)
EMPIRICAL RELATIONS
In the case of heat exchanger under dynamic operation
where big variations can occur during operation, it is
hard to see shift in the overall heat transfer coefficient
that can be related to diminishing efficiency in the heat
exchanger. In [15] it is proposed to use empirical
relations of U to make a heat exchanger model valid
over a wide range of operating conditions. The heat
transfer coefficient can be written as
September 5
th
to September 7
th

307

(5)
Where i is constant, ii is temperature dependent and iii
is mass flow dependent. In this study only mass flow
dependency was considered, since it has been shown
by previously by [15] that the temperature dependency
can be neglected. The relation for the heat transfer
coefficient can therefore be written as
(6)
By assuming that Eq. (6) applies to both the hot and
the cold side and neglecting the thermal resistance in
the separating metal, the overall heat transfer
coefficient, U, can be written as

(7)
where y is the exponent of the Reynolds number. In [1]
it is recommended to use y=0.8 for turbulent flow,
which is expected in a heat exchanger.
It can be practical to normalize U with a reference
mass flow.
The overall heat transfer coefficient according to the
reference mass flow and is similarly

(8)
After inserting Eq. (7) and (8) into Eq. (4) to make it
mass flow dependent and normalizing, the estimated
overall heat transfer coefficient will become

(9)
The overall heat transfer coefficient in Eq. (9) is the
variable that is used to detect the fouling in the heat
exchanger.
RESULTS
As mentioned above the method was applied to the
same data set as was used in [4].
Measurement errors were added to the inlet and outlet
temperatures as well as the mass flows to make the
measurements more realistic. Measurement errors of
0.2 C were assumed on the temperatures and 12%
measurement errors to the mass flows.
The effect of the empirical relations can be seen in
Figure 2. It can clearly be seen that including the
empirical relations really helps to reduce the variations
in the overall heat transfer coefficient.

Figure 2. The figure shows the evolution of the number of
transfer units and the overall heat transfer coefficient with
and without the empirical relations.

To detect fouling a CuSum chart is used, see [16]. The
CuSum chart was chosen since it is known to be
effective to detect shift in mean values. When using
CuSum charts it is necessary to define two CuSum
parameters, a decision limit to prevent false detection
and a reference value for deviations. Detection is made
when the cumulative sum of deviations goes over the
decision limit.
It can be seen in Figure 3 that the method is very
consistent in detecting diminishing efficiency. Figure 4
shows the detection if no empirical relations are used.

Figure 3. The CuSum chart quickly detects the shift in the
overall heat transfer coefficient.
September 5
th
to September 7
th

308

Figure 4. The CuSum chart quickly detects the shift in the
overall heat transfer coefficient.

Comparison of Figures 3 and 4 shows that it is possible
to detect fouling in heat exchangers operating in
dynamic condition with quite good accuracy by using
the NTU method and empirical relations.
In Table 1 a comparison between the method in [4] and
the method presented in this paper is shown. From the
table it is apparent that the method presented in this
paper gives better results. The fouling detection interval
for the drift corresponds to fouling factors on the
intervals [0.00002, 0.00004] and [0.00001, 0.00003]
respectively for the fast and slow fouling. The method
is therefore giving considerable better results than the
method described in [4].
Table 1: Comparison of detection time between the two
methods, where method 1 is from [4]
Method 1 Method 2
Percentiles Fast
2.5% 0.59 0.26
50% 0.83 0.35
97.5% 0.98 0.40
Percentiles Slow
2.5% 0.63 0.23
50% 0.81 0.30
97.5% 0.93 0.35

Typical fouling factors are, as stated above, on the
interval [0.0001, 0.0007]. The results therefore indicate
that the method can be used to detect fouling in cross
flow heat exchangers that are operating in non-steady
state condition prior to the time a typical fouling factor
heat exchangers are designed for is reached. The
method seems also to be very stable in detecting the
fouling.
DISCUSSION AND CONCLUSION
The results indicate that the method proposed can be
used to detect fouling in cross flow heat exchangers
operating under dynamic condition by using
measurements that can be obtained under normal
operation. The detection method is based on the well
known method of Number of Transfer Units, with
addition of empirical relations to make the method valid
over wide range of mass flow rates.
By monitoring the calculated overall heat transfer
coefficient, it is possible to detect changes that are due
to fouling or changes in the working fluid. Unlike
conventional methods, this method can detect fouling
in heat exchangers that are not operated in steady
state conditions. The fouling detection is performed
within the designed fouling factor interval.
Further work will include application of the method on
data from a real heat exchanger.
ACKNOWLEDGEMENT
This work has been supported by the Environmental
and Energy Research Fund of Orkuveita Reykjavkur,
National Energy Fund and Energy Research Fund of
Landsvirkjun.
REFERENCES
[1] J. P. Holman Heat Transfer. Ninth edition, McGraw
Hill, 2002.
[2] M. Mishra, P. K. Das and S. Sarangi. "Effect of
temperature and flow non-uniformity on transient
behaviour of crossflow heat exchanger".
International Journal of Heat and Mass Transfer,
2008, p. 2583-2592.
[3] H. Kou and P. Yuan. "Thermal performance of
crossflow heat exchanger with nonuniform inlet
temperatures". International Communications in
Heat and Mass Transfer, 1997; 51(9-10):357-370.
[4] O. Gudmundsson, H. Palsson and O. P. Palsson.
"Simulation of fouling in cross-flow heat exchanger
and a fouling detection based on physical
modeling". In: Proceeding of The 50th Conference
on Simulation and Modelling, Fredericia, Denmark,
7-8th of October, 2009.
[5] W. L. Pope, H. S. Pines, R. L. Fulton and P. A.
Doyle. "Heat exchanger design "why guess a
fouling factor when it can be optimized?". Energy
Technology Conference and Exhibition. Huston,
Texas, 1978.
September 5
th
to September 7
th

309

[6] A. Nejim, C. Jeynes, Q. Zhao and H. Mller-
Steinhagen. "Ion implantation of stainless steel
heater alloys for anti-fouling applications". In:
Proceedings of the International Conference on Ion
Implantation Technology, 1999;2:869-872.
[7] P. K. Nema and A. K. Datta. "A computer based
solution to check the drop in milk outlet
temperature due to fouling in a tubular heat
exchanger". Journal of Food Engineering.
2005;71:133-142.
[8] S. Sanaye and B. Niroomand. "Simulation of heat
exchanger network (HEN) and planning the
optimum cleaning schedule". Energy Conversion
and Management. 2007; 48:1450-1461.
[9] G. R. Jonsson, S. Lalot, O. P. Palsson and B.
Desmet. "Use of extended Kalman filtering in
detecting fouling in heat exchangers". International
Journal of Heat and Mass Transfer, July,
2007;50(13-14):2643-2655.
[10] O. Gudmundsson, O. P. Palsson, H. Palsson and
S. Lalot. "Fouling detection in a cross flow heat
exchanger based on physical modeling". In:
Proceeding of Heat Exchanger Fouling and
Cleaning, Schladming, Austria, 14-19th of June,
2009.
[11] B. Bansal and X. D. Chen. "Fouling of heat
exchangers by dairy fluids a review". In:
Proceeding of Heat Exchanger Fouling and
Cleaning Challenges and Opportunities, Kloster
Irsee, Germany, June 5-10, 2005.
[12] F. Fahiminia, A. P. Watkinson and N. Epstein.
"Calcium sulfate scaling delay times under sensible
heating conditions". In: Proceeding of Heat
Exchanger Fouling and Cleaning Challenges and
Opportunities, Kloster Irsee, Germany, June 5-10,
2005.
[13] M. W. Bohnet. "Crystallization fouling on heat
transfer surfaces 25 Years research in
Braunschweig". In: Proceeding of Heat Exchanger
Fouling and Cleaning - Challenges and
Opportunities, Kloster Irsee, Germany, 5-10th of
June, 2005.
[14] MathWorks http://www.mathworks.com/. 20th of
April 2010.
[15] G. R. Jonsson and O. P. Palsson. "Use of
empirical relations in the parameters of heat-
exchanger models". Industrial and Engineering
Chemistry Research, June, 1991;30(6):1193-1199.
[16] NIST/SEMATECH e-Handbook of Statistical
Methods, April 30, 2009,
http://www.itl.nist.gov/div898/handbook/.
September 5
th
to September 7
th

310

HEAT LOSS ANALYSIS AND OPTIMIZATION OF A FLEXIBLE PIPING SYSTEM
J. Korsman
1
, I.M. Smits
1
and E.J.H.M. van der Ven
2

1
Liandon B.V.
2

ABSTRACT
The object of this paper is to evaluate heat losses of a
flexible PB-PE-PE piping system in the field, compared
to a conventional rigid Steel-PUR-PE piping system.
The flexible system is optimized in both insulation
quantity (thickness) and quality.
The heat loss for pairs of pipes in the field, with 70 C
supply and 40 C return temperature, is based on heat
loss measurements in the laboratory and has been
evaluated using the multipole method.
Since the hydraulic properties of Polybutylene and
steel medium pipes differ, hydraulic calculations of a
demonstration distribution network, fitted with either
system, are made.
Total system heat losses for this demonstration
network are calculated by summing the product of the
heat loss per pair of pipe and the amount of pipe used.
INTRODUCTION
Flexible piping systems for district heating and cooling
have several advantages when compared to rigid
piping, mainly during installation, some even in use.
Flexible pipe can be utilized much like cable, arrives on
large reels, requires less engineering and fewer has
joints. However, flexibility comes at a price. It seems
harder to reach comparable levels of insulation, see
Smits et al. 2010 [1].
The reason for this lies in the specific properties of the
material most commonly used for insulation:
Polyurethane foam. PUR foam has a crystalline
structure and tends to be quite rigid. It is not very
suitable for flexible applications. Bending may lead to a
breakdown of the crystalline structure and may also
compromise the bonding between foam and medium
pipe, thus creating a channel. This channel may
accelerate the exchange of foaming agent and air with
the environment, thereby speeding up the ageing
process. Flexible variants of PUR are available, but do
not seem quite as good. Insulation foam made of
polyolefins show ample flexibility and quite good
insulation properties for small diameters. Furthermore,
aging typically is a faster process than in rigid systems.
As for the medium pipe, metals may be flexible enough
for the smaller diameters, but are too rigid for the
bigger pipes. Again, using polyolefins like Polyethylene
(PE), cross linked Polyethylene (PE-X), Polypropylene
(PP) or Polybutylene (PB) improves flexibility. From
these, PE does not have adequate strength at higher
temperatures and PP is rather stiff. This leaves PE-X
and PB, of which the latter can be welded without
difficulty. This is therefore the material of choice for this
study. In accordance with the temperature duration
profile mentioned in the BRL5609/EN15632, PB is
suitable up to a maximum temperature of 95 C.
As with other plastics, PB is prone to some diffusion of
oxygen and water vapor. These effects have been
investigated by Korsman et al. 2008 [7]. To prevent
oxygen diffusion, an EVOH oxygen diffusion barrier
may be used. Unless fully submerged for years, the
diffusion of water vapor will not be much of a problem.
When kept completely under water, it will take at least
30 years for all cells to fill with condensate.
It is not all that easy to compare the heat loss of a
conventional SteelPURPE piping system to a flexible
PBPEPE piping system. Internal diameters differ, as
does the friction coefficient, because PB is smoother
than steel. A pipe for pipe comparison yields skewed
results. One way around this problem is to compare
complete distribution systems, as was done in
Korsman et al. 2008 [2]. A demonstration (or reference)
network is used to design and compare similar
networks. For reference purposes, the same network is
used in this study.
The differing properties that complicate comparison
between piping systems, can also be used to minimize
distribution system heat loss. The object of this study is
to reach comparable heat loss for the flexible system,
by exploiting specific properties, whilst transporting the
same amount of heat with comparable pressure losses.
1. HEAT LOSS IN THE GROUND
Heat losses have been measured on test rigs as
described by van Wijnkoop et al. 2010 [3] and have
been evaluated by van der Ven et al. 2010 [4].
With the results of these tests, the in-ground heat
losses are calculated using the multipole method by
Johan Cleasson and Camilla Persson in 2005 [5]. Note
that the mentioned heat losses are calculated for a pair
of pipes, run at 70 C supply and 40 C return
temperature.
For rigid piping, some room between the pipes is
required for welding, see Fig. 1a.
September 5
th
to September 7
th

311

Fig. 1a, Supply and return pipe
For flexible systems, there is no space requirement
between the pipes for installation purposes, and pipes
are best installed right next to each other, see Fig. 1b.

Fig. 1b, Supply and return closely together
When supply and return are installed closely together
the temperature profile in the ground is altered in
benefit of the in-ground heat loss.
Dependant on the refill, it may be difficult to achieve
defined compaction when supply and return are
installed too closely together. However, similar
temperature effects can be reached by installing
likewise in vertical orientation.
As mentioned before, the in-ground heat losses are
calculated using the multipole method by Johan
Cleasson and Camilla Persson in 2005 [5].
See Fig. 2 for a calculated temperature profile.

Fig. 2, Temperature profile supply and return closely
together in the ground
The installation of a supply and return closely together
in the ground has a small, but positive effect on the
heat loss of the pair, as is shown in Fig. 3.
0
5
10
15
20
25
30
35
40
45
50
16 20 25 32 40 50 63 75 90 110
Nominal diameter [mm]
H
e
a
t

L
o
s
s

[
W
/
m
]
PB std insulation thickness PB close

Fig. 3, Heat loss per pair, standard and close together
For the rest of this paper, flexible pipes are supposed
to be installed closely together.
3. PIPE PER PIPE COMPARISON
Even though internal diameters and friction coefficients
are different between Steel-PUR-PE and PB-PE-PE
and therefore will lead to a different selection of
diameters in the engineering process, an approximate
comparison can be made, see Fig. 4.
0
5
10
15
20
25
30
35
40
45
50
16 20 25 32 40 50 63 75 90 110
H
e
a
t

L
o
s
s

[
W
/
m
]
St std PB close

Fig. 4, Heat loss per pair, flexible PB versus Steel
On the left of the graph, two diameters are included for
which no steel counterpart has been incorporated. The
reason for this is that PB allows for higher fluid
velocities. However, in the current range, using the
smaller diameters does not create a heat loss
reduction.
September 5
th
to September 7
th

312

4. SELECTION OF INSULATION THICKNESS
The reason for the relatively high heat losses for the
two smallest diameters is explained by Fig. 5.
0
5
10
15
20
25
30
35
40
45
16 20 25 32 40 50 63 75 90 110
T
h
i
c
k
n
e
s
s

i
n
s
u
l
a
t
i
o
n

[
m
m
]
PB std insulation thickness PB extended insulation thickness

Fig. 5, Insulation thickness, standard and increased
In red, this graph contains the insulation thickness of
the current PB-PE-PE product range. The somewhat
erratic distribution of insulation thickness over the
range is caused by the use of customary dimensions
for the outer casing. It can be seen that for the two
smallest diameters the insulation is rather thin, which
explains the relatively high heat losses.
In purple, the graph in Fig. 5 shows a modified range,
with increased insulation thickness for some of the
smaller diameters, as in general it is easier to achieve a
good insulation quality for the smaller dimensions. The
heat losses of the modified range were calculated and
are presented in Fig. 6.
0
5
10
15
20
25
30
35
40
45
50
16 20 25 32 40 50 63 75 90 110
H
e
a
t

L
o
s
s

[
W
/
m
]
St std
PB std insulation thickness
PB extended insulation thickness

Fig. 6, Heat loss per pair, including increased insulation
thickness
The graph in Fig. 6 shows that the modified range (in
blue) is much closer to the reference (Steel-PUR-PE, in
purple). Heat loss is (almost) proportional to the
diameter, which seems about right. However, the
modified range of insulation thickness in Fig. 5 (purple)
still shows a somewhat erratic distribution, which
suggests that the current range of customary
dimensions for the outer casing does not lead to an
optimal distribution of insulation thickness. It may prove
worthwhile to develop a new range of outer
dimensions, adapted to the diameter of the medium
pipe.
5. IMPROVEMENT OF INSULATION QUALITY
Due to the new testing facilities described by van
Wijnkoop et al. in [3], the process of product
improvement has been speeded up considerably.
During the course of the investigations, resulting in this
paper, it is becoming clear that further improvement of
the insulation quality is feasible. The measurement
principle used for the determination of heat losses does
not allow for direct measurement of the insulation
properties of the foam; however, some sort of
equivalent lambda can be derived from the data by
calculation. As explained by van der Ven et al. in [4],
insulation quality differs for different diameters. For
production reasons, it is not expected that insulation
quality will reach the same level over the entire product
range. Typically, the higher values will be reached in
the smaller dimensions. Still, an educated guess can
be made as to which levels are feasible from a
technical viewpoint, see Table 1.
Table 1, Improved insulation quality, equivalent or
synthetic lambda at 50 C mean temperature
Type Area Lambda fresh Lambda Degassed
50A25 1074 0.0283 0.0326
63A32 1701 0.0287 0.0330
75A40 2364 0.0291 0.0335
90A50 2993 0.0295 0.0340
90A40 3670 0.0300 0.0345
90A32 3886 0.0301 0.0346
125A63 6204 0.0316 0.0363
160A90 10790 0.0345 0.0397
160A75 12611 0.0357 0.0411
200A110 16879 0.0385 0.0442
Please note: The lambda values in Table 1 are not the
measured lambdas of samples of the insulation foam,
and may not be interpreted as actual physical
properties of the insulation material. The values were
calculated on the basis of heat loss measurements of
sections of pipe according to EN15632, and therefore
are some sort of synthetic system lambdas.
September 5
th
to September 7
th

313

The values presented do of course largely depend on
the actual physical lambdas (in W/m.K) of the insulation
material, but the underlying measurement data suggest
that other factors come into play as well, such as the
geometry of foam in combination with the temperature
dependence of the physical lambda of the foam.
Therefore, the values in Table 1 are valid only for
calculation / prediction purposes, in exactly the same
calculation model from which they were derived (also
according to EN15632). The values in Table 1 are
supported by experimental data on four samples at the
time of writing this paper.
When all parameters are known, equation 1 can be
used to calculate heat loss:
o
i
( )
2t T
probe
T
casing
( )
s
ln
d
2
d
1
|
\
|
|
.
i
ln
d
3
d
2
|
\
|
|
.
+
1
c
ln
d
4
d
3
|
\
|
|
.
+
:=
Where:
- T
probe,
T
casing
represent probe (medium) and casing
temperature
- d
1
to d
4
represent inner/outer diameters of service
pipe and casing
-
s
,
i
,
c
= heat coefficient of service pipe,
insulation and casing
In this case,
s
and
c
are known:
s
= 0.19 W/m.K
and

c
= 0.40 W/m.K. On a test rig, T
probe
, T
casing
and
heat loss are measured, so for specific test samples,
eq. 1 can be used backwards to calculate synthetic
values for
i
in Table 1.
For the Steel-PUR-PE reference, see [1], values can
be determined in a similar fashion. Synthetic
i

values for PUR foam, determined from
measurement of samples, were typically in the range
of 0.030 to 0.032 W/m.K, with
s
for steel 50 W/m.K
0
5
10
15
20
25
30
35
40
45
50
16 20 25 32 40 50 63 75 90 110
H
e
a
t

L
o
s
s

[
W
/
m
]
St std
PB std insulation thickness
PB extended insulation thickness
PB impr. fresh

Fig. 7, Heat loss per pair, including improved insulation
quality
The red graph in Fig. 7 represents the predicted heat
loss values for the combined effect of both increased
insulation thickness and insulation quality
improvement. For most diameters, these are on par
with or slightly better than the reference in Steel-PUR-
PE. These data are valid only for the recently produced
or fresh product. As there is no experimental data
available on the rate of degassing and therefore the
rate of ageing, it is difficult to predict heat loss over the
life time of the product.
However, it is possible to speed up the process of
ageing artificially, until all the foaming agent has been
replaced by air. The predicted values for this condition
are also presented in Table 1, as lambda degassed.
These are synthetic as well, and suitable for
calculation purposes only. Calculated heat loss results
with these values are presented in Fig. 8.
0
5
10
15
20
25
30
35
40
45
16 20 25 32 40 50 63 75 90 110
H
e
a
t

L
o
s
s

[
W
/
m
]
St std PB impr fresh PB impr degassed
Fig. 8, Heat loss per pair, including improved insulation
quality, fully degassed
In Fig. 8, the purple graph represents the reference,
Steel-PUR-PE as measured, see Smits et al. 2010 [1].
The red graph represents the prediction of improved,
fresh PB-PE-PE and green the prediction of fully
degassed PB-PE-PE. The values vary a bit, but are
generally in the same range. During the lifetime of the
product, heat loss is expected to increase from the red
values to the green values.
Of course, ageing is also applicable to the reference
product, but not included here for two reasons. First,
the ageing process for rigid systems is expected to be
significantly slower than for flexible systems, and
second, the reference samples were not fresh, as could
be judged by the gas content. Therefore, it is not likely
that the values presented for the reference system will
deteriorate much further during lifetime.
Ageing can be slowed down considerably if measures
are taken to prevent the exchange of blowing agent
with the environment. If successful, these measures
(1)
Pump I, power: 16 kW
14 %
2 %
84 %
Capital
O&M
Energy

September 5
th
to September 7
th

314

would result in the red heat loss values during
lifetime. Moreover, a new generation of blowing agents
is under development. These new agents aim at lower
conductivity values for the gas and larger molecules.
This may result in lower conductivity values for the
product as well as a slower ageing process.
6. HYDRAULIC CALCULATIONS
The pipe per pipe comparison between Steel-PUR-PE
and PB-PE-PE as demonstrated in Fig. 8, gives an
indication of field results, but is not conclusive. Internal
diameters differ, as do friction coefficients. Therefore,
for the comparison between distribution systems fitted
with either pipe, hydraulic calculations are needed. To
this end, a reference network is introduced in Korsman
et al. 2008 [2]. The same network is used here. It is
installed in a housing estate near Arnhem, the
Netherlands, and has been designed using Pipelab,
developed by Prof. Dr. Pll Valdimarsson in 1995 [6].
See www.pipelab.nl. Standard design criteria were
used. A total of 247 houses are connected by 3.02 km
of DH network (6.05 km of pipe), 12.2 m per house.

Fig. 9, Aerial photograph of reference housing estate

Fig. 10 shows an output graph of Pipelab.

Fig. 10, Design pressure drop steel network
The graph in Fig. 10 represents the pressure in the
supply network (in m water column), as a function of
the distance from the source. For standard symmetrical
networks, the return network is similar, but mirrored
over a horizontal axis.
Using the flexible and smooth PB pipes allows for
smaller diameters, mainly because PB is less prone to
the transmission of hydraulic noises. This is due to the
low modulus of elasticity of PB when compared to
steel. In contrast, a steel pipe filled with water is quite a
good conductor of sound. To prevent noise caused by
high flow velocities, these are limited in the design for
steel networks to 1 m/s.
A network, specifically designed for PB, is shown in
Fig. 11. Smaller diameters in the periphery of the
network as a result of a higher permitted fluid velocity
causes higher pressure drops. This has to be
compensated by bigger pipes closer to the source to
reach the same overall pressure drop.

Fig. 11, Design pressure drop PB network

In the design of district heating networks, the maximum
design point is chosen considerably below the sum of
the installed power in the connected buildings. It is not
uncommon to have a design point of 50% of the total
installed power for larger numbers of connections,
depending on the experience and the courage of the
designer. A design point of 50% of the total installed
power was used in both designs in this paper. In
practice, no problems have arisen with this design
point, partly because not all installed power is used at
the same time. However, this statistical effect does not
apply to individual connections. Therefore, a design
trick is used in the periphery of the network, to prevent
problems in the service pipes connecting the buildings.
The flow in these pipes is raised artificially above the
design point, up to 100% load. The result of this
calculation is shown in Fig. 12, which can be compared
to Fig. 10.

September 5
th
to September 7
th

315

Fig. 12, Pressure drop steel network with increased
service pipe flow
There is not a lot of difference between both graphs in
Fig. 10 and Fig. 12. The reason for this is that the
design maximum fluid velocity is rather low for steel.
This may prove different for the PB network, which is
designed with smaller diameters in the periphery. See
Fig. 13, which can be compared to Fig. 11.
The graph in Fig. 13 indeed shows an increased
pressure drop in the service pipes connecting the
houses, when the flow in those pipes is artificially
increased to 100% of the installed power. However, the
total pressure drop stays within the same limits as does
the steel network under similar conditions (Fig. 12).

Fig. 13, Pressure drop PB network with increased service
pipe flow
The result of both design calculations is plotted in
Fig. 14, steel in red and PB in green.
0
200
400
600
800
1000
1200
16 20 25 32 40 50 63 75 90 110
P
i
p
e

L
e
n
g
t
h

[
m
]
(1) ST.PUR.PE Ref (2) PB as measured std eng.

Fig. 14, Pipe length histogram Steel and PB
As a result of the use of smaller diameters with PB, the
distribution of pipe lengths generally shifts to the left in
the pipe histogram. As heat loss increases with
diameter (see Fig. 8) this should have a positive effect
on the total distribution system heat loss.
This shift to the left may be taken one step further,
since the wall thickness of the smallest PB medium
pipes currently is chosen a bit larger than the strength
class (SDR11) requires. This is done for ease of
installation. If the thickness of these pipes is chosen no
larger than SDR11, there is a slight additional shift to
the left, see Fig. 15.
0
200
400
600
800
1000
1200
16 20 25 32 40 50 63 75 90 110
P
i
p
e

L
e
n
g
t
h

[
m
]
(1) ST.PUR.PE Ref (2) PB as measured std eng.
(3) PB impr. Fresh all SDR11

Fig. 15, Pipe length histogram steel, PB and PB SDR11
networks

September 5
th
to September 7
th

316

6. TOTAL SYSTEM HEAT LOSS
To calculate the total system heat loss, pipe lengths as
shown in the pipe length histograms are to be
multiplied by the respective heat losses per pipe pair,
as shown in the heat loss value histograms.
Fig. 15 multiplied by Fig. 6 leads to the total system
heat losses in Fig. 16.

0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
system [-]
H
e
a
t

L
o
s
s

[
k
W
]
ST.PUR.PE Ref
PB as measured std eng.
PB increased insulation thickness
PB increased insulation thickness all SDR11

Fig. 16, Total system heat loss, current insulation quality
The graphs in Fig. 16 show the reference heat loss for
steel-PUR-PE in red and the currently measured heat
loss for PB-PE-PE with non-optimized insulation
thickness in orange. In green, total system heat loss is
shown for current insulation quality but with optimized
insulation thickness. The exclusive use of SDR11
(blue) has a rather small effect.
Improving insulation quality, as described in paragraph
5 and shown in Fig. 7, leads to slightly lower total
system heat loss for freshly produced PB-PE-PE when
compared to the reference, see Fig. 17.
0.0
10.0
20.0
30.0
40.0
50.0
60.0
system [-]
H
e
a
t

L
o
s
s

[
k
W
]
ST.PUR.PE Ref PB impr. insulation thickness/quality

Fig. 17, Total system heat loss, reference and prediction
for improved insulation quality
There is currently no experimental data on the rate of
ageing of PB-PE-PE as a result of the exchange of
blowing agent with air. However, it is possible to
calculate a worst case situation (see fig. 18), using the
predicted values plotted in fig. 8.

0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
system [-]
H
e
a
t

L
o
s
s

[
k
W
]
ST.PUR.PE Ref
PB impr. insulation thickness/quality
PB impr. Insulation thickness/quality degassed

Fig. 18, Total system heat loss, including worst case
In practice and over time, the predicted total system
heat loss will slowly shift from the fresh value in purple
to the worst case value in blue. Average heat loss
during lifetime will be somewhere in-between.
FUTURE RESEARCH
Hydraulic calculations in combination with insulation
thickness form an interesting optimization problem:
what diameter to select and which insulation thickness
to choose?
Current design strategies for hydraulic networks,
aiming at linear pressure drop with distance, seem too
adventitious to be optimal. In addition, heat loss
calculations using standard casing dimensions show
rapidly diminishing yields with each step up in
insulation thickness, suggesting the optimum is
somewhere in-between.
First attempts have been made to use Pipelab [6] in a
double optimization routine, trying to find optimal
hydraulic performance in combination with optimal
insulation thickness distribution over the network.
Given the specific hydraulic properties of PB (high fluid
velocities permitted) and the specific insulation
properties of PE foam (better at small size), this may
lead to rather different design strategies when
compared to conventional rigid piping systems for
district heating and cooling.
September 5
th
to September 7
th

317

FURTHER INFORMATION
Liandon B.V.
www.liandon.com
www.pipelab.nl
Dijkgraaf 4
6920 AB Duiven
The Netherlands
www.thermaflex.com
Veerweg 1
5145 NS Waalwijk
The Netherlands
CONCLUSION
Flexible Polybutylene piping, insulated with
Polyethylene foam is a recent development, leaving
ample room for product improvement. Experimental
data shows rapid improvement in heat loss
performance.
By optimizing both quantity (thickness) and quality of
flexible Polyethylene foam, and by using the specific
hydraulic properties of Polybutylene piping, the heat
loss performance of conventional rigid steel piping
systems insulated with polyurethane foam, is within
reach.
If the current rate of improvement of the PB-PE-PE
flexible piping is maintained, total distribution system
heat losses will be comparable to conventional rigid
Steel-PUR-PE piping. The evident benefits of flexibility
would become available without the current heat loss
penalty.
ACKNOWLEDGEMENT
Special thanks are due to Ivo Smits who did all the
calculations for this paper and made all the graphs. A
small change in approach may result in a lot of
recalculation. Thanks also to Camilla Persson for
supplying us with a MathCad implementation of the
multipole method back in 2008 and Pll Valdimarsson
for the invention of Pipelab, which set me off on this
road at the symposium in Helsinki, 1995.
REFERENCES
[1] I.M. Smits, J.T. van Wijnkoop, E.J.H.M. van der
Ven, Comparison of competitive (semi) flexible
piping systems by means of heat loss
measurement, in Proc. of the 12th International
[2] J. Korsman, I.M. Smits, S. de Boer, System
optimization of a new plastic piping system, in
District Heating and Cooling, Reykjavik, Iceland
(2008).
[3] J.T. van Wijnkoop, E.J.H.M. van der Ven,
Verification of heat loss measurements, in Proc.
of the 12th International Symposium on District
Heating and Cooling, Tallinn, Estonia (2010).
[4] E.J.H.M. van der Ven, R.J. van Arendonk, Heat
loss of flexible plastic pipe systems, analysis and
optimization, in Proc. of the 12th International
[5] J. Claesson, C. Persson, Steady-state thermal
problem of insulated pipes solved with the
multipole method, Chalmers University of
Technology, Report 2005:3. (2005)
[6] P. Valdimarsson, "Graph-theoretical calculation
model for simulation of water and energy flow in
district heating systems", in Proc. of the 5th
International Symposium on Automation of District
Heating Systems, Helsinki, Finland. (1995).
[7] J. Korsman, S. de Boer, I.M. Smits, Cost benefits
and long term behaviour of a new all plastic piping
system, IEA DHC|CHP Annex VIII research report
(2008)
September 5
th
to September 7
th

318

FREE OPTIMIZATION TOOLS FOR DISTRICT HEATING SYSTEMS
Stefan Gnchtel
1
, Sebastian Gro
1

1
Institute of Power Engineering, Technische Universitt Dresden, 01062 Dresden, Germany

ABSTRACT
At the Technische Universitt Dresden, Institute of
Power Engineering, Chair of Power Systems
Engineering as part of the project LowEx Fernwrme
Multilevel District Heating [1] supporting by the
Federal Ministry of economy and technology (FKZ
0327400B), two public available and cost free software
tools have been developed, which enable the user to
find
A) the optimal unit commitment of district heating
generators: FreeOpt
B) the optimal pipeline route with the optimal pipe
diameter of district heating networks: STEFaN
at a minimum of costs. Both software tools are very
easy and intuitive to handle. Lead time required to
learn how to operate the programs is short. Both tools
intend to support general design decision pro district
heating systems. In this paper an overview on their
advantages and fields of application, as well as
example calculations are presented.
PROGRAM FREEOPT
FreeOpt is an optimization tool to find the optimal unit
commitment of district heating generators for a given
time domain. Optimal decisions are found to minimize
total costs related to thermal and electric loads.
Local district heat networks are becoming more
common, so it is important to know how to operate
even small systems in terms of minimal costs and
highest efficiency. When should which generator be
switched on or off? How to handle the storage? Which
influences have contracts for electric power?
There already is a lot of existing software for unit
commitment. As commercial and generalised software
for large systems it is mostly very expensive. So
FreeOpt has been developed for any cases of district
heating networks for most efficient operation of all heat
and power generators. Mainly operators of smaller
supply areas purchasing an expensive software
solution would be uneconomical can reach monetary
savings.
A stable version of FreeOpt is already finished. The
programs power is demonstrated with a simple
example determining cost optimal operation of a
specific district heating network. Furthermore with the
help of parameters and figure lines it is very easy to
modify the system configuration to check how the
system reacts under new conditions and how the
operating costs will change. In all cases the tool gives
valuable information.
So FreeOpt provides help for any municipality that
need a first guess on the feasibility and operating costs
of a new district heating network or who need to
improve the operation of an existing one. And of course
it does not matter if the network is supposed to be
extended or build from scratch or an existing system
just to be analysed.
How does the software tool work?
First the simulation time is divided into time steps. For
every time step several variables (i.e. generated heat
and electric power, amount of fuel, energy
consumption) exist inside given boundaries as well as
individual costs and proceeds are stated (i.e. for fuel,
CHP refund). Different kinds of generators can be
chosen like heat plants, combined heat and power
plants, solar thermal plants or heat pumps. Hot water
storage tanks and electricity transferred in or out the
grid via contracts improve the flexibility of the whole
system (Fig. 1 ).
Block heat
and power
plant
Combined
heat and
power
plant
Heat
plant/
boiler
Grid
connection
Heat
demand
System boundary
E Qfuel
Qdem
Pdem
Solar
thermal
system
Heat
pump
Heat
storage
Power
demand

Fig. 1 System boundary and interaction plan

The modular design of FreeOpt allows to form easily
any generator system. All one has to know are the
respective figure lines and parameters for all available
generators and the network. After fixing the demand for
September 5
th
to September 7
th

319

heat and electric power as well as the parameters the
optimization problem is defined automatically. FreeOpt
calculates the minimal costs to satisfy the demand. For
testing purpose there is no limitation of the values.
The mathematical model belongs to the mixed integer
problems. So the objective function and all constraints
are linear, all variables are continuous or discrete. It is
written in the mathematical modelling language GMPL
[2] and solved with the COIN-OR brunch-and-cut solver
CBC [3]. Both GMPL and CBC are open-source
software under GNU GLP license [4]. An intuitive user
interface enables to enter all input data like variable
boundaries, cost coefficients, starting values, demand
values or figure lines in a very easy way. The internal
data flow between user interface, optimization model
and solver is realised with help of txt-files (Fig. 2).

Fig. 2 FreeOpt user interface

Running the generators cause costs. Therefore the
main aim is to minimize the total cost
(1)

in which are the following operating costs (and
proceeds):
- costs for fuel
-

costs or proceeds for transferred
electricity
-

costs for start-up procedures
- costs for network access
-

costs for maintenance
- costs for CO
2
-cerfiticates
-

CHP-refund
-

EEG-refund
-

proceeds for avoiding network access
- costs for electricity tax
- penal costs for balance violation
(virtual costs)
As already noted several variables exist for every time
step limited by some boundaries and connected by
parameters in lots of equations and inequations.
The most important ones are the two balance
equations for power (2) and heat (3).
(2)
stands for the generated electric power, for
power transferred in or out the grid via contracts,
for the electric power demand and for the own
consumption of electric power.
(3)
stands for the generated thermal power, for
the heat demand, the heat disposed in the auxiliary
cooler and for the heat transferred in or out of the
heat storage.
Example
As an example the following heat network of a local
energy supply company is given in Fig. .

CHP boiler
storage
CHP 1 Heat plant
Heat
storage
Electricity network
CHP 2 CHP 3

Fig. 3 Flow scheme of a local heat network

The heat demand of the customer is provided by
- 3 CHPs (Block Heat & Power Plants) base
load
- Heat plant (natural gas) peak load
- Heat storage used for optimization
and is given for every hour of one year.
The electricity demand is not directly considered
because of intern clearings inside the energy supply
company. The whole generated power is transferred in
the grid and refunded as well as used to satisfied the
own consumption. It is also possible to transfer
electrical power out the grid when all CHPs are
switched off.
September 5
th
to September 7
th

320

Tab.1 gives an overview on the design parameters of all
heat and power generation units as well as on the cost
assumptions for fuel and the refund of the energy tax.
Tab. 1 Design parameters of generation units

One specific characteristic of the energy system is that
all three CHP cannot be operated in part load, so the
value of the maximum and the minimum power have to
be equal.
The base own consumption of the whole energy
system is 50 kW. In the 100 m
3
heat storage it is
possible to store 4000 kWh. Tab.2 gives an overview of
all intern electricity contracts.

Tab. 2 Electricity contracts

The CHP-refund is 0.0056 /kWh and costs for network
access are 0.0386 /kWh.
Analysis operation mode
First the real operation mode of whole year 2008 is
analysed retrospectively (case I) and the optimal
operation mode is determined with the help of FreeOpt
(case II). Following the annual operation costs are
calculated for both cases (Tab. 3).
The calculations in Tab.3 show that the costs for fuel
increase but so the proceeds through electricity sale
and CHP-refund increase too.
Tab. 3 Comparison annual operating costs and proceeds
(case I and case II)

Annual operation
costs and
proceeds / 10
3

Real operation
mode
(case I)
Optimal
operation
mode
(case II)
Costs fuel 1326.12 1331.13
Proceeds
electricity
contracts
468.48 473.65
CHP-refund 58.17 58.67
Start up costs 10.01 6.62
Costs network
access
2.83 2.47
Total costs 812.30 807.90

Furthermore costs for start ups and network access
can be reduced. But altogether the total annual
operating costs for case I and case II are nearly equal.
In this particular example savings are under 1% of the
operation costs. So the energy system is operated
nearly in an optimal way.
System configuration changes
The energy supply company considers to replace the
CHP 2 through a new one. Of course the new CHP has
new parameters (Tab. 4).
Tab. 4 System configuration changes CHP 2
Generation
unit
CHP 1 CHP 2 CHP 3 Heat
Plant
/ kW 911 774 911 -
/ kW 911 774 911 -
/ kW 1200 1020 1200 200
/ kW 1200 1020 1200 2000
/ %
0.35
( )
0.35
( )
0.35
( )
0.87
( )

0.76 0.76 0.76 -
Own
consumption

2.0 %
(of )
2.0 %
(of )
2.0 %
(of )
0.5 %
(of
)
Costs fuel /
/kWh
0.04444 0.04444 0.04444
0.045
74
Refund
energy tax /
/kWh
0.0055 0.0055 0.0055
0.002
2
Start up
cost /
/start up
12 12 12 6
Electricity purchase Electricity
sale
6am 22pm 22pm 6 am
Costs /
/kWh
0.08150 0.05700 0.04735
Generation unit Old CHP New CHP
/ kW 774 404
/ kW 774 404
/ kW 1020 535
/ kW 1020 535
/ % 0.35 0.36
Start up cost /
/start up
12 9.5
September 5
th
to September 7
th

321

All other parameters are the same like the old CHP 2.
Main difference is the smaller range of performance.
The remaining energy system left unchanged except
for the CHP-refund expiring the next year.
The optimal operation mode considering the old
(case III) and the new (case IV) parameters of the
CHP 2 are determined. Case III serves as a reference
case. The calculated operation costs per week are
shown in Fig. 4 for both cases.
0
5
10
15
20
25
30
35
J F M A M J A S O N D
T
o
t
a
l

c
o
s
t

/

1
0
3
Month
Operating costs per week
Old CHP 2 (case III) New CHP 2 (case IV)

Fig. 4 Operation costs per week (case III and case IV)

In Fig. 4 it can be seen that the operation costs for
case IV are below the operation costs for case III in
every week, especial in the summer months because
the smaller size of the new CHP 2 suits better to the
heat demand. Altogether total operating costs
amounting to about 27770 (3.2 % of total costs) can
be saved (Tab.5). Main reasons are the huge fuel
savings which settle easily the decreasing proceeds
through electricity sales. By given investment cost it is
very simple to check if the renewal of the old CHP 2 is
economic reasonable.

Tab. 5 Comparison annual operating costs and proceeds
(case III and case IV)

As example the heat balance curve on one summer
week (168 hourly time steps) is shown in Fig. 5. Such a
curve helps to determine the recommended operation
mode.
The red line marks the given heat demand which is
satisfied at all time steps. The small-sized CHP 2
operates continuous. When the heat demand is higher
than the output of the CHP 2 the heat plant is switched
on or the heat storage is discharged mostly. Charging
the heat storage takes place in low demand times and
by switching on one of the others CHPs (CHP 1 or
CHP 3) for up to four hours. Using the CHP is more
economic than using the heat plant but start up costs
and the size of the heat storage restrict the operation of
a second CHP.

heat
plant
heat demand
0
200
400
600
800
1000
1200
1400
1600
1800
0 12 24 36 48 60 72 84 96 108 120 132 144 156 168
h
e
a
t

f
l
o
w

/

k
W
time / h
Heatbalance
CHP production storage load heat plant storage unload heat demand
charge of
storage
by CHP
discharge of storage
CHP production

Fig. 5 heat balance curve on one summer week

Finally it can be summarized that first experiences and
calculations show, how the FreeOpt allows in an easy
and quick way to check beforehand if certain system
configurations are useful or contra productive.
PROGRAM STEFAN
Application field
The network optimization is a special case of the
research-main focus optimization of the technical
structure of district heating systems of the 5th energy
research program of the German Federal Government.
Due to the relatively high net costs of district heating
systems it is necessary (beside the application of
actual piping systems) to optimize the nets concerning
their design parameters, in particular the pipe diameter
and the pipe routing.
Therefore the software tool STEFaN has been
created for the combined pipeline routes and diameter
optimization. This Windows program for the support of
the application of the district heating has interfaces to
geographical information systems (GIS) and is
complementary with these. Its application is possible in
3 planning phases:
Annual opera-
tion costs and
proceeds / 10
3

Old CHP 2
(case III)
New CHP 2
(case IV)
Costs fuel 1323.87 1208.79
Proceeds
electricity
contracts
462.87 367.12
Start up costs 7.72 1.20
Costs network
access
1.92 0.00
Total costs 870.64 842.87
September 5
th
to September 7
th

322

1. In the conceptual design planning phase for the
cost estimate and the principle decision for the
district heating (yes or no).
2. In the detailed planning phase the localization of
the pipeline route occurs for the approval
planning.
3. In the execution planning phase the final
determination of the dimension occurs, but no
mechanical calculation (stress-strain analysis) is
done by the program. Still the required proofs
according to e.g. EN 13941 have to be done.
In addition, the program can be used for the hydraulic
calculation of existing district heating networks.
Model
The hydraulic calculations establish the technical basis
which performs constraints of the optimization model.
In district heating systems a distinctive turbulent flow
can be presumed. In this case a good approximation
with the surface roughness of the pipe and the
coefficient of friction is applied (4).
(4)
In the mentioned planning phases the coefficients of
drag are included blanket into the pressure loss
according to (5):
(5)
where as a extra charge of length.
For a pipe of the length and the diameter (6) arises
for the pressure loss of a plain pipe.
(6)
Thus the following mathematical optimization model
arises:
The investment costs of every new route come into the
objective function (investment costs, annual costs or
net present value). They are included in the form of (7)
in the model.
(7)
The bracket of the first summand contains the
investment costs of the route per meter as a total lump-
sum price and must be multiplied according to by the
length of pipeline. In the second summand
"obstacles" can be included as direct costs dependent
from the diameter. The exponent is set =1 in the
present program version for linear dependence. The
parameters and are input data.
If the edge is not used for the site development ( ),
then holds.
As variables are required beside the diameter further
variable than auxiliary variables to the formulation of
the constraints:
- vector of the mass flows of the edges
- vector of the pressures of the vertices
- binary variable to the capture of the jump at
Thus the constraints can be formulated. These are the
equation (6), local and technical limitations as well as
equations. (8) and (9).
First Kirchhoffs law: Point rule. The sum of all mass
flows in a vertex is equal zero. ( - vertex matrix)
(8)
Second Kirchhoffs law: Mesh rule. The sum of the
pressure losses along a mesh is equal zero ( - mesh
matrix):
(9)
This mathematical model is simple to describe, but
difficult to solve (already for medium-sized graphs). If
the diameters are eliminated by the equation (6) as a
variable, the variables and whose impact on the
objective function is discussed in detail in [5] remain.
The principal dependency of the objective function on
the vector of the mass flows is displayed in Fig 6
schematically.

0
1 6
0 2 7
K konkav
K Baum

Fig. 6 Schematic dependence of the objective function on
the mass flow

On the abscissa the circulatory mass flow of a mesh
is displayed. The ordinate shows the non-convex
objective function which shows jumps by the binary
variables with the rhombuses (the filled rhombus is the
function value).
September 5
th
to September 7
th

323

This shape leads to the fact that as a solution under the
requirements mentioned above a graph free of
meshes, thus a tree, arises.
This complex course of the objective function requires
especially suitable methods.
Development of mathematical procedures
Different mathematical methods are used:
a) The classical non-linear optimization which is
applied for the diameter optimization for fixed
development ways.

b) Topological optimizations to the determination of
the shortest ways (shortest path problem) and the
shortest networks (spanning tree problem) which
are combined under use of the procedure from a)
to a special iteration process.
c) Stochastic methods for the improvement of the
optimization results of the algorithm of b): A special
implementing of the Monte Carlo Method and a
special implementing of the Evolutionary Algorithm.

Fig. 7 step 1
Project processing
The help file and the user's manual of this program
contain detailed instructions to its operation and for the
project processing which occurs typically in six steps in
the change of STEFaN (steps 2, 3, 5 and 6 a) and of a
GIS (steps 1, 4 and 6 b).
These necessary steps 1 to 6 are demonstrated at a
fictive example. As a GIS system the recommended
and provided program ShapeUp (www.nilione.com) is
used. It is freeware too.
Example
Step 1: Gathering of Geographical information
With the GIS layers (themes) with geo-referenced
information (vertices for the source and for the
customer as well as edges for consisting and possible
routes) invested and in a special standardized format
(MIF MapInfo Interchange format) exports:
a) The geo-referenced background image is imported
( OpenStreetMap, pale colors in Fig. 7). The
figures of the buildings (darkly) and the courses of
the streets (white) allow a good orientation.

September 5
th
to September 7
th

324

b) Directly input (or import) of the heat source (red
pentagon on the top left in the Fig. 7, through violet
circle marked) and the sinks (customer yellow
flags in the Fig. 7, as house service connection
stations in the house lines arranged) with the given
attribute for the heat demand.
c) The input of possible routes (thick green lines in
Fig. 7) by using the mouse and the assignment of
the attributes (table in the right section of Fig. 7),
for quite available pipes with attribute
(branch pipe from the source in
DN 150 input value: ) and
for the laying procedure ( for street and
ground, for cellar corridor and available
channel).
Step 2: Providing the non-geographical data
The files with the non-geographical data (general
entries to the network as for example media
temperatures as well as economic data) are entered on
forms.
Fig. 8 step 4

Step 3: Generating the network topography
After the import of the files invested by the GIS the
network topography of the possible routes is created by
the program. Gaps between the inputted routes (step
1c) are complemented by the program to a graph with
entire and varied development.
Step 4: Verification of the generated network
topography
With this step the files generated by the program can
be imported in the GIS: The generated edges (thin blue
lines in Fig. 8) and vertices (blue dots in fig. 8)
complete the entered network topography and can be
checked.
Step 5: Determination of the optimal development
The route optimization is carried out by the program.
Step 6: Evaluation
a) Output of a result report and export of the
optimization results to the GIS.
September 5
th
to September 7
th

325

b) The local representation and if necessary
treatment of the results by the GIS: The result can
be visualized in the GIS (Fig. 9). The different
colored lines in the left section show the
ascertained route planning, and in the right section
of Fig. 9 a part of the data base with the site-
related data is displayed: Length of pipeline in m,
the diameter in mm, the mass flow in kg/s
and the pressure difference in bar.

Fig 9 step 6
CONCLUSION
Stable versions of FreeOpt and of STEFaN are already
finished. The programs power was demonstrated with
simple examples. They can be downloaded from [6].
FreeOpt calculates the optimal operating solution of
district heating networks at a minimum of costs to estimate
saving potentials. With the help of parameters and figure
lines it is very easy to modify the system configuration to
check how the system reacts under new conditions and
how the operating costs change. In all cases the tool gives
valuable information.
Unfortunately, the user guide, the help file and the
manual are only available in German at the moment for
both programs.
REFERENCES
[1] www.bmwi.de: LowEx Fernwrme Multilevel
District Heating, Frdergeber: Bundesministerium
fr Wirtschaft und Technologie, FKZ 0327400B.
[2] www.gnu.org/software/glpk:
GNU Linear .Pro-gramming Kit
[3] www.coin-or.org: Computational Infrastructure for
Operations Research.
[4] www.gnu.org/licenses/gpl.html: GNU General
Public License
[5] S. Gnchtel, Ein Beitrag zur Strukturoptimierung
von Fernheiznetzen, PhD thesis TU Dresden
(1981)
[6] http://tu-dresden.de/die_tu_dresden/
fakultaeten/fakultaet_maschinenwesen/iet/ew/
forschung_und_projekte/mldh/download_ml

12th International Symposium On District Heating and Cooling

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

12th International Symposium On District Heating and Cooling

Uploaded by

Copyright:

Available Formats

12thInternationalSymposiumonDistrictHeating

can also be described by equation (4) and

insulance of the convective heat transfer

You might also like